Visit the Architecting Microsoft Azure Solutions course recordings page
United Arab Emirates - Architecting Microsoft Azure Solutions
WEBVTT
-->
That's it.
-->
Thank you.
-->
That's it.
-->
Okay, next question.
-->
I think like a lot of people are that like, oh, that's right.
-->
I'm gonna do something that's gonna help to make it better, right?
-->
I think that's what I'm gonna do with this is like what I think is better
-->
than just thinking paper is about when everything's all done at once, right?
-->
Like I feel like if I can bring stuff in there and then I can house all this
-->
Okay guys, good morning everyone.
-->
How's it going?
-->
Okay, can you hear me all?
-->
Okay, okay, awesome.
-->
Okay, okay, so good morning everyone.
-->
Let's start today's session.
-->
So yesterday we were discussing.
-->
Oh, yeah, go on.
-->
Oh, yeah, yeah, thank you.
-->
Okay.
-->
Okay, so yesterday we were discussing the designing of a governance solution.
-->
So in this case study, we have this company, Tailwind Traders, which has two business lines or units, one for apparel and other for sporting goods.
-->
So this company in addition to those business units have three departments, product development, marketing, and sales, and which business unit and seed unit will be responsible for tracking their actual spend.
-->
So here's my first question for you guys.
-->
What does this line means regarding subscriptions?
-->
If this line is in there, how can you or let me change or rephrase the question.
-->
How can you track the actual spending of each business unit and an or each subunit?
-->
What will you do?
-->
There is a straightforward solution and there is an additional solution, which may be a little bit more complex, but it's valid as well.
-->
So what will you do in order to solve this specific highlighted point?
-->
I've got one sticky pink.
-->
What about lean?
-->
What do you think?
-->
In order to track the Azure spending of each one of those business units and subunits, what do you think it can be done?
-->
Perfect.
-->
That is the more straightforward solution.
-->
So we can have each one of the business units to have a Azure subscription.
-->
So with that information or with those subscriptions, we can track effectively how much they are spending.
-->
So that is one solution.
-->
Which other solution do you think that we may have available besides having a subscription for business units?
-->
I don't know who wants.
-->
OK, go on, Jason.
-->
No, I think that is super valid.
-->
Actually, I was thinking about another way to do it.
-->
But yeah, that works.
-->
That one works for sure.
-->
I don't know if anyone has any other suggestion that can come up.
-->
Yeah, for sure.
-->
So we have one solution, maybe a subscription for each one business unit.
-->
Another solution may be using our resource group per business unit and per subunit.
-->
So those are perfectly valid.
-->
And so this is the goal of tracking Azure spending.
-->
So which other one do you think it may be applied?
-->
For example, this one, tagging.
-->
You may have tags so you can tag all the resource groups or all the resources according to which department are those resources in.
-->
So it may happen that you have only one subscription for the whole company, which is not ideal, but it's super common.
-->
And with tags, you tag all the resources, which belongs to a specific department.
-->
And that way you can track all the information of the spending for each one of those departments.
-->
So that is a good option to do that, right?
-->
So if you are wondering, OK, but how should I do that?
-->
I'm going to access through the desktop.
-->
So we can just take a look into the training room.
-->
Actually, I don't see anyone in the training room.
-->
So please, guys, join the training room so we can have a view of your systems.
-->
And also to make it easier for you to watch what we can do.
-->
In the meantime, while you are connecting to the training room, I'm going to show you here.
-->
For example, this resource group, any resource group that is here, and there is the tags.
-->
Here are the tags.
-->
So I can create a tag, for example, leaning account.
-->
This can be wherever you want.
-->
The cost center depends on the company, cost center.
-->
And this is marketing, for example.
-->
So this is a tag.
-->
So that is how you may want to do it.
-->
You may be able to do that.
-->
If you tag the full resource group, well, everything that is within that resource group,
-->
will get tagged with the cost center, which is in the sample.
-->
So that's it, okay?
-->
Now, let's go back here.
-->
So the final goal of this thesis study was to, or for you guys,
-->
was to provide a company-wide Azure Cost Reporting Tool.
-->
So the questions that were here for this study, for this case,
-->
was which ways could you use to organize the subscriptions and the management groups?
-->
Which would be the best to meet these requirements that we have discussed?
-->
And they proposed two alternative hierarchies to explain the decision making.
-->
So I would love for you guys, if you can share what you worked yesterday.
-->
The first group that had some, that used Gemini as a discussion partner
-->
or assistant partner in order to discuss which will be the best topology.
-->
But I would love if you guys can share that conversation with Gemini
-->
so all of us can see it, because I think that is really valuable to leverage on AI
-->
to enhance our powers, our capacities.
-->
So I don't remember which group was it.
-->
Sorry about that.
-->
But if you guys, you know who you are.
-->
So any of you, if you can share what you did and how it ended up.
-->
So please go ahead. I will stop sharing my screen.
-->
Do you have sharing permissions or is it not there?
-->
Good morning, John.
-->
That is correct.
-->
No, that is okay.
-->
We need to embrace AI, at least in my opinion.
-->
So if you can show your screen and show us the conversation, that would be awesome.
-->
Perfect.
-->
That is perfect.
-->
And those units have the permits. Awesome.
-->
Thank you.
-->
I agree. I'm still working.
-->
Okay. So let me see if we can do a really quick poll here within the Teams meeting.
-->
So let's see. Let me launch the instant poll here.
-->
Let me see if I can. Okay.
-->
So sorry.
-->
Let me delete this one.
-->
Because...
-->
Okay. So let me see.
-->
Let's vote everyone with two options.
-->
So which one would you prefer?
-->
Business unit focus, which is the one on the right.
-->
No, let me see what happens here.
-->
Okay. This one.
-->
So you can click on the thumbs up or thumbs down for the business unit option.
-->
So let me see how many people prefer the business unit option.
-->
You should have the thumbs up in your screen in your Teams.
-->
So let's see if you guys have it.
-->
Okay. Perfect. Perfect.
-->
There are four responses.
-->
So now let's see who else have this one.
-->
Okay. Well, it seems that this one has a good number of votes at least.
-->
And let's see the other one.
-->
Let's see with hearts.
-->
Which of you guys prefer the department focus, which is the one on the right.
-->
So in the instant poll.
-->
Yeah, it makes sense.
-->
But it seems that people prefer the business unit focus.
-->
I mean, there is no right or wrong answer in this case because that depends a lot into the business.
-->
But for example, in this specific situation, if you take a look,
-->
there is the business units are apparel and sports.
-->
And those are like subordinated to the department in the department focus.
-->
So it doesn't necessarily make a business sense because the business unit that is providing the budget.
-->
Probably it's apparel from one side and sports from the other.
-->
So perhaps in this specific scenario, this is not the best approach.
-->
However, it may happen in your company or in a different company that is different.
-->
It's different because the product development has several lines to work with.
-->
So they are the ones with the budget.
-->
So sales and marketing.
-->
So it really depends.
-->
It really depends on the company.
-->
So how this works.
-->
So yes, I do agree that probably the business unit focus is the best approach to manage the sure key in this situation.
-->
But that doesn't mean that the other option is wrong.
-->
It's just that it needs to be like applied to the right or to a more appropriate scenario here.
-->
The key is who is the owner of the budget?
-->
Who is the one who is in charge of the budget?
-->
So that way you can follow as they say, follow the money.
-->
Just follow the money.
-->
Okay.
-->
So let's see who the other group has.
-->
If there is any other group who has a different solution and want to share that solution.
-->
Or if any other group that has it can be the same solution.
-->
Because what I want to do now is with one of those solutions to let's place where the subscription will go.
-->
How the management group will be defined.
-->
Okay.
-->
So who wants to share the three, the here key?
-->
There was a group that was already doing that yesterday.
-->
The last group, the group four, I think.
-->
So if you want to share that with us, please.
-->
Yeah.
-->
I don't agree.
-->
Yeah, I didn't.
-->
I think I don't have the list.
-->
Let me see if I can see.
-->
No, I have the list here.
-->
I forgot.
-->
Okay.
-->
It can be Kevin or Kempi or Pierce.
-->
Any of you guys?
-->
Awesome.
-->
Thank you.
-->
I have the list here.
-->
I forgot.
-->
Oh, man.
-->
You are the greatest.
-->
So, okay.
-->
So he already has the group four with Kevin, Kempi, and Pierce.
-->
Already have the management or the root groups.
-->
So these management groups.
-->
So go ahead and explain that to us so we can discuss.
-->
This is actually what I was wanting to do.
-->
Okay.
-->
I'm sorry.
-->
Notice that in this approach, the apparel and the business units and the sports unit, sorry,
-->
doesn't have a resource group or a subscription.
-->
She is using the subscriptions and resource group directly in each one of the subunits.
-->
And that is perfectly valid.
-->
However, it may happen that, for example, apparel has some shared services that are used by all of the subunits.
-->
So in that case, you may want to have a subscription for apparel, another subscription for a sports unit,
-->
so they can provide those shared services.
-->
Also, remember that you may want to have a subscription for what it is production,
-->
and at least another production for what it is development or QA or wherever.
-->
So that may another approach, or not approach, but component in this solution.
-->
You may, perhaps it will be super easier just to have one subscription per business unit
-->
for development and testing purposes, so at the upper level,
-->
and that can be used for all of the subunits that may be perfectly valid as well with the sports unit as well.
-->
So that is what you may do here.
-->
That is really good, guys.
-->
So I will recommend you, if you guys can do that, share this file, for example, in the chat,
-->
as well as the vote diagrams from the first group, from John, Kevin, and me,
-->
and you guys, Kevin, competing first, if you can share what you did in the chat,
-->
so everyone can take a look into that and ensure that.
-->
And also, the other two groups, the one from Jason and Alexander,
-->
if you want to share something as well, it will be super awesome.
-->
Let me show you just a few more things, or if any, before doing that,
-->
if any of you guys have any questions.
-->
If there is any questions here, otherwise, let me show you this.
-->
So here we do have something that we may think about.
-->
Okay, so how do I create a management group?
-->
So here in the resources, in the resources, where is it?
-->
I can see the governance.
-->
Go to governance.
-->
So here in the documentation, we can create a management group in all services,
-->
management, and governance.
-->
Why I didn't see it?
-->
Here, let's see.
-->
All services.
-->
So all services, management, and governance.
-->
So here are all the different options for management and governance.
-->
So these are the management groups.
-->
So by default, there is one management group.
-->
So there is one management group, which is the group, the tenant group.
-->
And in this environment that we have here, we don't really have like a proper
-->
design and tenant.
-->
So perhaps in this example from other training.
-->
So in these all services management groups, you can have that hierarchy that we just created.
-->
So you can create the tenant group, which will be like the basics of the company.
-->
The organization example.
-->
I'm going just to create the example here.
-->
Organization example.
-->
So here, this is the organization example.
-->
If we want to do the example, we can do that.
-->
So this can be like the management group.
-->
Ah, where is it?
-->
Here I found the separate.
-->
So.
-->
I'm not sure what happened.
-->
Okay, so create here.
-->
Okay, let's see.
-->
Okay, let's see.
-->
So all this is the tenant work group.
-->
And under that is where you create your management group or your organization.
-->
You can add this description here, or you can create the instructor.
-->
So there was the apparel.
-->
And there was the sports.
-->
So let's create a little bit like this apparel.
-->
And here, so I can create the marketing.
-->
And sales.
-->
So that is what we are doing here.
-->
So with this organizational unit or with this management group, what you should do is.
-->
Okay, I'm trying to copy what the group one designed.
-->
So here I will add the subscription.
-->
So I'm not going to do it.
-->
I will just have one of those subscription and move those subscription to the management group.
-->
And I can even create a new subscription.
-->
So here is where I create the subscription or assign the subscription to this management group.
-->
And with this unit or within this unit is how I can give access to the different.
-->
Actually, I'm going to keep using this desktop so we don't have to change the screen as well.
-->
So that way is how I do implement this kind of stuff.
-->
So let me erase it.
-->
And all the policies and all the stuff goes in there.
-->
So yeah.
-->
Okay, so that is what we were supposed to create.
-->
So in this case, this will be the management groups.
-->
This will be the subscriptions.
-->
This will be the resource group.
-->
This will be the policies, the groupings and the tax.
-->
So these are like the objects that we use in order to achieve that goal that we have in the case study.
-->
The next lab, the lab 16, was related to these tagging and policy stuff.
-->
So in this case, what we wanted to do here is to, okay,
-->
so how are we going to ensure that the costs related with the project in particular are captured?
-->
And the virtual machines, which naming should be used?
-->
Or how rules are we going to use in order to have this manageable?
-->
Like being able to manage to track the cost for any new project,
-->
to be sure of the compliance of sizing and menus,
-->
and how are you going to discuss this?
-->
Please notice what is suggested in the case study.
-->
So that they want to, once you have this,
-->
you define like a workshop in order to work with the case study as well.
-->
So basically here, what we are referring to is this naming convention.
-->
Either using this template that helps you to define a policy
-->
for your applications, projects, business units, cost centers, projects, everything,
-->
to for tagging, for cash, and your naming.
-->
How are you supposed to name the sources?
-->
So the objectives of that second part of the lab are fulfilled.
-->
And here is where the final names that are created.
-->
So this template should be at least one of the deliverables
-->
of the design with a proper adoption of the framework that we are working on.
-->
And the other components here that is really valuable is this, the tool,
-->
that allows you to generate these names.
-->
So with this tool, remember that you can set up how this is going to work
-->
if I have units or not, if I want to use projects and services and functions.
-->
So you define how do you want to create the structure of your name
-->
and your environment and everything.
-->
So I really recommend you guys to take the time to play with this
-->
and see what you can produce because the amount of options is really big.
-->
In the example from yesterday, it was a name that we generated
-->
for great technologies in Linux.
-->
And we used everything that was available in here,
-->
but perhaps it's not really necessary for the use case or for the company.
-->
So this is something that needs to level the business with technology
-->
and see what the needs of cost accounting and all that stuff,
-->
how are we going to work with that.
-->
This is what you need to take a look into this.
-->
Okay, let me know if you have any question in the slide number 208.
-->
You will find the links to the different documentation related to this chapter.
-->
So you can get a little bit more in depth with the options as well.
-->
Really quick, this other lab will help you to list all the role-based
-->
access control available for a resource group.
-->
So you can see what is in there.
-->
So here you can see the specific resource group.
-->
You can see the access controls that are available.
-->
So you have access if you want to check access to other users.
-->
So this is that.
-->
Just to clarify a little bit, this training is not supposed to see
-->
how to operate, how to do this kind of stuff,
-->
but to understand what the rationale behind this.
-->
So then once this is designed, you can operate or implement this kind of stuff.
-->
Just for you to have the idea.
-->
Same with the policies.
-->
So the policies are here.
-->
What we will define the policy here is a policy here.
-->
What are you?
-->
The policies are already here.
-->
Most of them are already there.
-->
Remember what I showed you?
-->
That there were some blueprints that if you applied those blueprints,
-->
your cloud configuration will be ready to be certified for ISO 27,000,
-->
for example, the most global one.
-->
So here are the certifications.
-->
What you see here is that there are 26 certifications.
-->
Okay.
-->
So we need to discuss this a little bit.
-->
So basically those are policies that are already in place.
-->
So obviously the best option of practice is to start in the very, very basic.
-->
Otherwise, you may have a lot of issues because you are not allowed to do certain stuff
-->
that you need to do and because of the policy,
-->
you need to change the definition of something else.
-->
So it's important just to take it slow as well.
-->
And okay.
-->
So let's go.
-->
Let's move on to this.
-->
Okay.
-->
So now what we're going to discuss is the signing of app service or web app.
-->
So in order to do that, we do have this decision tree, which is super useful.
-->
And this is something that at the end you want to have like not memorized,
-->
but you need to have it as clear as possible in your mind.
-->
Because this decision tree allows you to choose which kind of resource you need to use
-->
for your workload to move your workload from on premises into the cloud.
-->
So which would be like the best solution.
-->
Okay.
-->
And remember that we discussed that when you are moving into the cloud,
-->
when you are adopting the cloud, perhaps from the five hours from Gartner,
-->
the most common one is just to use the virtual machines and shift and go there.
-->
But that way is like this challenging hosting for guidance.
-->
And you are losing all the benefits of having the cloud.
-->
So this decision tree helps me to realize the value of moving into the cloud.
-->
So here, let's do the process.
-->
So this is page 211 for the ones that are following the slides.
-->
So the start.
-->
So if you are going to migrate, remember, migrate or innovate, right?
-->
So in this case, migrate or build a new instance.
-->
So if you want full control and this refers to the operating system,
-->
plus the applications on the middle work that you put into this virtual machine.
-->
So this is a way to do that.
-->
If you don't do that, go ahead, Kempi.
-->
Do you have any questions?
-->
If you don't require full control.
-->
So it doesn't matter if it's Windows or Linux or whatever.
-->
So it gives you a different set of questions.
-->
For example, if it's a high performance computing workload,
-->
if it's a microservice architecture, if it's event-driven workload,
-->
if it requires full orchestration,
-->
if it requires integration with .NET or Spark on Windows.
-->
So all of these options will provide you a specific actual resource or service.
-->
So you can run your application.
-->
And that is the most optimal way.
-->
So if you invest, that is in the case.
-->
Go on, go ahead.
-->
OK.
-->
If you are migrating, the questions are already there,
-->
but the outcome is kind of the same.
-->
So are you going to migrate?
-->
So are you going to lift and shift?
-->
So probably the virtual machine and make it run on the cloud?
-->
Or you are going to optimize for the cloud?
-->
So if you are going to optimize for the cloud,
-->
please apply the same set of questions applied.
-->
Is it high performance?
-->
Is it microservice?
-->
Is it event-driven?
-->
Is it requires orchestration?
-->
If it's .NET integration or if not, the Kubernetes service is happy.
-->
If you're just going to lift and shift,
-->
the first question will be can this be contained?
-->
So containerized.
-->
So it turns out that it is surprising how many monolithic applications can be containerized.
-->
I mean, PHP application, Java application,
-->
even .NET applications can be containerized.
-->
So that means put those applications in another container.
-->
So it's something that is not that hard to achieve with a lot of applications.
-->
So this is something that is worthy to analyze and see if it's possible or not.
-->
So if it can be done, the question here will be do you need full orchestration?
-->
If not, go here.
-->
If you do need orchestration, you will have two options.
-->
Use the .NET service or use the app service again.
-->
So it depends on what you want here.
-->
If the application cannot be containerized,
-->
you can see the app service depending on if the language is one supported by the service or if not, the automation.
-->
So the point here is that the only way in which you may end up with a virtual machine are kind of the same cases.
-->
So there are two options for virtual machines.
-->
But after discarding everything else, almost, right?
-->
But in the practice in real life, this usually is the first option.
-->
So it's the most common option.
-->
But it doesn't really need to be like that.
-->
I mean, a lot of applications can be run as app services.
-->
And a lot of applications can be run as containers, either in Kubernetes or as container instances.
-->
Although this service was appreciated.
-->
There are some changes in this service.
-->
The common is more the Kubernetes one and the app service one.
-->
But the point is that virtual machine is not like the first option.
-->
But in the decision tree.
-->
But in real life, it is probably popular as well.
-->
So with that discussion in planes, let's go.
-->
Let's see a few more stuff before going to our next phase study.
-->
So here there is another decision to help you choose when to use an app service.
-->
So the app service is mostly web applications.
-->
But it can be also be used for, for example, mobile backends or REST APIs or jobs that we run from time to time.
-->
It supports different languages.
-->
This is this kind of thing.
-->
This can be connected to your CI-CD pipeline.
-->
Either a gift card, actions or edge drops, whatever.
-->
So you can use this to deploy in there.
-->
There are some considerations that you need to take into account.
-->
The most important one is which is the plan that you are going to use.
-->
Because that is the cost.
-->
That is what defines the cost and the scaling capabilities.
-->
So let me show you here.
-->
For example, this is the report sheet.
-->
So this is a resource group with app services.
-->
And here we have what is called the app service plan.
-->
So this app service plan is basically you can call it like kind of the subscription of your web applications.
-->
So this app service plan and let's see where is it.
-->
This service plan is in the free time.
-->
So it doesn't cost at all and gives you small instances.
-->
And in this case it only has one instance.
-->
But it can be more instances if you require them.
-->
So this is an instance of that plan.
-->
So an instance is basically an application that can be run in a specific container.
-->
So let's give it a minute for this to load.
-->
So this app service plan is like the hosting plan or the former hosting plan that we have in our distribution data center.
-->
So let's wait for it to load.
-->
So in this app service, you have like all the options that can be configured, not only the deployment.
-->
So this is linked to CICD, to your partner, the configuration.
-->
So you have cast backups, domains, certificates.
-->
You can scale up or scale out your system.
-->
And it was started and going to stop it.
-->
So it has a domain here.
-->
It is a PHP application.
-->
This has the IP address.
-->
It's a Linux server.
-->
If you want to scale this up so you can have like more power of the system.
-->
So this is, I mean, this is instead of having a server with Apache server with the PHP page,
-->
you just deploy this app and you have the same.
-->
So this is the features of this kind of workloads.
-->
So the ones for development and testing that, for example,
-->
the free one gives you a maximum of 60 minutes per day for free of service of this application,
-->
which means that it can only be like one hour use during the day for a maximum of one hour.
-->
And the productions ones with the ACU, which is like the capacity per unit.
-->
So they have the data.
-->
So this is for production workloads.
-->
But you can see that we are taking care about one CPU or four CPU.
-->
In this extreme case, 32 gigabytes, 250 gigs, the SOA that you have available here.
-->
And the cost of these things is super cheap per month.
-->
So the thing is that these can have a lot of websites or backend services available around this environment.
-->
So this or that is perhaps one of the most important things that you need to define when you are choosing your web service app.
-->
This is an example of an architecture using one of these.
-->
So you have the directory, the DNS for resolution, and you have the app service plan with the web app and different environments.
-->
This is the testing one.
-->
And the last good production or last good operated environment database backend.
-->
So your application connects to this backend.
-->
And this is monitoring.
-->
So if you have monitoring enabled, if you pay close attention to what I was telling you,
-->
you will have noticed that this doesn't have application enabled in this machine.
-->
Because obviously that costs and this is just like a demo stuff.
-->
So there is no need to have it that way.
-->
So that is like the web application.
-->
We have a few labs here to deploy web application using this guide, the .NET store guide.
-->
But there is this other one that is also super nice because there are multi, I mean, this one allows you to deploy a .NET web application
-->
either from business to the code or directly from the portal.
-->
So I don't know why it is loading there.
-->
So it's just like a sample application to deploy.
-->
Or there is this other lab that deploys a more bigger application here in DevOps.
-->
So let's give it a try.
-->
Let's have like a few minutes in here.
-->
So we choose App and Play.
-->
There are some applications here.
-->
I have suggested any of these.
-->
My shuttle, how many parts I needed, the smart portal.
-->
So what we're going to do is let's go again to our breakout rooms.
-->
And as we have multiple applications, what we're going to do is that the group number one will deploy my shuttle.
-->
The group number two will deploy my health clinic.
-->
The group number three will deploy parts unlimited.
-->
And the last group will deploy the smart portal.
-->
It may happen that the application doesn't deploy because there was a change or something like that.
-->
I mean, this is changed a lot.
-->
So it may happen sometimes that it fails or something.
-->
But do not worry about that because what we want to see is what this creates, what this generates.
-->
So we can have a better understanding.
-->
So I'm going to send you guys to the groups.
-->
Although there is something that's here.
-->
I mean, yesterday I see that Alexander was working with Charmaine.
-->
And Charmaine is going to, I mean, she's a rock star because she has attended like more courses, trainings that you can imagine.
-->
So let me, I'm going, Alexander, I'm going to move you to a different group so you don't be alone.
-->
No, it's okay.
-->
Okay, cool.
-->
No worries.
-->
So let's open the groups and go from there.
-->
This is page 217.
-->
We can work like for 10 minutes.
-->
And then we will get back here as well.
-->
So, yeah, go on.
-->
Yeah, as soon as you get in the room, we will show you this room number.
-->
Okay.
-->
So let me see.
-->
I think we have 40 seconds here.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
Okay.
-->
I was having issues with my things, but now I'm back.
-->
So how are you doing?
-->
Are you deploying already something or for which one?
-->
Awesome.
-->
Okay.
-->
Cool.
-->
Cool.
-->
Cool.
-->
And take a look.
-->
This is the board, but you also can take a look into the repos and pipelines.
-->
And actually, that is the application.
-->
And take a look into the pipeline.
-->
So this has the proper application created.
-->
So you can run it, run the pipeline.
-->
Okay.
-->
Choose the...
-->
Start with, for example, the Docker group.
-->
Choose that one.
-->
And run pipeline.
-->
This pipeline creates a...
-->
Or should create a Docker image.
-->
A Docker image for the application.
-->
So...
-->
And then go to releases as well in the left column.
-->
And you should see here the release of a container.
-->
So as this project is fully set up, everything is in place.
-->
From the board, to the backlog and everything, to the source code,
-->
to the pipelines, the CI pipeline that creates the images from the binaries,
-->
and the CD pipeline that deploys the application into the work.
-->
So this is super beautiful because it has the whole package.
-->
All the idea that you hear about, okay, let's do DevOps, let's do automation,
-->
or let's move it to the cloud, blah, blah, blah.
-->
So that is in here.
-->
So with the pipeline, the first one that you'll watch, go back there.
-->
See if you've already finished it or...
-->
It may take a while.
-->
It's still running.
-->
Ah, it got cancelled.
-->
Let me see.
-->
It may fail because perhaps it's too old to be honest.
-->
Yeah, it wasn't able to...
-->
I mean, it's root testing for 16.04, and that doesn't exist anymore.
-->
But that pipeline creates the binary, and that, in this case, will be an image,
-->
a vector image, and the release will take that image and deploy that into,
-->
in this case, some app service.
-->
So click on Create Release.
-->
It will fail as well because it doesn't have an image.
-->
Okay, it's using the...
-->
If you want to go to pipelines, and let's see the other one.
-->
There was...
-->
I haven't tried to run that one yet.
-->
Okay, no, there is no pipeline in there.
-->
Let's see if it shows up.
-->
Let's see if it shows up.
-->
Yeah, these projects may be outdated.
-->
They have a lot of stuff open up there.
-->
Let's try with another of the pipelines I learned.
-->
I should try with the Docker, and then I should probably build to see if it works.
-->
Okay, this is more recent.
-->
We can try that.
-->
Try to run this one to see if it runs.
-->
And perhaps you can change the Ubuntu 18 to a more recent one.
-->
Let's see if it works.
-->
Try to Ubuntu.
-->
We want to make this open.
-->
These projects can be adjusted so they work,
-->
but it will take too much time.
-->
It's just to have an idea.
-->
Let's see if it works.
-->
Okay, let's give it a try to figure out what was done in there.
-->
And if you want to, you can take a look into the other deployment of our web app to see how it works.
-->
It seems it's running.
-->
Click on phase one and you can see the level.
-->
Okay, it seems it's running.
-->
As soon as this finishes, go to releases and find the corresponding one.
-->
Okay, it plays the maiden.
-->
Click on the maiden.
-->
This is the life of a DevOps engineer.
-->
I mean, what you're seeing here is what a DevOps engineer does all the time.
-->
Because the pipeline is failing, so what you will have to do as a DevOps engineer is to...
-->
Okay, let's see why it's failing.
-->
It was because, I don't know, the version or the library or whatever.
-->
So we need to adjust and do whatever.
-->
Or there is a need of variable that is not set or something like that.
-->
So that is like what you do as a DevOps engineer.
-->
Try to use a different image.
-->
The 2004, perhaps because, I mean, probably as this project is old,
-->
the versions of the libraries in the new one doesn't work as expected.
-->
You can give it a try and see if it works.
-->
And you can also try to deploy the other, that net application as an app service to see how it works as well.
-->
Okay, so see you in a while.
-->
Okay, guys, how are you doing?
-->
Let me show the error.
-->
It seems that it's asking you for...
-->
Okay, it's asking for the name of the project, the organization.
-->
Which organization are you using?
-->
Okay, that's the application.
-->
So you need to give a name for that application.
-->
So name that and dash your name so it can be unique or perfect.
-->
Group two, you are two.
-->
Uh-huh, uh-huh.
-->
No, no, you cannot use spaces.
-->
And, uh-huh, perfect.
-->
It may fail to compile because these projects are kind of old,
-->
so it not necessarily will compile completely.
-->
But you can see the full structure of how the DevOps is implemented and all the components
-->
and hopefully the web application will be deployed at the end.
-->
So just explore.
-->
I'm going to be back here in a while.
-->
Let me go to the other groups to see how they are doing.
-->
Hello, guys, how are you doing?
-->
Oh, very good.
-->
We're able to deploy.
-->
Yeah, because I was telling you, awesome, awesome.
-->
That is great news.
-->
Because the thing is that these projects may be kind of old
-->
and not necessarily will be deployed or will have some failures while compiling.
-->
But in the basic sense, it works.
-->
So super cool, nice.
-->
So I'm going to use you as the demo later.
-->
Okay, keep going there.
-->
So see you in a while.
-->
Hello, guys, how are you doing?
-->
Okay, it's creating.
-->
Oh, you are creating them, I hope, as well.
-->
Okay, cool.
-->
Let's see how it goes.
-->
So not necessarily it will complete because these projects are kind of old.
-->
So perhaps it doesn't deploy or is able to deploy the full application.
-->
But you just can see all the components that are in there.
-->
Let me see the pipelines.
-->
This one, which one is this?
-->
Okay, let me see the pipelines.
-->
Okay, there are no pipelines.
-->
And it releases.
-->
Okay, very sorry.
-->
Let's try to launch a release.
-->
Yeah, create a release.
-->
I think why it's going like that.
-->
Click on deploy.
-->
When you click on the My Health Clinic E2E in the middle column where it shows,
-->
in the middle column, it says My Health Clinic E2E.
-->
Click there.
-->
It doesn't show anything.
-->
Here, here, here, where we are, here.
-->
Right button, what does it show, the right button?
-->
The right button, the right button there.
-->
It doesn't show like a menu, no.
-->
Click on the blue button that says create release and see if it works.
-->
The blue button on the data base.
-->
Yeah, it shows that.
-->
It's not showing you any action.
-->
Is it testing?
-->
Yes, it's testing.
-->
Ah, okay, the artifacts deployed, yeah.
-->
Okay, there are no.
-->
Okay, yeah, because what I see is that the pipeline is not there.
-->
Go back to pipelines.
-->
Okay, it wasn't created.
-->
Is the demo generator already finished?
-->
Ah, okay, okay, okay.
-->
We can see.
-->
Ah, exactly.
-->
The pipeline is what it was failed.
-->
Ah, that is true.
-->
Let's, okay, let's go.
-->
Yeah, the thing is that not all of these demos work.
-->
But one of the teams were able to play the demo, so we are going to see.
-->
Give it a try to see if it works.
-->
And I'm going back in a few seconds.
-->
Or I'll call you back to the name in a moment so we can check.
-->
Okay.
-->
Okay, guys.
-->
So, well, some of the projects didn't deploy.
-->
However, the team from Alexander and Blair, they were, they advanced a lot.
-->
So, I mean, it was able to deploy the web absolutely.
-->
So, if you can show your screen so we can take a look, all of us.
-->
And also, I'll show you mine to see.
-->
Okay.
-->
Yeah, please.
-->
Yeah, well, kind of.
-->
Let's see if it works.
-->
Okay.
-->
Okay.
-->
Awesome.
-->
Okay.
-->
No worries.
-->
Let me show this.
-->
So, let's go to the DevOps so we can see the project.
-->
Oh, no, this is the portal.
-->
So, go to the DevOps, which is the next tab that we have.
-->
No, there.
-->
Okay.
-->
So, let's just.
-->
Okay.
-->
Okay.
-->
Oh, yeah.
-->
Okay.
-->
So, let's see the work first.
-->
So, this lemon generator that this creates is a full project with the backlog and the
-->
five, the repos from the source code and the pipelines and all that stuff.
-->
So, this is just an example, but it gives you, like, a general idea of the whole
-->
process as well.
-->
So, okay.
-->
This is the backlog.
-->
Let us show us, please, the repos or the source code, please.
-->
Uh-huh.
-->
So, this is the repo for the application.
-->
In this case, it's an ASP net application, a web net application.
-->
And something interesting here is that file that says Azure pipelines, but it has
-->
a lot of branches.
-->
No.
-->
Before going there, look in the middle column.
-->
There is a file that says Azure dash pipelines one, oh.
-->
In the middle column, in the column of the middle.
-->
Uh-huh.
-->
At the end.
-->
In the middle.
-->
The column of the middle.
-->
The middle column.
-->
The last.
-->
The last.
-->
No.
-->
The other.
-->
To the other side.
-->
In the middle.
-->
There are three panels.
-->
So, the panel in the middle.
-->
And the one.
-->
No, that is the third panel from the second one.
-->
And the last one.
-->
In the same.
-->
In the other panel.
-->
Uh-huh.
-->
Last file.
-->
Last file.
-->
So, this file is the general file of the pipeline.
-->
So, this is like the instruction set to deploy this application.
-->
So, if we go, please go down.
-->
Down, down, down.
-->
A little bit more down.
-->
Here, I think.
-->
Line 81.
-->
So, the line 81 is the deployment task of the Azure app service.
-->
So, this deployment task, what this creates is the app service that we were discussing a few months ago.
-->
So, the subscription.
-->
The name of the web app.
-->
Where it's going to be deployed.
-->
The resource group.
-->
The package that is going to be deployed.
-->
So, that is, that goes about ten lines.
-->
Are the key of the deployment into the web app.
-->
Before that is all that process, no?
-->
So, the scope of what we are doing here.
-->
So, please have in mind how big, how huge this is.
-->
Because the only scope of what we are discussing is which are the trade-offs.
-->
Which are the design decisions in order to decide what to put in those ten lines.
-->
So, all this discussion of the CAF and the WAF and app service and what comes next.
-->
Is just to decide what to put into those ten lines.
-->
So, this is huge.
-->
Yeah?
-->
So, this is huge.
-->
I mean, this is a file with, I don't know, one hundred something lines.
-->
Two hundred something lines.
-->
And we are discussing with those that would protect.
-->
Because before that is the developers part and the infrastructure guys and all that stuff.
-->
Now, you can go to the blue rocket in the first panel.
-->
The pipelines.
-->
No, that one.
-->
Okay.
-->
This pipeline is, click on there please.
-->
This is the pipeline that we just saw.
-->
We can run it here.
-->
Don't do it because as the deployment runs by itself, we may break that.
-->
It's already broken, but I want to show you what it is.
-->
So, when we run this pipeline, what we are doing is executing that Jamf file.
-->
And this would create the app service and deploy the application into the app service.
-->
The application is just a zip file.
-->
For example, imagine a PHP application, which perhaps is simple.
-->
So, the PHP at the end, it's a folder with a series of PHP files and a configuration that points towards a device.
-->
So, that is going or that goes into a zip file.
-->
And that zip file gets unzipped into this container or these app services.
-->
And that is how it's served to the public.
-->
This is like the general idea.
-->
So, that is what this does.
-->
And go to releases.
-->
You just need pipelines where you are.
-->
In the same rocket.
-->
And releases.
-->
So, these releases is when the pipeline creates the zip file or the binary or the finish.
-->
This is sent to the app service.
-->
So, the release part is the one in charge of that.
-->
Probably in the test plans, in the purple icon, there are some test plans.
-->
Yeah, okay.
-->
This is a test plan.
-->
So, it checks the application.
-->
For example, here you can check the application.
-->
It has errors in the code.
-->
Stuff like that.
-->
And you can automate that.
-->
And the artifacts.
-->
And the artifacts in the last icon.
-->
We found a test plan.
-->
Artifacts.
-->
Probably you will see the zip file for this one.
-->
Yeah, okay.
-->
Okay.
-->
It doesn't show anything.
-->
Well, it should be there somewhere.
-->
I have to see another.
-->
But no worries.
-->
Don't waste too much time here.
-->
Let's go now to the portal.
-->
To the portal.
-->
To the top of the portal.
-->
No, the portal.
-->
Uh-huh.
-->
So, here, can you go to Chrome, please?
-->
So, and resource groups.
-->
Can you type resource groups?
-->
No.
-->
Uh-huh.
-->
And you should have a resource group for this application that you created.
-->
Which is, okay, that is the name.
-->
So, here is where we have the objects.
-->
So, we have the plan.
-->
Which is the hosting plan.
-->
And it defines how are we going to handle the loads.
-->
So, can you click there in the plan, please?
-->
No, in the middle.
-->
In the last panel.
-->
Where it says app service plan.
-->
In the two objects.
-->
Uh-huh.
-->
Down there.
-->
Down there.
-->
No.
-->
In the two objects.
-->
In the middle.
-->
The biggest.
-->
The big panel.
-->
Uh-huh.
-->
The big panel.
-->
Uh-huh.
-->
Uh-huh.
-->
And up it goes.
-->
So, the second one.
-->
The app service plan.
-->
Go back one time to the network.
-->
And the plan.
-->
Where it says.
-->
Uh-huh.
-->
That is the plan.
-->
So, this is the plan where you can have multiple deployments here.
-->
You can say that this one is like the machine where your application is going to run.
-->
So, this shows you that.
-->
And in the left column, you will see apps.
-->
The icon apps.
-->
So, here are the apps that are running into your plan.
-->
Into your machine.
-->
So, now you can click there.
-->
And this is the actual deployment that was made.
-->
So, it shows you the URL.
-->
This is how we use the machine.
-->
This is the runtime stack.net.
-->
And the.
-->
So, all the information.
-->
The deployment center.
-->
The line.
-->
Where.
-->
So, if you go to the default domain.
-->
And open that.
-->
Which is the page that we showed you.
-->
Show us the very beginning.
-->
Oh, no.
-->
In that page, it says default domain.
-->
There.
-->
There.
-->
Default domain.
-->
Uh-huh.
-->
Open that in the top.
-->
Uh-huh.
-->
So, that is the application.
-->
And it is not loading.
-->
But.
-->
I think that you have one.
-->
Uh-huh.
-->
Show us.
-->
Show us the one that.
-->
Uh-huh.
-->
There you go.
-->
Okay.
-->
Yeah.
-->
It's the same.
-->
So, here it's failing.
-->
And it shows that the error is because it seems that it's not finding the web
-->
config file, I think.
-->
Uh-huh.
-->
But.
-->
It doesn't matter right now.
-->
So, you will like the idea.
-->
So, what we wanted to see here is an example of how to.
-->
Of how to.
-->
Well, in this case, to see all the debug process.
-->
Which is outside of the scope.
-->
But it's good to have it in mind.
-->
So, we can know where we are standing up.
-->
And see finally that web app service.
-->
That we were talking about as well.
-->
Okay.
-->
I tried to deploy the smart hotel application.
-->
That is failed as well.
-->
Let me see.
-->
So, I don't know if any of you guys were able to launch the application.
-->
Or if any of you launched the app service directly.
-->
To the other left.
-->
Any of you did that.
-->
It's okay if you don't.
-->
Okay.
-->
So, maybe.
-->
So, that is what the app service is about.
-->
Okay.
-->
Okay.
-->
So, these two components.
-->
Well, this box.
-->
This component is the one that we were deploying.
-->
The app service plan.
-->
The app service web app.
-->
These two slots are the ones from the pipelines.
-->
From the DevOps environment.
-->
So, you can have multiple deployments.
-->
Depending on the environment.
-->
Or the stage of development of the application.
-->
So, this is this component of the whole solution as well.
-->
And the criteria that is needed here.
-->
In order to work with this kind of applications.
-->
This one.
-->
What kind of service.
-->
If you are going to use some kind of deployment.
-->
If you are going to have apps.
-->
Mobile apps.
-->
Using this.
-->
Web app as a backend.
-->
You can have authentication.
-->
You can have REST APIs here.
-->
You can use these to run jobs.
-->
So, this is like a really flexible way.
-->
And it's not that expensive.
-->
Compared to having a.
-->
A what?
-->
A full built machine running that.
-->
Because in this case.
-->
You have.
-->
If you take notice.
-->
Almost all of these were Linux.
-->
But the first limited was with a Windows application.
-->
So, at the end.
-->
You want nine.
-->
Because the service is able to understand.
-->
Based on the code.
-->
Which kind of image needs to be assigned.
-->
In order to deploy this application.
-->
And make it work.
-->
Okay, so we're going to do a small coffee break.
-->
And we are going back.
-->
In order to discuss all the storage items.
-->
And the designing decision.
-->
Of storage solutions for Azure.
-->
So, we will discuss this.
-->
The storage counts.
-->
The redundancy.
-->
The block storage.
-->
The files.
-->
The distribution.
-->
The security.
-->
And we will have another case study.
-->
This case study.
-->
This one.
-->
This case study.
-->
The idea is to.
-->
For the same company.
-->
That we are working on.
-->
Which kind of storage solution are we going to create.
-->
Or to assign for them.
-->
Provided that they have, for example.
-->
Media files.
-->
With photos and videos.
-->
Marketing literature.
-->
Which are mostly PDF files.
-->
With suppliers and stuff.
-->
And corporate documents.
-->
Which, and this is a touch.
-->
Have a difference with sensitive information.
-->
In common office formats.
-->
So, that is the next case study.
-->
That we are going to work with.
-->
But first, or before that.
-->
We are going to discuss all these components.
-->
So we have an idea.
-->
So, let me know.
-->
If you have any questions.
-->
And if you are good to go.
-->
Let's have page 15 minutes.
-->
So.
-->
So we can ask.
-->
That will be.
-->
11 AM with time.
-->
Okay.
-->
Okay.
-->
So let's see page 15.
-->
Right.
-->
Okay.
-->
Okay.
-->
Okay.
-->
So let's continue then.
-->
Discussing the storage components.
-->
And let me know if you have doubts.
-->
Or whatever.
-->
Regarding what we have seen previously.
-->
We have discussed it previously.
-->
Otherwise, let's just.
-->
This is just regarding these storage components.
-->
Or services available.
-->
Okay.
-->
So, first we need to.
-->
But there are two kinds.
-->
Kinds of data.
-->
The first one is non-relational data.
-->
Which refers to files.
-->
And the second one is non-relational data.
-->
Which is either media files.
-->
Or documents.
-->
Or PDF files.
-->
And relational data.
-->
Which is basically.
-->
So.
-->
So.
-->
Components.
-->
The first thing that needs to be done.
-->
Is to classify.
-->
Which kind of data.
-->
Are we going to store.
-->
In the service.
-->
For example.
-->
If it's on a structure.
-->
Which is a transportation.
-->
If it's a security structure.
-->
Which could be with SQL.
-->
Or if it's somewhere in the middle.
-->
In the case with semi-structure.
-->
So.
-->
The first one of the.
-->
Of the components.
-->
Is the.
-->
Storage.
-->
The storage accounts.
-->
These storage accounts.
-->
Are the.
-->
Are the basics.
-->
Of how to handle.
-->
Non-structural data.
-->
These storage accounts.
-->
Allow you to.
-->
Have or to define.
-->
A series of.
-->
Pictures.
-->
That are related.
-->
Of the way you handle the data.
-->
These features are related.
-->
For example.
-->
Where do you want to locate your data?
-->
And which.
-->
Are you going to use.
-->
To locate some of the data.
-->
If you are going to replicate.
-->
The data.
-->
In order to have.
-->
A redundancy.
-->
If you are going to provide.
-->
Disaster recovery options.
-->
For storage.
-->
So that is.
-->
Related to location.
-->
Regarding compliance.
-->
If you are going to have.
-->
Storage.
-->
With sensitive data.
-->
Of storage.
-->
You can mention that.
-->
Because of some legal.
-->
Requirements.
-->
To be stored.
-->
For 5 years.
-->
Or 20 years.
-->
So that is.
-->
The use limit of storage.
-->
So that doesn't change.
-->
Under the period.
-->
Is the cost.
-->
If you want to have.
-->
Code access.
-->
So it's storage of files.
-->
That are permanently being used.
-->
Or it's just archived storage.
-->
That is stored.
-->
For the.
-->
Or eventually use that.
-->
Or.
-->
Code archive.
-->
That is not.
-->
Frequently accessed.
-->
Accessed for information.
-->
Also.
-->
If you want to have.
-->
Different kinds of replication.
-->
Either local redundancy.
-->
Or some redundancy.
-->
Or geolocation redundancy.
-->
Or real time.
-->
Access redundancy.
-->
So there is different replication.
-->
For this.
-->
Regarding the administrative.
-->
Overheat.
-->
How are you going to file this data?
-->
If there is like.
-->
After a certain while.
-->
The data in this.
-->
Container condition.
-->
In this account is going to be moved.
-->
To code archive.
-->
Or is going to be.
-->
Raised.
-->
It needs to be indexed.
-->
It needs to be available for.
-->
Searching.
-->
Et cetera.
-->
So how it's going to be handled that.
-->
How sensitive.
-->
Is the data.
-->
So if you want to have versioning.
-->
If you want to have.
-->
If you want to have.
-->
In time.
-->
Also.
-->
The handling.
-->
Of the authentication.
-->
Of authorization.
-->
If you are going to.
-->
An active directory.
-->
Or if you are going to use.
-->
A SAS key.
-->
To access the data.
-->
Which kind of service level.
-->
If you are going to have 99%.
-->
Or 99.
-->
99%.
-->
If you.
-->
Are going to have.
-->
Which kind of account.
-->
Are you going to have.
-->
Or a premium.
-->
Account for the rules.
-->
If you have.
-->
Features.
-->
Et cetera.
-->
These are different options.
-->
That are required.
-->
And that are needed.
-->
In order to design or define.
-->
Each kind of storage.
-->
Are you going to use.
-->
So first of all.
-->
Deciding.
-->
On which type of storage account.
-->
Are you going to have.
-->
So.
-->
The same.
-->
Choice with the standard.
-->
One by default.
-->
The standard.
-->
For.
-->
Services.
-->
This is important.
-->
So you have.
-->
Account time.
-->
And you have services.
-->
That are related.
-->
Or are classified.
-->
In this account time.
-->
So depending on.
-->
The level that is available.
-->
So this one is.
-->
Can be used in almost every.
-->
Other scenario.
-->
That is available.
-->
And there is a premium ones.
-->
If you have the block under the legs.
-->
The files.
-->
And the face blocks.
-->
Those issues are for.
-->
Specific use cases.
-->
So.
-->
The premium block blocks.
-->
Is for small transactions.
-->
Or high transaction rates.
-->
So something that needs to be.
-->
Store with a high.
-->
Interaction rate.
-->
Story.
-->
The actual files.
-->
In your story.
-->
And the files.
-->
In a.
-->
Traditional.
-->
Resource chart resource.
-->
In the network.
-->
Or the blocks.
-->
The face blocks which are required.
-->
For example.
-->
To provide a.
-->
Low latency.
-->
Storage and high.
-->
Performance.
-->
In particular.
-->
What else do you need to have in mind?
-->
The location.
-->
To the closest.
-->
To your customer.
-->
Or your clients.
-->
Or your business.
-->
Which regulator.
-->
Which guidelines.
-->
Or what they can actually.
-->
Fulfill the costs.
-->
Of these services.
-->
The premium services.
-->
Are more expensive.
-->
Than the other ones.
-->
If you are going to have reputation.
-->
You are going to have different.
-->
Requirements.
-->
For these.
-->
And both management.
-->
Activities.
-->
So one of the most important items.
-->
To have in your mind.
-->
Is redundancy.
-->
So the ability to replicate.
-->
People's storage.
-->
So you can have.
-->
Within a regional fashion.
-->
Redundant storage.
-->
Locally to the region.
-->
Or redundant.
-->
Among different zones.
-->
Within the same region.
-->
So this means multiple data centers.
-->
But in the same region.
-->
Or you can have.
-->
Or additional.
-->
You can have a secondary region.
-->
Which means that you can have.
-->
A coffee.
-->
Or a replica.
-->
Of the storage.
-->
In a complete.
-->
Different region.
-->
In the world.
-->
For the data center.
-->
You can have.
-->
As an example.
-->
In India.
-->
In the India region.
-->
In Australia.
-->
A replica of that data.
-->
So you can have access to both.
-->
In case of a failure.
-->
A catastrophic failure.
-->
In the India region.
-->
So that is the.
-->
Is available here.
-->
So this is like an example.
-->
Of how this can be seen.
-->
So the replication of the storage.
-->
You have the data center.
-->
With the storage account.
-->
And in the primary region.
-->
And this will be.
-->
Locally redundant storage.
-->
So multiple copies.
-->
In the same data center.
-->
When it's sewn redundant.
-->
If you have multiple data centers.
-->
Different data centers.
-->
With multiple copies.
-->
Of the data center.
-->
But this is in the same region.
-->
This is India.
-->
And this is India as well.
-->
But three different data centers.
-->
If you think.
-->
Or if you map that.
-->
To the own premises.
-->
The scenario.
-->
The traditional scenario.
-->
That is something that.
-->
Is almost.
-->
Impossible to achieve.
-->
Without love.
-->
Or a huge investment.
-->
For a common company.
-->
You need to take out.
-->
Three data centers.
-->
In different countries.
-->
That is not something.
-->
Easily done.
-->
So that is why.
-->
This is so important.
-->
But the answer is tricky.
-->
In the same way.
-->
This here.
-->
Doing that is super simple.
-->
Super easy.
-->
But it has a cost.
-->
Do I really need.
-->
To have a replica.
-->
In another argument.
-->
Is that really needed.
-->
For my requirements.
-->
Or just having the.
-->
Copy multiple data centers.
-->
Is enough.
-->
But what happens if you lose.
-->
Access to the.
-->
Internet connectivity.
-->
To that country.
-->
You need to take that.
-->
To account.
-->
With the price.
-->
In this example.
-->
This is the.
-->
Geographical replication.
-->
So this is one region.
-->
This is a different region.
-->
And here is like.
-->
The same situation.
-->
But with some.
-->
Replication works.
-->
So in different regions.
-->
Secondary region.
-->
This is like the highest.
-->
Option available.
-->
So those.
-->
Regarding the type.
-->
Of storage.
-->
So you can have storage.
-->
That is available.
-->
Like object storage.
-->
That is available.
-->
Using an API.
-->
So the lowest.
-->
Storage.
-->
Will be like the object storage.
-->
Which is like the most.
-->
Common one.
-->
The one that you use.
-->
But there is also.
-->
Available for cable storage.
-->
Which is.
-->
Kind of.
-->
Not necessarily.
-->
But it's just like a table.
-->
A matrix for storage.
-->
Is the same.
-->
Or.
-->
And key storage.
-->
Is in a segment.
-->
With publishing subscription.
-->
Mechanisms.
-->
To access the data.
-->
So these.
-->
Are programmatically.
-->
Accessed.
-->
This requires a client.
-->
To use an API.
-->
To access the data.
-->
And this other one.
-->
Is the.
-->
Like the more.
-->
Traditional way to access.
-->
The data.
-->
Which will be the file storage.
-->
Like SMD or SIPs.
-->
And the bit storage.
-->
Which is like.
-->
Which is a hard to search.
-->
So.
-->
So here is where the.
-->
Blob storage shows up.
-->
So the blob storage.
-->
The criteria.
-->
Of designing.
-->
Or the blob storage is a file.
-->
So if it's streaming.
-->
Which is the.
-->
This is for.
-->
For performance.
-->
It needs to stand down.
-->
In the sense that.
-->
It is used frequently.
-->
Or if it's cool.
-->
Or if it's hard type.
-->
When it's used by.
-->
Backup.
-->
Or long time startup.
-->
So these are like different.
-->
Fires.
-->
And the ones that we use.
-->
Usually are.
-->
At the startup.
-->
Depending on your use case.
-->
Also.
-->
You can have policies here.
-->
To enable storage.
-->
So this can be.
-->
Only read it.
-->
But not read it.
-->
So you need to store.
-->
Documents.
-->
Because of a layout requirement.
-->
Or a policy requirement.
-->
So.
-->
It can be just write.
-->
Write once and read multiple times.
-->
So here are two examples.
-->
On how to use.
-->
The blob storage.
-->
So you have a built-in thing.
-->
And have a tab available.
-->
In a blob storage.
-->
So this will be like any.
-->
Store in there.
-->
Or in this case.
-->
This will be kind of the same.
-->
But using a separate file.
-->
But it is like same.
-->
How to use.
-->
The fire.
-->
So basically.
-->
Having workload.
-->
Performance requirement.
-->
Into those requirements.
-->
And the rest can go to.
-->
Depending on the.
-->
What is needed.
-->
Because of all.
-->
Of the standard.
-->
Now.
-->
For actual files.
-->
Here is the discussion.
-->
Here is the discussion.
-->
So when.
-->
You should use actual files.
-->
Versus.
-->
Blob storage.
-->
So depending.
-->
On the kind of.
-->
That you are requiring.
-->
Depending.
-->
On the file sharing.
-->
Scenarios.
-->
That you are thinking about.
-->
Is when you use actual files.
-->
So actual files.
-->
Is the one.
-->
That gives you access to.
-->
Features like performance.
-->
So.
-->
300 megabits per second.
-->
Or up to.
-->
5 gigabits per second.
-->
Of performance and access in these files.
-->
Handling of identity.
-->
So these files are still.
-->
The way to access these.
-->
Either.
-->
Like a standard.
-->
Share with the system.
-->
With the machine.
-->
Or by synchronization.
-->
Of the files.
-->
Like the service running and synchronizing the file.
-->
With different redundancies.
-->
And different files.
-->
Exactly the same as the one.
-->
That we discussed previously.
-->
So they use cases.
-->
For Azure files.
-->
Or versus.
-->
So.
-->
When you are replacing the files.
-->
So this is like the most traditional one.
-->
Because here we are talking about.
-->
Not structured data.
-->
So this is what we have here.
-->
And.
-->
And it's accessible.
-->
By traditional means.
-->
Usually SMB.
-->
Or NFS.
-->
Or web app.
-->
And it has a little.
-->
Interrupted.
-->
Per second.
-->
And the log is more like.
-->
Perhaps.
-->
Media.
-->
That is like a content.
-->
Depending on the network.
-->
Or analytical data.
-->
So it can be used.
-->
With NFS.
-->
Or with database.
-->
So this is more like a toolkit.
-->
But it's more like an application.
-->
Or systems data.
-->
The structure.
-->
Of these types of files.
-->
So you always have your storage account.
-->
The service Azure files.
-->
And the shared resources.
-->
The corresponding
-->
directories of those.
-->
Shared resources.
-->
And the specific files.
-->
That are available.
-->
Regarding performance.
-->
So you have.
-->
The latency.
-->
Depending on the tire.
-->
If it's standard.
-->
But you have to.
-->
Take into consideration that.
-->
This mostly depends.
-->
On you.
-->
Internet connectivity.
-->
Or how are you connected.
-->
To the Azure cloud.
-->
So no matter.
-->
If you're using standard.
-->
If your internet connectivity.
-->
Is not up to the task.
-->
You won't notice the difference.
-->
Because it will probably be.
-->
Slow.
-->
So this is something that.
-->
The limitation is not properly in the cloud.
-->
But in view.
-->
This is an
-->
architectural.
-->
Describing how to.
-->
Have for example.
-->
A file synchronization agent.
-->
With on-premises
-->
data center.
-->
And Azure cloud.
-->
Services.
-->
So the file here.
-->
Is accessed by.
-->
Virtual objects in the cloud.
-->
Or servers.
-->
But from on-premises.
-->
Information is accessed using.
-->
File synchronization services.
-->
So this is like.
-->
Like one of the.
-->
Options that are available here.
-->
You.
-->
Need to think about this.
-->
Because.
-->
These synchronization services.
-->
May have the issue of.
-->
When multiple users.
-->
Are accessing the same file.
-->
It may be challenging.
-->
Sometimes.
-->
So it needs to be.
-->
A big analysis of the use case.
-->
And see if it is.
-->
The right solution or what.
-->
You have.
-->
There are another solution.
-->
Which is with NetApp.
-->
That NetApp provides.
-->
An enterprise trust.
-->
With high performance.
-->
Service for file storage.
-->
So.
-->
These ones on top.
-->
Of Azure.
-->
And obviously it has like additional.
-->
Costs. But it's an option that.
-->
You may want to do.
-->
Now.
-->
The other resource that is available.
-->
Here is the.
-->
So the Azure disk.
-->
So the Azure disk is.
-->
Let's say.
-->
Like the ARM disk of your server.
-->
Or of your machine.
-->
So in this case.
-->
In this case.
-->
You may want to.
-->
I mean what do you need to think about.
-->
Here is.
-->
If I'm going to.
-->
Have a.
-->
A standard hard drive.
-->
Or an.
-->
Or a premium.
-->
Or an ultra disk.
-->
So this is the IO.
-->
IO. IO EPS.
-->
EPS.
-->
How are you going to have backups.
-->
With manual.
-->
Shops or fundamental.
-->
Or the service of backups.
-->
If you are going to encrypt.
-->
The disk. Either.
-->
At Azure level.
-->
Or at host level.
-->
So how are you going to do that.
-->
The keys for encryption.
-->
I mean if you use the keys.
-->
You lose the data.
-->
And if you are going to.
-->
Have action or sweeping.
-->
On the.
-->
So you can have like a better.
-->
Performance access.
-->
In your system.
-->
So depending.
-->
On the requirements.
-->
Or the use case. You may want to.
-->
Choose.
-->
Which.
-->
Type of this.
-->
Are you going to use.
-->
So.
-->
Databases.
-->
Or workloads.
-->
Are using.
-->
So this is the.
-->
As our.
-->
Web application.
-->
Usually go for premium.
-->
What is the difference.
-->
On whether you have.
-->
You see.
-->
The right option.
-->
For this as well.
-->
These are like.
-->
The premises.
-->
For the.
-->
IOPs.
-->
Or these.
-->
So.
-->
So it's important to have it.
-->
And finally.
-->
The security.
-->
How to secure the storage.
-->
So the most important thing.
-->
Is.
-->
To not expose.
-->
Your storage account.
-->
Or either your endpoints.
-->
To the public.
-->
To have policies.
-->
To have keys.
-->
Managed for.
-->
On your data.
-->
And the endpoints.
-->
To be divided.
-->
Another important thing.
-->
Is the life cycle.
-->
So.
-->
When.
-->
You can make.
-->
That your storage.
-->
Evolves over the time.
-->
So you don't.
-->
So you can.
-->
Optimize.
-->
The cost usage.
-->
And the access.
-->
To that.
-->
As well.
-->
So this is like an example.
-->
Of how to have.
-->
Some endpoints.
-->
By the endpoints to accessing.
-->
The storage.
-->
And.
-->
And those are.
-->
What we talk about.
-->
So.
-->
What we are going to do in this case study.
-->
Is that we have.
-->
Our company.
-->
With.
-->
Our requirement of what.
-->
To do.
-->
This is page 255.
-->
What to do to design.
-->
A storage solution.
-->
So you need to think that.
-->
As I told you before going to.
-->
Our project.
-->
We have.
-->
Pictures.
-->
Files.
-->
Office documents.
-->
But also sensitive.
-->
Documents.
-->
To.
-->
How to.
-->
Use the storage costs.
-->
How do we.
-->
To take.
-->
The content.
-->
And integrate that content.
-->
Into the cloud.
-->
So we're going to be the files located.
-->
How frequently.
-->
Are we going to access that information.
-->
So.
-->
Maybe as to discuss.
-->
What to do.
-->
Or how to.
-->
Store that.
-->
Or set up the storage.
-->
For the customer.
-->
Also.
-->
There is some exercise.
-->
For those.
-->
Who want to.
-->
Do the actual creation.
-->
Of the object.
-->
Of the resource.
-->
In Azure.
-->
There is an exercise here.
-->
To create.
-->
A storage account.
-->
For these requirements.
-->
So you can also take a look.
-->
Into that.
-->
So.
-->
Let's do that.
-->
I'm going to put this in the.
-->
In the.
-->
In the groups.
-->
And we are going to have like.
-->
Half an hour for this.
-->
We're going to have to have this here.
-->
And this is.
-->
You know.
-->
Okay.
-->
So.
-->
Let's see.
-->
Let's move to the.
-->
To these rooms.
-->
I'm going to open those.
-->
So you can work.
-->
And I will go.
-->
So let's start.
-->
Okay.
-->
Okay.
-->
Okay, hello guys. How are you doing? Okay, no? Okay. Yeah, it's okay. Perfect. So, what
-->
type of data is represented in the examples for this case? Is that structural,
-->
not structural? Okay, you're just correct. Those are files. So, in that sense, we need to take a look
-->
into, okay, regarding files. What kind of criteria should be used for the design? For example,
-->
where I'm going to locate those files? If there is any compliance or regulatory requirements? What
-->
about the performance of the access in those files? Am I going to replicate the storage in
-->
different locations? So, I can have a disaster recovery. So, those kind of stuff we need to think
-->
about. Also, we need to think about what kind of service I'm going to use. Okay, if it's files,
-->
I have the global storage. I have the Azure files. Which one am I going to use the Azure
-->
files for everything? Or perhaps the global storage may be used for some of those files?
-->
Kind of things that are in there. So, try to take a look into that to see what you come up
-->
with regarding that as well. Okay, so take a look. No worries. We can keep going for a while.
-->
So, I'll go to another group. Okay. Hello, guys. How are you doing? Hello, guys. How are you doing?
-->
Okay. No worries. It's okay. It's okay. Perfect. Right. That is correct. That is a good idea.
-->
Okay. Awesome. No worries. Awesome. Perfect. That is great.
-->
Perfect. Okay. You're in the right track. You're doing perfect. So,
-->
keep discussing. And if you can come up with a design, you can draw IU on the icon. So,
-->
that represents this stuff that we are discussing here. That would be super awesome.
-->
Okay. Okay. Okay. Okay. Cool. So, see you later, guys. Thank you.
-->
Hello, guys. How are you doing here? Hello, guys. How are you doing? Okay. That is super
-->
perfect. You're taking notes. Awesome. Awesome. Perfect.
-->
Yeah. Yeah. That would be great. I mean, discuss this exactly like you're doing right here
-->
and try to diagram that discussion into draw IU. So, you can get familiarized with the
-->
icon and how to represent this ADS into an architectural diagram. Because sometimes that
-->
is the hard part. But that is the idea. So, try to create a diagram representing those
-->
actual files that you're talking about, those actual blocks that you're talking about.
-->
Remember that in draw IU, you have the icons for office documents, for databases,
-->
for media files. So, you can use that as well. So, yeah. You're doing great. So,
-->
keep doing that and try to translate that into an architectural diagram for this.
-->
Okay. Awesome. Awesome. See you later, guys.
-->
Hello, guys. How are you doing? Okay.
-->
You have your notes. Awesome. That is important. Perfect.
-->
Yeah. That is. Well, is it semi-structured or un-structured? What do you think?
-->
Okay. That is correct. Yeah. You need to think about that. That is more like
-->
structured data. So, that is correct. And okay. So, keep doing that. I mean,
-->
try to take notes into each one of the points and see what applies and whatnot.
-->
And the idea is that at the end, you can
-->
illustrate in a diagram, in an architectural diagram, the things that you discussed here.
-->
So, remember that you have access to the icons of Azure files and Azure global storage.
-->
So, you can put that into a diagram and put in front of that, for example, a PDF icon or a
-->
media video icon. So, you can represent the different kinds of objects that are going to
-->
be stored here. And if you want to have redundancy, you can use that,
-->
create a diagram to do that. So, try to reflect your discussion into a diagram,
-->
which is hard to express those ideas in a diagram. But it's a good exercise because
-->
at the end, what is seen and what can be used when you are implementing the solutions in Azure.
-->
Okay. Awesome, guys. I'll see you later.
-->
Okay.
-->
Okay.
-->
Let's see.
-->
So, who wants to share what they have done so far? Who wants to share? Go ahead and
-->
share your discussion on the design that you have done so far.
-->
What about Jason? Jason, do you want to share what you guys have done?
-->
Thank you.
-->
Yes, we do.
-->
Mm-hmm.
-->
Okay. Okay. Okay. Okay. Okay. Okay. That is good.
-->
Perfect. Awesome.
-->
Okay. Awesome. So, okay. Let's just have a few observations or if someone wants to chip in.
-->
So, John, go ahead. Let us know your thoughts about...
-->
Okay. Okay. Okay. Never mind. Well, I will just ask the question to all of you guys.
-->
What do you think about the VN within this design? Is it really necessary to have a VN
-->
to store those maybe files, legal documents? What do you think about it? It's valid. I mean,
-->
it can be done heavily, but what do you think? Who you have chosen another service with Azure
-->
to do that same purpose? Anyone wants to discuss or share what they have done?
-->
By in their team regarding that part? Okay. Go. Yeah.
-->
That will be the glove storage. The glove storage, it's like the equivalent.
-->
Mm-hmm. Yes. That is correct. So, okay. Let's see. I don't know. Perhaps Kenki, what do
-->
you think, guys, if you want to share the discussion that you had and if you were able
-->
to create a diagram, if you want to share it with us. Very good. Very good tasting team
-->
and very good girls. So, let's see another group. Kenki, do you guys want to share
-->
what you did, the discussion that you had? No matter if there is no diagram yet,
-->
we can see the discussion. Okay. Cool. Thank you.
-->
Okay. No worries. It's okay. Yeah, absolutely. Absolutely. We are building. Again,
-->
I highly recommend you guys to share what you did. All the groups, not only Jason who already
-->
showed us what his group did, but the other groups, just share what you did in the chat
-->
because that will be useful for everyone else to have. So, go ahead. Tell us about
-->
what you did, guys, and the discussion that you had, for sure. Okay. Perfect. Perfect.
-->
So, the data is an instructor. They have a training by Sheriffs. They are planning
-->
to do some data redundancy. Awesome, guys. So, okay. Let me share one, another
-->
possible solution that we may have here. As I told you, it's not like there is only one
-->
solution for this kind of designs, but there are some suggested stuff. Let me show.
-->
Let me find where it is. So, this is the storage solution. So, here first,
-->
okay, I think that every one of us agreed that this is non-instructor data. However,
-->
something that we need to think about as well is regarding which data is accessed by whom.
-->
So, for example, all the marketing material, if you take a look into that marketing material,
-->
it seems that this material is something that is going to be published in the website,
-->
in the web pages. So, for example, that a customer wants to download a case study,
-->
or a data sheet, or a product, or all that stuff. So, this material here is stored in a
-->
storage account. This storage account can be, for example, a blog storage, and that is
-->
accessible for the web application, for the website. So, it can be accessed from the
-->
internet, through the website, and internally from the main office using a file share.
-->
So, this is like an example of how can this part be designed or discussed about.
-->
And in the details of this access, I mean, yeah, you need to be like,
-->
see how to put that together, because that is like the tricky part here. So,
-->
this function that is here, if we go a bit down, you can say, okay, in order to do that,
-->
we will have the internet, an application gateway. We haven't discussed that,
-->
it's a way to access a web application, like a load balancer. So, you have the app service
-->
with the web application, and this app service has access to a storage account,
-->
and leaving that storage account, you have two services. One is a blog storage,
-->
and the other one is five actual files. So, this storage that is here is like the same
-->
that you are seeing here. This is like a functional view, if you want to call it like that.
-->
And here are the access files that, well, it will depend on the policies that you define,
-->
and how do you manage the lifecycle of those files that are in here.
-->
And this Azure files endpoint that is here is, so this user that is internal with the company
-->
can access the documents that are in here. So, this is a way to publish it, but the key
-->
discussion here is, okay, when to use the blog storage, when to use the Azure files,
-->
discussing which access files do you have, you always will need to have a storage account
-->
in order to access these resources. And actually, if we take a look into these,
-->
into the portal, directly into the portal. So, let's see. So,
-->
we have a storage account already created. So, let me see if I have one. This is something
-->
that is already there. So, this is the storage account, and this storage account, for example,
-->
has containers. And these containers, if I click here, so this is like a blog storage
-->
I can have information. So, well, this is empty, there is nothing in there, but I can
-->
put files together. I can have this file share in this storage account. So, this will be like
-->
a folder that is available, and I can browse it here. Well, it has some stuff in here,
-->
some files, but this is like the file server, and see that you can have snapshots here,
-->
and you can clear backups here, and also the queues and the tables. So, these are
-->
types of storage. So, this container is the blog storage, and the file share. So,
-->
and also you do have this storage browser, which is a tool that allows you to see what is inside
-->
of your sources. So, the blog container, the file share, and this is empty because
-->
there is not much to see. But there are some tools here that, for example, if you want to
-->
data from one place to another, from the premises to the cloud, here is how you can do it.
-->
There is this data management stuff that allows you to set up the read and see
-->
what is the storage. So, how do you want to create this, if it's locally, if it's
-->
redundant? Let's see if we can change this. Well, I would like to continue that.
-->
I will have to save that, and there will be another data center available here.
-->
So, this is the way to manage all of the stuff that we were discussing, the inventory.
-->
For example, this one for a separate website. If you want to have a landing page, you can have
-->
this available here. So, this storage account is like the object or the resource
-->
in which you can create all the stuff that we were discussing now.
-->
Okay. So, in this, before going to the launch break, let me go into the two other use cases,
-->
which are the SQL storage and, well, the structured storage and the semi-structured storage.
-->
So, we can, as soon as we are back from the launch break, to execute this case study as well.
-->
So, before going on, please let me know if you have any questions of those design principles
-->
for this storage, this kind of storage. Okay, awesome. So, let's discuss a little bit
-->
regarding relational data or structured data. So, here is where we discuss
-->
H2 SQL databases. So, usually, the design criteria for this kind of storage will be
-->
which database service are we going to use, which file, which size, which option of scalability
-->
are we going to apply, if there will be any encryption where the data is located or when
-->
the data is being transmitted or being used, which solution applies for our relational data,
-->
which higher availability solution are we going to use. So, these are like design criteria and
-->
the items that we are designing here will be SQL database, we are going to use SQL edge,
-->
so database at the end point, or if you are going to use CosmoDB for unstructured data
-->
or the table storage that we execute earlier. So, here with H2 SQL, you have multiple options
-->
to make this or to design this service. So, you can have a virtual machine, like a
-->
digital virtual machine with SQL, with the operating system, so you can access to this
-->
machine and have your database in here, that is an option. And in this case, you usually are
-->
the one who manage this virtual machine. The only thing that happens here is that it is
-->
installed by or provisioned by Azure, and that in Azure well is provisioned
-->
from there, so you don't need to install the actual database in the virtual machine,
-->
that is what happens here. There is this other option, which are the managate instances.
-->
So, these are single instances of SQL server, which are kind of a virtual machine, but here
-->
you only have access to the SQL, you don't have access to the operating system,
-->
so you have only the SQL access here. And you can have either one instance,
-->
or you can have access to a pool of instances, that is when you are sharing resources with other
-->
customers or SQL instances to optimize the price. So, you have in the end,
-->
at the end, this is a virtual server, but you only have access to the instance that is assigned
-->
to you to have your database, so the machine is being used by multiple SQL instances.
-->
So, this is also to manage as well. And finally, there is the proper database,
-->
which is the hyper-scale, and this proper database is when you're talking about a
-->
big database of just 100 terabytes, and this is an instance that has all the services provided,
-->
it is accessed only by you. And also too, you can have an elastic pool that gives you
-->
the ability to optimize the pricing of the resources, but also gives you more performance,
-->
which is kind of similar to these two offerings. So, you can see that here you have a grid,
-->
or there is a recognition, but you don't see, and here is just the database running on the
-->
technology provided by Azure Azure. So, there is also a business model of pricing model for
-->
these offerings. So, the data transaction unit is a scale between compute and storage that
-->
gives you a certain amount of CPU, memory, reads and writes, so you can pay for
-->
according to how much power your database needs, or there is regarding the virtual
-->
calls that are assigned to the system, and the ability of those virtual calls.
-->
All the serverless, which is whenever there is a unit of your database,
-->
you will have access to it, and it will be built according to the usage of the top.
-->
So, it is great, but it may be, I mean, it won't be always be there,
-->
it will be there whenever it is requested, and that request may take a few seconds to be
-->
available. So, you need to use the function, but it's one option that is in there,
-->
or you need to design the applications having that in mind, so you don't have those kind of
-->
issues. So, there is a strategy to scale databases. So, there are different scenarios
-->
or solutions for this kind of scaling. So, if you have multiple Azure SQL databases,
-->
you can scale with Elastic Schools. If you have one application with just one database,
-->
but you want to test that, so you can use Azure SQL database or a managed instance,
-->
or if you want to optimize the price versus what you have in database,
-->
but having Elastic Systems, you can also have the same Elastic School.
-->
So, it really depends on how you want to handle the pricing. So, this needs to be
-->
like, be analyzed in order to avoid soft prices. Now, regarding the high availability
-->
of your database, there are some tires. There are four post-standard tires. So,
-->
this is one of the three service tires that Azure SQL has. So, these two general
-->
proposal standards provide you with a common base for any application to have access to
-->
the database. It's balanced between complete and storage. It has some nodes available with
-->
capacity. So, if there is a need to have an instance to take over the binary
-->
particular, it can be done. It has some storage for data and for logs, and it has
-->
some backup in the standard storage available to tell you a situation or
-->
a failure of your database. So, you can recover from that. There is another level
-->
issues there, the business critical opinion tire. So, in this case, you have not only
-->
a binary replica, but you have secondary replicas ready to do this. So, in this case,
-->
you have the backup. So, in this case, what happens is that one node with a spare capacity
-->
will take the backup and provide it with the service, but it will be a slow process
-->
to recover. In this case, you have a primary replica, and in this case, the
-->
this other instance that is already running the data, practically the data online
-->
because it's been replicated all the time. So, this is the opinion tire for construction
-->
applications. The high transaction, low IOU,
-->
have the highest experience scenarios. It has the ability to have an always-on
-->
service. So, this service is available. It uses SSD and replica,
-->
and then there is the high-per-style history tire. This is for very large
-->
large construction databases. It can out-scale the storage and the input. It has
-->
some tiny backups. It can be restored in minutes. It can scale up or down according to the
-->
workload changes. So, this is like the services that this has. It uses storage. It uses
-->
phasing servers. It uses some primary computes and secondary computes to read and to read only.
-->
Read and write and read only, and those are like the three-folders. Also, you can have
-->
or you can design a very powerful strategy. So, in this case, you can have a front-end.
-->
In this case, with a balancer, and you can have two primary SPL servers, with a secondary
-->
SPL server with some new replication of SPL in different regions. So, you can have
-->
other systems to use without any need. So, SHIELD is the strategy to similar database.
-->
So, if you are integrating SPL to Azure, so you can choose any of those. If you want to
-->
have the minimum cost possible, you can use this kind of pricing. If you want to have
-->
replication, you can have these failover groups. You can have elastic pools to
-->
minimize and speed the most possible. If you want to have higher level databases, you can use
-->
this hyper scale as well. So, here in the lab 22, you can take a look later. So,
-->
you can create an SPL database and see all the stuff, all those options that are available
-->
in the screen. Now, regarding the security of the database, there is, if you need to use
-->
or apply a regular approach to the data. So, there should be network security with virtual
-->
nets, firewalls, firewall links to access the database. There should be identity and
-->
access management, authentication, robot role-based access controls. There should be
-->
data protection here. So, encryption, when it is used, when it is stored, when it is
-->
lost on the internet. So, this is something that needs to be decided because everything
-->
should affect something. If you have encryption, right, you should be aware that this affects
-->
the usage of CPUs and it will require more resources. And if your pricing model depends
-->
on the cost and the use of the cost, you will increase the pricing, right? So, you need to
-->
Okay, I'll do 500 that. So, that is simple. And security management, there is some tools
-->
additionally that allow you to have some of these, have three protection laws and
-->
a few assessments to support the database as required. Now, regarding authentication,
-->
the SQL supports the authentication against the active directory
-->
and also the local authentication of SQL as well. So, the application from
-->
There is another product which is the Azure SQL Edge but so you can have
-->
satellite databases for ingress data on the Edge. They are usually used for IoT
-->
but have been changed in the last few years. Now, we're going on a sector data. We have
-->
the Proxmox DB. So, in this case, what we have is the SPIN database which is a relational
-->
approach that allows you to have a lot of tools and stuff like that versus non-SQL
-->
database in which you have either t-value or tolerance or documents, graphs or graphs
-->
to store the data. So, these are like the common differences. It's being used
-->
over the time more and more, the no-SQL option. But probably SQL is the team that
-->
no-SQL is getting more commonly used. So, these are the copies and the pictures of each
-->
one of those. So, the biggest difference is that these are more flexible than these ones.
-->
So, that is one of the reasons that this is an interaction. And also, the kind of
-->
documents that are getting more and more popular as well. This is the Azure Proxmox DB
-->
which supports all the non-spectral data instances, the t-value program,
-->
documents, graphs, and it supports a lot of languages to access it from as well.
-->
This also has, one of the biggest features of this is that it gives you
-->
scalability across the world and full tolerance across the world because this
-->
has the response time. So, this is really good for that kind of stuff. So, low-latency
-->
is one of the two options that you have on your application. This is something that needs
-->
to be considered because achieving that in SQL is not possible at all.
-->
So, there are some use cases for this kind of databases, whether on mobile applications,
-->
with Gain IoT, for example. Here you have a decision tree to decide which kind of
-->
non-spectral database you should use for your application in particular or for your
-->
space as well. So, this lab can actually also create one of these databases.
-->
So, you can have like the idea of how this is created as well.
-->
Now, finally, the algorithm to choose which kind of structured data product are you going
-->
to use. So, if you want to control the database software and the operating system, so you go
-->
using a virtual machine deployment. If you don't need that, you can choose which kind of
-->
database. If it is non-spectral or spectral, you can use SQL. So, you can use one of these.
-->
If not, you can use, you can see here how Microsoft
-->
own it, the SQL terminal, right? Because they use SQL to Microsoft SQL.
-->
This is the more likely scenario. But when they share this, they are asking if it is
-->
Microsoft SQL. If not, it is MySQL or MariaDB or Postgre. So, these are like options
-->
available. And in the case of SQL, you have the option with Azure machine or directly
-->
SQL database with Azure machine. And this other is the options
-->
for when you don't need this kind of stuff, if you want to go to SQL
-->
guidelines. So, what we are going to do is as soon as we come back from the
-->
launch break, we're going to go into the breakout rooms to execute this case study,
-->
which is to design a relational storage case study. And the idea here is to design a
-->
solution including authorization, authentication, pricing, performance,
-->
high availability, and diagram what is decided for this kind of solution.
-->
So, this is like the guide for this case study, design the solution. You will have a
-->
application. You will want to have a MySQL server. These are some features that are
-->
expected. And here, these are the details. So, this is the template for the whole solution
-->
of this case study. So, this is a really good loop. We probably won't have the whole
-->
time that is needed in order to design this whole solution. However, it means that you
-->
take this and read carefully what is explained in here and try to use this template to
-->
different components for having the solution working. So, just to give you an example,
-->
here is like what it's showing in this slide.
-->
So, what we are seeing there in that template is this front door is called this service
-->
and the load balancers to provide a failover. So, these are the different options that you
-->
may want to find because it is for general purpose, product, for enterprise level,
-->
environment. So, let's go for our lunch break. We have like one hour. Let me see. You guys
-->
can have one here. So, we can come to 2 p.m. and we start as soon as we arrive to the
-->
room. Let's meet in one hour. See you then. Thank you.
-->
Oh, I love your outfit. What are you wearing?
-->
I'm going to send you guys to the breakout rooms.
-->
We can start with the case study. Remember, let me check which are the files.
-->
Here is the 295. So, it's the 295 as like 295 up to 298.
-->
So, see you in the breakout rooms. We will have at least 30 minutes, 30-40 minutes
-->
because this is kind of a long life. See you then.
-->
Okay, guys. So, how is it going?
-->
Hi, guys. So, how is it going? What do you have? What have you seen so far?
-->
Okay. Okay. Okay. It's okay. Keep going and discuss and try to see what fits in and what
-->
doesn't fit. So, see you later, guys. Thank you. Hi, guys. How are you doing?
-->
Oh, okay. Cool. Okay. Perfect. Awesome. Awesome. Keep going and we see each other.
-->
You know what? Thank you. See you later. Hi, guys. How are you doing? Oh, okay. You
-->
know what? Okay. Awesome. See you later. Keep going.
-->
Keep going. Oh, Kempis, you're by yourself. What happened? Oh, okay. Okay. Okay. How
-->
are you doing? Okay. Awesome. Awesome. Okay. Cool. Keep going and anything. Just
-->
ask me if you need anything. Okay? Okay. Okay. Thank you. See you later.
-->
Okay. Okay, guys. Let's see if we are all here.
-->
Okay. Let's continue then with the other process. So, who wants to share what they
-->
have designed or no matter that it's not complete yet, we can guess what you have
-->
so far. So, go ahead. Who wants to share what they did during this space? Go ahead,
-->
Alexander. For sure, if you want to, go ahead. Please share your screen and tell
-->
us about it. Okay.
-->
Awesome. It's a really good design. I like a lot that it's...
-->
It's simple and that doesn't mean that it's bad or something, no? Remember the
-->
principle of KISS, which is keep it simple. So, this works. I mean, you have the
-->
application service. You have the backend for the database with a reputation in multiple
-->
ways. So, and you have authorization. It works. It's very good. Really good.
-->
Congratulations, guys. Okay. Awesome. Awesome. Awesome. So, perfect. So, who
-->
wants to share their design as well? So, we can enhance and see what we have
-->
produced in this afternoon. Who else wants to share that? Don't make me call
-->
you. You know who. Go ahead and share your screen and let us see what you
-->
have done. Who else has the design? I know that you guys did several designs
-->
over there. Just share your screen and let's see it. Because in the previous
-->
design, it's simple. It fulfills the requirements, but you can add some
-->
stuff in there to make a more robust website and database backend, for
-->
sure. So, let me show. Who wants to share their design? Hi, guys. No?
-->
No one? Let's see. Let's see. Marfibel, a lady that I haven't
-->
heard you guys talk. Marfibel, do you want to share the design that you
-->
guys did? Hi. Okay. She can share it. Ah, you were with the same team,
-->
with Blair and Alexander? Ah, okay. Okay. Okay. Ah, with Jason. Ah,
-->
go ahead. Go ahead and share that and let us know how it is. Okay. No
-->
worries. No worries. It's okay. It's okay. Awesome. Okay. Yeah, we have.
-->
Okay. So, okay. You have the first part and the backend. That is okay.
-->
We all know this. Okay. Okay. Oh, I got it. Okay. I got it. Okay. No
-->
worries. Okay. Okay. Perfect. Got it. So, this is interesting because, I
-->
use CosmosDB as the main database and what they are doing is to move data
-->
from CosmosDB to SQL for the long term. That can be done. Function can
-->
be used to do that. It is correct. But here, as we are part of the
-->
team, in real life, these kind of decisions are made with the developers,
-->
the architects, the cloud architects. So, it's not just only us who define
-->
or decide what is going to be in this solution. So, what you will find
-->
usually in the wild is that the SQL database is used for a lot of
-->
the structure of the website. Information, users, catalogs, a lot of
-->
stuff. And the MongoDB or the CosmosDB is used for certain information as
-->
for example, sessions or indexing or search of web pages and keywords
-->
within the website. So, usually what you have is a mix of both of those.
-->
So, let me show you one proposal of solution for this implementation.
-->
This is a... Where is it? Let me see. Where is that?
-->
So, one proposal of solution for this design because you can see here
-->
that, okay, there is a lot of stuff in this template to be filled. So,
-->
they talk about the two regions for redundancy. So, that means that you
-->
will have redundancy across regions that you will want to replicate
-->
the website. So, for example, we are discussing here mostly databases,
-->
but if you mix or merge what we had from the previous case study with
-->
this new case study, perhaps you can add here an Azure files storage or
-->
an Azure blog storage for serving those PDF files and that stuff as
-->
well. So, that is something that goes here as well. They always
-->
work because of the servers, of the vacants, the database. They want the
-->
database to be high level again. And the performance, the searching pages.
-->
So, this case is almost every time use is solved with MongoDB,
-->
because indexing and searching in this kind of non-structured databases or
-->
semi-structured databases is way much faster than a query in a standard
-->
SQL database. And regarding security, well, it's the active directory
-->
authentication. So, here is a design with all the components. So,
-->
the high availability with a traffic manager that sends traffic to either of
-->
both of these websites or just one website. And in case of failure,
-->
it sends this to the other one, the web tire with the app service.
-->
This is something that we didn't discuss, but it's good for you to
-->
learn about it. It's a service that is the web is touching or main
-->
cache. It's another option. This is a service that is specifically designed
-->
to create a session cache. And what is the idea of this kind of cache
-->
already that is here or main cache, which is the other typical part that
-->
you see here. For example, if this is the web application for a bank
-->
and the user is here working in this, doing some transactions in this
-->
web tire and for some reason this server went down or this environment
-->
goes down. And the user that was making, let's say, a transfer,
-->
a money transfer, a wire transfer. So, the user gets redirected to this
-->
other environment, this region, and that is done by the traffic manager.
-->
And the magic that happens here sends to this caching release or this main
-->
caching is that when that failover occurs, the session information
-->
for that user regarding what she was transferring on all of these things
-->
are not lost, but are stored in this cache and they were replicated.
-->
So, the user can continue with whatever he was doing and he will not even notice
-->
that there was a change of location in the experience that this person had
-->
on the website. So, that is what this kind of caching allows you to have,
-->
which is super nice. You can have this same architecture either using the
-->
caching release. What would be the difference? That if the user was doing
-->
a transaction here and this fails, when she is moved or this person is moved
-->
to the other region, they will need to start over the transaction as it has
-->
never occurred because there is no information here about what she was
-->
doing at that moment. Because as this person has not finished or had not
-->
finished the process, it wasn't stored yet in any of these databases.
-->
So, that is the function of this caching release. So, you can have it in that.
-->
So, going after that is the data tire. So, this data tire, here you have the
-->
skilled database with the replication. Please notice that there is a
-->
link. So, these regions are internally connected, are not exposed to the
-->
internet. So, this is a secure replication. And there is a Cosmos DB
-->
for that indexing, for that distribution experience that they want to
-->
remove. So, the Cosmos DB is in charge of that stuff here with the network
-->
security groups, the defenses, defenders. So, for any tricks that happen
-->
on the internet. So, this is like an example of this diagram or this
-->
architecture that you can create. But the approach, it's, I mean,
-->
what we did was perfect because first you need to go simple. The simple,
-->
the better. So, the first, actually both the ones that we saw are able
-->
to give you something from working deployment. So, you have the database,
-->
you have the web service, you have a redundancy or reputation. So,
-->
you're good to go. And you can enhance or build on top of that. So,
-->
so that is important. So, you can wish this kind of stuff. And please
-->
have this in mind, the caching release, because as I told you,
-->
really, really, really, really nice. Okay. So, we are going to jump a
-->
little bit because I want you to work with the compute case study
-->
because this is huge. So, perhaps this will be the last module for
-->
today. But let's see if we are able to develop another module.
-->
However, this module is super important because this is like the core
-->
of any solution in the cloud. Okay. So, this is the compute. So,
-->
here we are going to discuss the different compute-based services that
-->
are available within Azure. So, the virtual machines, as I told you,
-->
as we have discussed, are the most common ones, compute services used in
-->
Azure, but they are not necessarily the feed-out solution. So,
-->
there are additional options that should be taken into account whenever
-->
you need it. So, there are the batch for running jobs, parallel jobs.
-->
There are the app services that we already discussed previously.
-->
There are the functions to run codes, the containers, the service
-->
fabric for microservices, and the container, the Kubernetes for
-->
containers as well. So, this is the algorithm that we already discussed
-->
earlier today. So, this algorithm is super important. We need to have it
-->
really available and start to apply it. And over time, you will know which
-->
is better and why. So, just keep that in mind.
-->
Now, regarding virtual machines. So, the virtual machines, perhaps you
-->
can say that or you can think that it's simple because, okay, it's
-->
just a machine and we usually deploy virtual machines all the time.
-->
However, when we are discussing virtual machines in the cloud, we
-->
need to think a little bit about it, how this works. So, from the name,
-->
remember the standard, the naming standards that we discussed yesterday,
-->
the location of this machine, the sizing, the price, the availability,
-->
the kind of disks that you are going to use, the image, if we're going to use
-->
an image provided by the marketplace or if we're going to use a custom image
-->
by ourselves, created by ourselves, the operating system, and who is
-->
responsible for this machine, if you or Microsoft. So, these are
-->
the stuff that needs to be taken into account whenever you are choosing
-->
a virtual machine. So, when you use a virtual machine, when you want to press
-->
something, I'm going to give you a super simple example, but real life.
-->
In my personal case, I need to, I mean, I use a Mac, a Mac, but a Mac
-->
OS X, and I need to deploy a Windows machine because there is a software
-->
that only runs in Windows. So, I mean, if I can't have a virtual machine
-->
on my laptop, but it will take space, I don't use it like every two,
-->
three months, wherever, something like that. So, it's not worthy to have
-->
all that space and that set up in place. So, it's easier just to access
-->
the portal Azure, deploy a Windows virtual machine, install the Windows
-->
software that I need, obtain the results, and destroy the machine.
-->
So, it's quickly testing something, finishing up something. So, that is
-->
something that is super useful in this type of machines. If you want
-->
to do something really particular with something that is available there,
-->
like GPU that you can use from Azure, if you want to extend what you
-->
already have, but you don't want to purchase any additional hardware,
-->
so you can just deploy some machines. So, the greatest example was the
-->
pandemic. So, I mean, basically, all the clouds reached their limits during
-->
the pandemic because everyone moved their workloads to the cloud and it
-->
took them almost a month to recover, to being able to provide service
-->
to everyone. So, it was really, really hard that time. If you want to run a
-->
legacy app that has the most modern hardware, or with software, or if you want
-->
to quickly integrate an application, so these are several reasons you want to
-->
use Azure Touch. So, what is the deployment that needs to be done?
-->
So, this is part of the decision thing. If you want to build a new one or if you
-->
want to integrate. So, the decisions that need to be
-->
defined, the network, the location, storage, they will erase the
-->
memory of this. There are several kinds of virtual machines.
-->
The general purpose ones, the compute optimizer, the memory optimizer, the
-->
storage, the GPU, the HPC. So, it's like its name
-->
says it all, but obviously, this is directly related to the price
-->
of each one of those kinds or family of information. There is
-->
a feature that is not that widely known, and it is the virtual machine
-->
spacesets. And these spacesets are a way to have higher reliability with
-->
virtual machines. I mean, you can explore your applications not in one machine,
-->
but in a virtual machine spaceset. So, this will allow you to increase
-->
the power of your application to have high availability using a traditional
-->
virtual machine. So, this will be interesting. This is powerful. This
-->
has a load balancer. It allows you to use space. It allows you to have a
-->
special image. You can define these availability sets, how they are going to
-->
be provided. You can define which zones you will be using to have this high
-->
availability solution. You can have a way to orchestrate this kind of
-->
deployments. You can have auto scaling, because if there is a lot of CPU
-->
usage, you have to have some large additional information to top up with
-->
the load that is being demanded. So, that is a really nice feature from
-->
Azure. So, you can take a look into that. Obviously, it has a trick.
-->
Your application needs to support this kind of operation. And not all
-->
applications can do that. So, you can have groups of virtual machines, or you
-->
can have a scale set of virtual machines. This is an example. In this
-->
example, we have the virtual machine, which probably has a network into the
-->
sky. They are using one of these. So, this is like the full design of
-->
the thing. And those are the complete services. There are these complete
-->
things and services. I will be showing what the ones that we will be
-->
using. The ones that are in the right box are the ones that are in the
-->
left box. Another component that is required in this kind of stack are
-->
the virtual networks. So, the virtual networks are the cloud network for
-->
deployment. So, these networks can have public endpoints and that can be
-->
accessed from anywhere on the internet. Or find out endpoints that only
-->
can be accessed from the network. You can have subnetting inside the
-->
virtual networks. And you can have connections to these networks.
-->
And you can have a design of the virtual network inside the cloud.
-->
So, additionally, in networking, you will find different networking
-->
services. For example, VPN networks. So, you can have a VPN network that
-->
gives you access to the virtual network in the cloud. And in order to
-->
access this VPN network, you will need your on-premises location, an
-->
endpoint for an IT set VPN to establish a tunnel. So, you can route
-->
traffic from on-premises to the cloud and vice versa. And here is
-->
something that it should not be overlooked, but it is the subnetting of
-->
both your on-premises network and your cloud networks. Because if you
-->
make the rookie mistake of having the same subnet here on-premises and have
-->
that same subnet here on the cloud, you won't have communication between
-->
both networks, right? Because there will be no way of routing those
-->
packets. Because it's the same thing. So, you need to think about how to
-->
route the traffic between those components. There is another solution
-->
that is the ExpressVault. This ExpressVault is an endpoint that is
-->
provided by Microsoft partner, usually a data center that has a direct
-->
connection to Azure. And they are able to give you access. So, you can have
-->
with a private connection for you, for your company, to access all the
-->
office servers, services, the public IT addresses of Azure, and all the
-->
traffic for the on-premises network. So, this is like a private connection.
-->
But it is provided by a telecom vendor that has this kind of arrangement
-->
to ensure in the country. In particular, at the end of last night.
-->
Another service that is available here is the DNS service. This DNS service
-->
has the ability of the DNS, or the FQD end-ends that you have created for
-->
the services that you are exposing to the internet. So, those are like the way to
-->
reach out to these services for the internet. There is another complete service
-->
that is the Batch. This Batch is usually used for high-performance computing.
-->
But it can be used for other kind of jobs, but you need just to check the
-->
But the idea is that this is something that requires high intensity or high power of
-->
complete power of processing the data. And then just once that job is executed,
-->
you reduce the machine and you don't use it anymore. So, that is like the idea of this
-->
Batch processing solution. So, it depends on the industry and the use case in particular.
-->
So, those are not that common. You'll see in the specific industries where you can see
-->
this kind of stuff. This is an example. The movie industry is a big user of this
-->
kind of solutions. So, all the rendering of 3D images can be done in this kind of stuff.
-->
So, the rendering, you know that rendering a 3D image, a 3D video can take hours of
-->
computing power. So, this is those kind of solutions that are used in that kind of stuff.
-->
There are some machines that are available within
-->
within Azure that are well suited for this kind of jobs. In the sense that there are machines
-->
that can use or can have a separate peak usage for a specific or a short period of time and
-->
just don't have enough power to be used at all or to be used in this kind of tasks. So,
-->
these are machines that don't guarantee the same level of performance all the time. So,
-->
those are cheaper than other ones, but it may adjust for the systems better because it's
-->
they are no quantity of performance. It can be way cheaper than any other
-->
kind of machine that is available. Also, there are the spot instances which are similar.
-->
So, the spot instances are like some, I forgot the word,
-->
so, when several users compete for having access to the compute, they make offers of
-->
how much they are willing to pay for having that compute power and regarding the different
-->
offers the prices settle and that way you can buy or have access to this compute power. So,
-->
this is like having dynamic pricing and obviously it can increase, but also it can decrease. So,
-->
this is another case for this kind of workloads.
-->
Another complete solution is the container systems. Although this was depreciated,
-->
so it's not that used, but this container is just a simple way to launch a container and
-->
get one. So, the containers are much more lighter than a virtual machine
-->
because they just have the image and the container portion of the image to one
-->
and the virtual machine has like the full operating system. So, that is here.
-->
So, this is an example of how to have a pipeline similar to the ones that we deployed
-->
previously. So, in this case, we have the container register and they are deploying
-->
in this example to QNH services those containers. These can be replaced with
-->
container instances if you only have one or a few images to one.
-->
And Kubernetes, which is like the most common. Well, not necessarily the most
-->
common, but it becomes more and more mainstream every day. So, it's becoming the factor way
-->
to launch workloads into the cloud because the Kubernetes has the feature to be cloud
-->
diagnostic. So, you can deploy on top of Azure Kubernetes, and you can do this same
-->
deployment in Amazon or in Google. So, you will have cloud support, and you will
-->
have a, like, and you will break the vendor locking that comes with it. So,
-->
that is what is being used more and more often. Another feature that comes with
-->
Kubernetes is the high availability by default. So, with Kubernetes clusters,
-->
you have a complete high availability. And if you merge that with the storage
-->
and the replicas, you will be able to have multiple clusters in different regions
-->
providing service to your applications with the backend fully replicated as well.
-->
And also, there are the functions. So, these functions are serverless
-->
applications for executing, for example, a web service, a web API endpoint, or an event
-->
to execute an event within your system. So, when to use Azure Functions? When you
-->
run something short term. So, the main characteristic of the Azure Function is that
-->
you, I mean, this is something that is supposed to run for no more than 10 to 15 seconds.
-->
If the execution of your function requires more time than that, this is not your solution.
-->
So, you need either to design the function to be able to finish the execution within
-->
that time frame, 10, 15 seconds, or otherwise, you need to choose a container or an app service
-->
or a different solution to run it. So, this is important because this is really useful for
-->
small stuff that you may need to do or to have within your system.
-->
This is an example of using Azure Functions. In this case, is to use different features of
-->
the AI services for translation or vision or replication within the cloud. It's an example.
-->
There is another example here with classification of documents
-->
with Azure. So, those are reference architectures that can be used. And in this case,
-->
is for functions in this table. It doesn't show, but what this does is just like a function
-->
running. And so, I can use some branches of a specific one.
-->
There is an additional kind of function that is the Logitech App. So, these Logitech
-->
Apps, these are workflows with Logitech workflows which are figured. So, for example,
-->
if a message arrives, this Logitech App will, for example, send that message to
-->
distributed domains or it will use that to send a tweet or send a record to Dynamics
-->
or send a file to Dropbox. So, those are the Logitech Apps that are available here.
-->
So, this decision tree allows you to decide if you want to use or not a Logitech App
-->
for your service. So, this is a great way to integrate these different systems that are not
-->
signed to be integrated, but you may need to do that. I mean, you want to store every tweet
-->
regarding the product launch that your company did last week into Dropbox resource
-->
because the marketing people wants to check those tweets.
-->
These are some reference architectures on example here for COVID-19
-->
and the comparison between Logitech and Logitech Apps and functions. So, Logitech Apps is more
-->
low cost function, it's cost. So, those are like ones of the difference.
-->
And this is like most likely to connect different platforms and this is there for events.
-->
And this is an example of using functions. We, in this case, are sharing location service.
-->
So, it is used here as well. So, the comparison is here.
-->
So, here is our case study. There is this, again, the decision tree, so you can take a look
-->
into that. And in this case, the case study is passed and the idea here is to design a full
-->
pledge app solution for this company because you require to have a funnel, a prior as a funnel
-->
in which you will have the web application or the stuff that needs to be in front of the
-->
system. And there will be a middle prior in which you will deploy the functions and the
-->
stuff that is required by this solution, by this application.
-->
So, this is the example of the solution. So, there is an application that requires some
-->
servers, some point users, Logitech.net and some databases. This application has a heavy
-->
demand. The servers can reach their performance limits in the same day, but the servers also
-->
sit here for a few hours. So, what should I do or what should I use to provide the service?
-->
Here are the middleware with six API calls from the funnel. And all the time, depending on
-->
the workload that is here, will vary the workload or the demand for these services.
-->
And there is a server that requires the best performance available here. So, which service
-->
are you going to recommend for each one of these carriers and try to create a diagram
-->
to explain the solution? You can create a few kinds of solutions for this. So, let's do this.
-->
This is like 386 up to 388. So, I'm going to put you guys in the
-->
record. So, let's work with this and define a compute solution. So, this is the decision
-->
tree. So, you can decide what will be useful here. You can use chapter 2 again to discuss
-->
as well to help you decide. So, go for it and let's give it a try. We have the USAIS
-->
compute solution. So, I'm going to send it to you and I'll meet you there in a while.
-->
Hello, guys. How are you doing? What have you found so far?
-->
You can add, I mean, it works. It's simple. You can enhance these, adding some replication to
-->
your database backend. So, you can be resilient to failures as well. But yeah, it looks good.
-->
Okay. Okay, guys. So, let's keep working and see you in a few minutes. Okay. See you.
-->
Hey, guys. I cannot see you.
-->
Yeah, it's okay. Yeah.
-->
Oh, awesome. Okay. So, tell me about your design, guys. Let me know how it is.
-->
Okay. Sounds really good. Sounds really good. Okay. So, okay. Keep working on that and we
-->
will meet later. All right. So, see you in a while. Thank you.
-->
All right. Hello, guys. How are you doing? How is it going? Awesome. Awesome. Let me see
-->
it. Let me see it. Yeah. I'm not seeing this. Oh, okay. Okay. Okay. Cool.
-->
Oh, this is really good. Okay. It's really good. I like it a lot. This is a good architecture,
-->
guys. Very good, though. Okay. Well, the service part usually is more like a message
-->
or when you are doing integrations. So, actually, what is happening is that it's being,
-->
I mean, what was used to be deployed on top of service power is now is being deployed
-->
on top of Kubernetes is what is happening, to be honest. But yeah, because it's a service bus
-->
at the end. So, it's like an integrator.
-->
Exactly. That is correct. That is correct. So, awesome. Awesome. Okay.
-->
Okay. No worries. So, see you later. And Blair, for sure, I'm going to ask you to
-->
show your design, guys, to the group. Okay? Okay. See you nowhere. Thank you.
-->
Let me, give me just a minute.
-->
Okay. I'm back. Finally. Oh, this is a cool design. I like it. So, talk to me about it.
-->
Let me, let me see how it is. Okay. Okay. Let me see the, I'm not being able to see
-->
it. Okay. So, you have a little bit over there. Okay. Okay. Okay. It's simple.
-->
You have the three tires for the front end, the middle tire,
-->
the back end, and you're using the end. Okay. Cool. So, awesome. So, we are going to go back
-->
to the main room and we're just going to expose both designs. Okay. So, see you in a few minutes.
-->
See you. Okay, guys. Welcome back. So, awesome. I saw the designs and I have to say that,
-->
I mean, almost, yeah, I would say that all of the designs are different to each other
-->
and they are really good. So, that is super cool. Please, please do share all the designs
-->
that you have done so far in the chat. So, the other, your other colleagues can take a look
-->
into what you have done. So, we can like, spread the additions that we have. So,
-->
I would love if you guys can take a few minutes to expose the designs that you
-->
have made. So, first, the team one, I'm not sure if it's Kevin or Lynn, who is the one who has
-->
the diagram. So, to show the screen and explain what you did, guys, what you guys did.
-->
So, John or Kevin or Lynn or Pierce, whoever you want to show the screen and explain it to us.
-->
Okay. Yeah, we can see that. Okay. Okay. That is good. Okay. So, please, guys,
-->
this is a straightforward design that has left two main levels. Perhaps what I will suggest you
-->
to take a look is into the middle tier that perhaps it will require an additional service
-->
or a function or a microservice to run that .NET application that is being discussed in the
-->
back. It works. I mean, you can have this and make this happen for sure. So, thank you for
-->
sharing. And let's go to the team two to see their design. So, I'm not sure if Jason or
-->
Murphyville, if you guys have your design and want to share it or explain what you did, guys.
-->
Yeah, it is there. So, okay, cool. Let's walk past through it.
-->
Okay. Okay. Okay. Okay. Okay. Okay. Cool. Interesting. Interesting. This is an interesting
-->
design. So, please be sure to share that again with your colleagues. Okay. Thank you so much.
-->
Very good. Now, let's see. Alexander or Blair or Amitri Charmaine wants to
-->
expose the solution that you guys designed. Okay. Perfect. So, it's the three tires,
-->
the app service for the unload balance for the website, the middle tier is Kubernetes with
-->
some IP gateways to access those services. Okay. Perfect. And the database is with
-->
redundancy replication, cross replication. Okay. So, that is super cool as well.
-->
Okay, guys. Very good. Very good. So, let's see the last team. So,
-->
Kevin Tempe-Pears, whenever you're ready, want to explain the design. Okay. Thank you.
-->
Yes, I think we can see the rest. So, please, guys, pay attention because this design
-->
adopts a different approach to what you have done previously. And it's valid. I mean,
-->
it's not, as I told you, there is not a bad or a good solution. So, just different approaches.
-->
And the idea is that you can improve and build on top of that. So, go ahead and explain that
-->
to us. Thank you. Okay. Awesome. Awesome. Perfect. So, very good, guys. So,
-->
what happens here? Okay. So, this approach is using more like a traditional approach of
-->
using the redundancy to deploy the different workloads at each one of the five. So,
-->
very good indeed. So, please, please share the designs. And let me show you
-->
two proposed designs or solutions for this case study. So, you can have also an idea of
-->
a different approach to solve this case. So, here it is. So, the web servers,
-->
the middle world with that net and the backend with the still server. So, in this case, we can
-->
have an infrastructure as a service solution, which actually is the one that was just being
-->
proposed. So, it's kind of like the same proposal that the system did. So, in this case,
-->
with infrastructure as a service, they have a load balancer with a front end running in
-->
sets of virtual machines. And also, the middle player will have another load balancer. And
-->
again, with the virtual machine sets to handle the load. And finally, another load balancer
-->
with SQL virtual machine running the backend. It has like a storage account and the
-->
directory. But notice, please, that it is the same idea, the same approach that was done.
-->
Obviously, as this is like your first exposure to Azure and all that stuff,
-->
there is a lot of things that can be drawn from these days. One of those is the
-->
icon set, the use of the icon sets. There is a specific icon set that needs to be used if you
-->
want to represent a state set. So, you can dig a little bit, explore a little bit, so you can
-->
have other idea how to use those. However, in Azure, you can have exactly the same
-->
architecture that is here. But in this case, you can see that they are using just
-->
virtual machines, multiple virtual machines, but it's like the same example of the state set.
-->
So, this is one solution. Another solution that is proposed here is kind of similar to the one
-->
that was proposed by two of the teams that we saw, that we discussed. Well, actually,
-->
the other two teams, I will say, China, because there were two teams that used
-->
microservices without QNX, which is super cool, and it's a valid option.
-->
Hold that thought. I'm going to discuss a little bit about that. However, just to discuss
-->
this approach, which will be like a platform as a service solution. It's using the gateway
-->
and the app services for the command, and it's using function apps for the middle
-->
world, for the webnet application. And in the back end, there is an SQL database for that
-->
storage of things. This is like the same design that we participated. So, this is another
-->
solution. So, if we take a look into what we have done this last hour, we came up with
-->
about five, six different solutions that fulfill the requirement. Obviously, you can improve,
-->
you can make it simpler, you can make it more complex, but in the general idea,
-->
different components that are in there. Now, regarding Kubernetes, personally, I really,
-->
really like Kubernetes. I find it really attractive. It's really beautiful. However,
-->
Kubernetes may be sometimes above what is needed, an over complex solution to what is needed.
-->
So, you can have the same solution simpler and less expensive than using Kubernetes.
-->
So, because Kubernetes requires a really big infrastructure in order to completely work. So,
-->
sometimes, I mean, it is useful when the size of your application is big enough
-->
and complex enough to justify the use of Kubernetes. And the reality is that
-->
the big amount of companies that are moving into the cloud, that are modernizing their services,
-->
they are not just there. They are just starting and so on. So, there is still a gap.
-->
Obviously, there are companies which are big enough that there is no other option but to
-->
use Kubernetes. However, there was this study from a few weeks ago that found that in general,
-->
the Kubernetes clusters are over-provisioned. So, people are using more nodes that are really
-->
needed. So, they are wasting money without the need of, I mean, they can be more
-->
adjusted to the requirements. So, that is something that has a lot of stuff to
-->
think about. Anyway, those are different, really good designs that you can work with.
-->
And what we are going to do is we are going to have a small coffee break for 10 minutes.
-->
And after that, we are going back just to wrap up and finish
-->
what today's session and the training. So, let's meet in 10 minutes.
-->
So,
-->
let's discuss about cost management or you can just go to the same cost management
-->
is like a hot topic to hold it again before the topic.
-->
So, let's discuss a little bit regarding cost management. So, there is some stuff that we
-->
need to be aware of regarding cost management. And first of all, I think I already told you
-->
about this, that the cloud is not for you to move into the cloud. It is to save money
-->
to be taken with a bit of style because it is not like you are going to save money.
-->
If you have a project and there are two options. Well, if you have a modernization
-->
for your, let's say, your hardware, your infrastructure. So, you will have two options.
-->
One option is to move into the cloud and the other option is to invest in
-->
the business infrastructure. So, in that case, probably going into the cloud will be cheaper
-->
if and only if the investment in infrastructure includes multiple data centers,
-->
multiple facilities to have higher availability, to have disaster recovery solutions and to have
-->
scalability and all that stuff that is already available in the cloud. So, that will be the real
-->
situation in which you can have a better
-->
infrastructure across the equation. Today, that can be achieved just with Kubernetes.
-->
We can create an infrastructure of Kubernetes and that will be the LSTC and high availability
-->
and all the stuff that the cloud offers. So, that is why Kubernetes is so cool.
-->
So, that is what you need to build into a cloud. But having a Kubernetes cluster is expensive,
-->
so just the design and all that stuff requires a lot of funds. So, that is something that
-->
you need to have in mind. Now, if you are into the cloud, there will be a lot of factors
-->
that affect the costs of having the cloud. Just the design that we are doing, the designs
-->
that we do in these two days have a strong component in costs. So, those designs that
-->
were using Kubernetes and were using our services and always on SQL servers and a lot of stuff
-->
that is available and that needs to be used, it turns out that when you run the calculator
-->
to those designs, you will hit a wall because it is super expensive and you won't get
-->
a solution. So, there are some factors that affect the cost.
-->
So, there is those types. So, these costs are very specific to each resource that you have.
-->
So, there are some metrics in each resource that shows you how the costs are calculated.
-->
The consumption. So, remember that by default is the pay-as-you-go model,
-->
which is the more flexible one, but also is the more expensive one. So, if you want to think
-->
in the long term to complement to one year, two years contract, that will give you easy
-->
50-60% of discount on the expenses in the cloud as well. The maintenance of the system.
-->
So, monitoring the system to identify where are gaps to be closed for different Kubernetes costs
-->
and stuff that you don't need. Information that can be forwarded. So, that is something
-->
that needs to be taken care of. Also, the geography, wherever. Depending on the area,
-->
you may find that the same resource is more expensive in a specific area compared to
-->
a different area. So, you may want to see, okay, what happens if we want everything in Australia
-->
or versus India. I don't know. So, depending on where are you, where you are located,
-->
you may find that the trade-off between costs and performance and closeness, the graphical
-->
closeness may be covered by the people. So, it is something that you need to take into account.
-->
Also, it's super important to remember that the network traffic and also the storage
-->
of data. So, we are discussing inbound data, but the cost of the outbound data,
-->
or moving data between resources, usually does increase the cost. Actually, when you are
-->
having backups in the cloud, if you don't apply an extraction strategy that minimizes the costs
-->
of recovering a backup from Azure, you may be paying a lot, a lot of money because
-->
as in the backups, you use cold storage. That cold storage, writing on it is super cheap,
-->
but reading and saving data from it is super expensive. So, there are some calculators that
-->
minimize that cost, but it implies that for storage, you will have to wait several
-->
hours, several days, or even weeks to recover the backup if you don't want to pay a super
-->
high bill. And we are talking about thousands of dollars. So, that is something that you
-->
need to take into account with the network traffic. And finally, the subscription,
-->
that probably is where the cost is taxed. Also, in the Azure Marketplace, you will always find
-->
different solutions that may be useful for anyone that you may have. So,
-->
this is always a nice place to find stuff that you might want. There is a tool,
-->
the price impact meter tool, that allows you to have an idea of how much it will cost,
-->
but this is not a precise science. Don't expect to create an estimation that will
-->
do the perfect for the stock price of Ruby, that will change always. So,
-->
you would only have a general idea of how much that will cost. So, we have here two labs to
-->
do. So, let me explain the other topic, and we can move on to those two labs. So,
-->
the first one is using the Azure pricing calculator. So, this calculator, there is
-->
this exercise. This exercise, one second, I will show you the webpage. It allows you to
-->
calculate and estimate how much that would cost to run a basic web application from Azure.
-->
So, that will be this exercise. The other involves another tool, which is the Total Cost
-->
of Ownership Calculator, which will help you to realize the costs of having the infrastructure
-->
on Azure versus having the same infrastructure on premises. So, this allows you to compare
-->
those two. So, these exercises create cost-comparison reports between both of them,
-->
and here is also this calculator. So, let me show that screen so we can see
-->
the generality of the exercises.
-->
Okay, so the first one is this.
-->
Oh, the link is wrong.
-->
So, I'm going to show this link because it's wrong.
-->
So, the link is elevated in the slides. So, I'm sharing the links here in the chat.
-->
So, the workload, the pricing calculator, and the TCR calculator.
-->
So, here is, there is the requirements, there is the architectural diagram. So,
-->
the idea is that with this information, go to the price calculator and create
-->
the specific architecture, all the components, and estimated costs
-->
for this with these categories that are here. So, we can see how much does it cost.
-->
This is, this is just to use these regions. So, you can play with those regions. And I
-->
will, I will suggest to you that use regions that are close to this here, geographically.
-->
So, you can compare how much does it cost to use the US data centers versus using the
-->
local data centers. Okay. And also, the other, the other calculator is the TCO one.
-->
So, in this case, you will compare to the conferences versus what you have in the cloud.
-->
So, let's, let's use our groups. But before sending you guys to the
-->
rooms, please open those rooms. Otherwise, you won't be able to see them in the lock.
-->
So, please open those rooms. And once those are open, I'm going to send you to the,
-->
to the, to the room. And we're going to have like 10 to 15 minutes
-->
stop for this exercise. So, we can get back together and see, have a quick overview of
-->
what are the resources of these populations and what about the training.
-->
Okay. So, I'm going to send you guys to the rooms.
-->
But, okay. So, no worries, no worries. Who wants to show us at least what you have
-->
advanced so far? We can complete it all together. So, go ahead and.
-->
Okay, Alvin, go ahead. Thank you. Yes.
-->
Oh, okay. So, okay. So, okay. You were at the, so you can see that this is sometimes
-->
the architecture design ends up just doing this kind of stuff. I mean,
-->
at the end, it degenerates in growing the diagram and creating an estimated
-->
parts. So, that's the real life, guys. So, okay. So, Azure Cosmos V. So, how is it
-->
you need to add? What else is missing there? So, there is a trick there. It turns out
-->
the adding the, the QNets service, what you're adding there, there is just the
-->
control plane. In order to have workloads on top of QNets, you need to add
-->
nodes. Okay. There are the nodes. This is a node. So, okay. So, you need to
-->
add one cluster there in cluster management. That needs to be one. And the nodes,
-->
I mean, you are going to have one node. But I'm not sure if it allows you to
-->
have one node. I think it's a minimum of three nodes. But maybe like that
-->
it goes. And, and, and calculate to see what, what is in there.
-->
So, okay. There are service. So, here you need to estimate how many
-->
per month are you having. So, so it can be, I mean, if you have 10, 100,
-->
1000 per day or 10,000 per day. So, that is how this works. So, usually this is
-->
per month. I mean, per person per month is nothing. So, probably the, I mean,
-->
in real life, it can be 10,000 per day easily. I mean, it depends on the
-->
application. But you can see, the first million are free. So, that is,
-->
that gives you like an idea of how this works. So, you will have the
-->
service for free if it's not requested more than one million times.
-->
Although I think that these are the first million. I mean, if you, I mean,
-->
if you consume 500,000 per month, starting the third month, you will have to pay.
-->
Okay. Did you really complete your estimation that you guys have done?
-->
I think we didn't have the time to discuss this part, but that is okay.
-->
And you have to see the databases.
-->
So, that is $1,000 in total. So, I mean, that is ridiculous compared to what
-->
would cost to create everything that is in there, as well as all the features.
-->
But obviously, there are some stuff that probably will need to be added here
-->
because you may want to have a secure SSL termination and the amount of
-->
the stuff that you're using. So, obviously, there is some stuff that
-->
needs to be fine-tuned in there, but it's okay. So, this is, that is part of
-->
the exercise to be able to translate a design and a feature into a spreadsheet
-->
with all the features that needs to be taken into account in order to have
-->
a working model, a working plan. So, that is good.
-->
This can be even exported to Excel or PDF, and you can save it into your account.
-->
And so, you can review it later. So, it's really useful there.
-->
Okay. Awesome. Thank you, Alvin. Is there anyone with the TCO calculator?
-->
So,
-->
we can just put a link into this TCO calculator to have an idea.
-->
So, this TCO calculator, it is able to define the infrastructure that we have,
-->
a course each, two courses per process hour. This is processes. So, two courses,
-->
two processes, and a course per each, and let's say 256 gigabytes. And that's
-->
easy work like Windows, Windows servers.
-->
And let's have Linux servers. So, Linux servers. There is a, let's say,
-->
a, other thing, servers, the same thread, two processes, the course, and we
-->
500 gigabytes and that's it. And now, database, let's see that we have
-->
a post SQL database, and running in Linux, I believe. And we have at least two, because
-->
I have availability to cause, two processes with a course, and because
-->
gigabytes, I have availability, and it will go to actual database code. And the storage,
-->
what we have here, and this is the storage, let's say, the sun.
-->
And we have, let's say, 20 terabytes. And the record is also 20 terabytes.
-->
And there is a file, let's say,
-->
and gigabytes, for example, how to calculate this, you need to have some data. So,
-->
let's say that it's outbound, no more than 20 terabytes, and let's do this.
-->
So, we can use this. And I'm sure something is missing here. Okay. So,
-->
which assumptions do you want to have if you are going to have some licensing, if you want
-->
your redundant storage, if you want to use built-in machines,
-->
and how much does it cost, the energy, the storage, the IP flavor, for the assumptions.
-->
And let's see in five years how this works.
-->
Okay. So, it didn't allow us to use this. So, let's see.
-->
Actually, it's saying that some services are not available in this region.
-->
I think it's doing this.
-->
Let's see if it works.
-->
Let's change this.
-->
So, that is our plot. We cannot see the result. Let me see if I have one here.
-->
Okay. So, this is just a quick example of how much it will cost on premises,
-->
and how much it will cost with issues. So, at the end, we need to decide, okay,
-->
this is about six, eight day of defense awarded, about nine thousand.
-->
So, this is like, this report is done. It works. So, it can be used like an idea.
-->
You can save it, you can download it, you can use that for your work.
-->
So, this is the TCO exercise that is in there.
-->
Okay. So, anyway, please, let me show my screen again because I need to show this.
-->
So, please remember, guys, that you can...
-->
But that you can copy this material, at least,
-->
if you haven't done it yet. Logging into the desktop machine and
-->
copying to the Gmail account or robots or whatever account you have.
-->
This folder, the materials folder, what is here, because this may be useful for you
-->
and also the slides will be useful for you because there are some additional materials
-->
that we were able to cover and the last that you may want to create to do as well.
-->
So, that will be like a good practice for you to have, guys. Okay. So, don't forget to do that.
-->
And no more to say for now, just thank you for your time
-->
for your dedication during these two days. I really hope that you have learned some new
-->
stuff and that you can use what we have discussed during these two days in your
-->
way today and you can grow from here in your future projects and challenges that you face.
-->
Okay. So, thank you, thank you so much and that's it. That's it for today.
-->
Thank you, guys. Thank you. My pleasure. Have a good one, guys. Bye-bye.