Azure Core Products

Core Products of Azure

Azure Compute 

On demand compute power which is provided by 

1. VM
2. Containers (more light weight than VM) /
3. Azure App service
4. Serverless Computing 

 (way to manage app without managing any infra)  Azure function  which allow to run small block of code.Azure logic apps which allow us to   configure workflow in cloud. and Azure event grid

Virtual Machine in Azure 

It comes under IaaS as you get full control, Downside is you need to manage including patching. 

When Creating VM

• Type of Image

There are many images available for windows (2012,2016,2019 etc) and Linux, Windows 10. There are other images also available with different software already installed. for example Windows with SFTP etc. Some other non MS image example are ArcGis Citrix net scaler. You can create your own image and upload it in Azure to use. 

• Size of Image(RAM processor)

While selecting VM there multiple option to choose from based on usage. For example filter by Size and Family options (General purpose, Memory Optimized, Compute Optimized). 

• Availability Options

No Redundancy required

Avilablity Zone

Availablity Set is a Grouping that you assigned to VM so that those VM are orginized under single DataCenter  to reduce potential for outage. 

If you want to create multiple VM and manage load you need to create a load balancer/VM Scale Set

Azure Batch is pool of VM used for High performance  computing like image video rendering , risk modelling, ETL operation or software test excecution.

Its even possible to move existing on premise VM to  Azure. (lift and shift type of Migration)

Azure site recovery 

Azure Migrate

So to create a VM, clik on Create a Resource go to category Compute or search the marketplace. 

With VM you have Disk, Virtual network and Storage account all associated. Moreover you manage auto shutdown, add new disc to VM configure back up and disaster recovery. You can use Boot diagnostic to investigate VM problem. 

Containers

dia Container is a way to wrap up server based application and services in it's isolated package. When a app is deployed using a container everything app needs to run successfully in included in container like runtime and dependencies so it's to move it around.

Containers emulate the underlying operating system rather than emulating the underlying hardware. Therefore they are smaller in size and easier to spin up as you are waiting for app to launch not OS.

As it's lighter in size you can have more containers so cost saving. 

Container is instance of container image. Image is ready only template with instruction on how to create a container. A container is runnable instance of a image. You can create new image by leveraging on existing image. 

Container Registry store the container images. Docker hub is the container registry on web.  Docker is runtime for containers  which is open source project. 

Local/On Premises Server: Install docker and prepare a container with all dependencies. 

Deploy container on own VM in Azure but you need to patch those VM. But if want environment to deploy container without need to maintain or patch the environment then you can use Azure Container Instances services. Its intended for small application so limited availability and scalability. 

AKS 

For complex architecture or higher control you can use Azure Kubernetes Services. (AKS) It's container management system. It's easy to scale out and you can monitor and deploy containers. 

Pods are group of one or containers with shared resources, Kubernetis run your pods on Nodes (VM). If pods crashes AKS can create new instance. You can leverage VM scale set , and Azure monitor service. So AKS uses other Azure services to provide a robust way of hosting your container based solution. 

Demo 

Create a Resource/Container Instances by selecting resource group , container name, region, container image.

 App Service 

App service is like a IaaS where you can deploy you applicaiton. It's like traditional wehosting where framework are already installed. 

When you create a new app service. If you are running windows services on you server so you can do that using WebJobs. 

You need to first create a App Service plan when creating App service. When creating app service plan you need to define size of VM Plan type is Pricing Tier. You can have more than one App service in a plan. 

Generally you have limited access on those VM, as you don't have to maintain or patch those VM. 

You have all the metrics information available. 

is about allowing Devs to focus on business logic and code rather than underlying infrastructure. 

Through App service and service for containers, we create abstraction of infrastructure. These services can be called Server less Computing. 

Azure functions are small peace of code triggered by Events/trigger. (C#)

Through Azure Logic apps you can write workflows right in Azure portal) You can call AFunc from Logic apps. You can connect to various end point like share point. 

Even Grid

Even grid connect data sources and even handles. This lets you create subscription to event and create automation. 

Example related to email with Azure function and logical App. With in logical app workflow we call Azure function to clean the email html contents and save email content and it's attachment to a Blob.

Core Networking products in Azure

 There are number of network products in Azure to create secure network for your VM and other Azure resources so those other resources. Though underlying network is managed by Microsoft but there are many network configuration option available on Azure portal. 

VM in different subnet can communicate but VM in different virtual network can only communicate through VNET peering in same or different region.

In order to VM to communicate to outer world VM needs to be assigned public IP. Public IP is a separate resource in Azure with it's own  config setting. And then this IP address is assigned to VM. 

We can have load balancer when there is incoming traffic from internet. You can have LoadBalancer for incoming traffic or even for internal private network as well for example in case of n Tier application. LB does not interact with incoming request but just redirect it to VM after checking it's health. 

LB can translate public IP address to public Ip adress and it's high performance solution that can handle lots of traffic but it's just load balancing and port forwarding engine. 

In case you need more control over incoming traffic use Application Gateway as it does much more than load balancer. 

In case you have on premises network connecting to Azure VM network you can have VPN gateway. This is like hybrid cloud. Through VPN gateway both network communicate in encrypted way securely over internet.

Another way for on Premise network to connect is

In Demo we looked at Vnet where we have ip address range. In subnet you distribute those ip address to different sub network. 

In case you want to make the VM as DNS server just give the IP address. Use peering to enable the resources to communicate.

Then in network security group we specify if rdp to vm is allowed from internet. 

For application gateway there is public ip address.  

Content Delivery Network (CDN)

CDN is a distributed network of servers which store the static data like CSS, pdf files, client site script icons, to minimize the latency. With this we also offload the traffic from source server

Using Dynamic site acceleration we can accelerate the dynamic data as well.   

CDN server can connect to various sources to cache the data like App service web app/blob storage. It also optimizes streaming of certain media files. In fact it connect to any publicly available server to cache the data.

CDN server are also called Edge server.  CDN servers are part of POP i.e. Point of presence. POP is part of CDN region. For example Asia pacific. Here you will have a CDC Region data center and in of the POP you will have CDN server which perhaps is holding your cached data. To increase coverage Microsoft partner with other vendors like AT&T.

Above is diagram of how it works, first request goes to Edge server if not available request goes to original server and gets it into Edge server from where it is served. It remains there on server for any further request for 7 days but can be deleted from Edge server in case new file is available so to make sure that new files are delivered. 

CDN profile consist of different cdn endpoints and there is pricing based on cdn service used. 

Dynamic Site Acceleration 

CDN mostly deals with Static data but even dynamic data can also be cached. This data is delivered based on user behavior. DSA is a optimization option in which data delivery is optimized to reduce latency. 

Below are few approaches for DSA.

1) Route Optimization : As in world of internet there is always multiple ways to deliver data. Fastest Route from Edge server to Original server is calculated by doing health check.

2) TCP optimization : Avoid TCP slow connection i.e. check bandwidth for high data packet traffic. 

3) Object Prefetch: We have HTML and related contents (CSS, image icon, scripts) are delivered together with html. In general they first html is delivered and then resources are requested. 

4) Adaptive image compression: In case of slow connection smaller files are delivered. Good for mobiles.