Learning Azure Cosmos DB

Azure provides platforms to clients to develop and deploy software without worrying about the server infrastructure and configuration. You can focus on the development of the application and leave the infrastructure worries to Azure.

One of the key perks of using Azure is the overall cost, which is low as the resources are allocated on demand and servers are automatically updated. Azure provides automatic updates to avoid security issues, and it does all that in the background without affecting the application.

Azure provides a managed compute service that gives complete control of the operating systems and the application platform stack to the application developers. It provides developers with control over the data center and allows them to manage it.

This is the ideal case for projects where complete control is required. The virtual machine can be completely altered to the requirements of the project or business. However, if not managed carefully, it could become troublesome; all the security updates then become the job of the developer who manages the infrastructure.

The maintenance of the legacy application also sometimes becomes difficult due to operating system and application stack updates.

As the name suggests, the job of this module is to provide a set of tools and services to do computing. Here is the list of services which are provided under compute modules:

As seen in the preceding screenshot, these components mainly consist of virtual machines and containers, but as per the requirements you can join them together as well.

For example, a virtual machine with more disks and snapshots along with containers is used for continuous delivery.

Let's look over the next component, the data management modules.

Data management modules consist of SQL Server, Azure Cosmos DB, Redis, PostgreSQL, and data warehouses. You can deploy these data management modules and use them in your app.

Here is the complete list of data management modules:

We are going to cover Azure Cosmos DB in an upcoming section. Microsoft Azure started to support MySQL and PostgreSQL in the current version (only at the time of writing this book).

Let's look over the networking module.

The Azure traffic manager is in charge of directing the request of a user cleverly to an accessible data center. The procedure includes finding the closest data center to the user who makes the demand for the web application, and if the closest data center is not accessible because of different reasons, the traffic manager deviates the demand to another data center.

Be that as it may, rules are set by the administrator of the application regarding how the traffic manager should act in such cases. The virtual network permits a system between nearby machines at your commencing place and a virtual machine in Azure's data center. IPs to virtual machines can be appointed for identity and accessibility purposes. Virtual network setup requires a virtual private network device.

Azure also provides an application gateway which is very useful in micro services application development.

Following is the list of networking modules Azure supports at the time of writing this book:

Let's look over the big data and business intelligence modules.

The extensive measure of data can be stored and retrieved using Windows Azure. Azure offers HDInsight, which is a Hadoop-based service to determine and sense out enormous information stored as big data. Companies regularly need to oversee extensive measures of information which are not fundamentally relational in nature. Hadoop is used by many organizations nowadays, and it is used in Azure too.

In this manner, Azure offers Hadoop benefits on their platform for their customers. The term big compute alludes to high-performing calculations. This is accomplished by executing code in parallel on many machines in the meantime.

The components offered by big data modules are shown in following screenshot:

As you can see, Azure offers big data services ranging from storage using Hadoop clusters to performing analytics over big data, performing log analytics to determine errors, and generating reports based on your data.

Azure also provides machine learning web services to be used as smart algorithms to train and derive sense out of data.

Let's look over the caching and messaging modules.

The Windows Azure caching service helps you to build a high-end, scalable, and dedicated cache for your website.

Azure cache is offered in three different types:

Let's look over the messaging module.

Windows Azure offers two options for handling the interactions between two apps:

Generally, service-oriented architecture or SOA-based applications use service buses to perform the communication between various services.

Distributed and independent services, such as micro services, use message queues to perform the communication between them. The developer can choose one according to their needs and requirements.

Let's look over the identity module.

This component is all about the administration of users, validation, and authentication/authorization. The active directory stores the data of clients getting to the application and, furthermore, the association's data within the organization. It can synchronize with the related data on nearby machines dwelling on-premises for effective sessions.

Here are the security and identity modules provided by Azure:

Microsoft Active Directory is one of the most mature and popular authentication and authorization modules on the market.

Let's look over the mobile service module.

Windows Azure offers a simple platform to create portable mobile applications. You can begin the development of your mobile app using the Azure mobile SDK. There are so many libraries already available in Azure SDK such as notification, authentication, and so on. You can focus on your business logic and Azure SDK will help you in the common routines to ship your app faster.

Windows Azure provides a very easy platform to develop mobile applications. You can simply start using mobile development tools after logging in to your account. You don't have to write big custom code for the mobile application if you use this service. The push notifications can be sent, data can be stored, and users can be authenticated in less time.

Please refer to the following screenshot for the full list of mobile service modules:

As you can see, the list of modules is enough for mobile app developers to bootstrap their application quickly and easily.

Let's look over the last, and one of my favorite, modules: the internet of things (IoT)!

IoT is a hot subject right now and Azure is one of the first providers in the cloud market to facilitate such services. IoT modules allow developers to deploy services such as IoT device communication, networking, and device allocation to the hub.

Refer to the following list of modules provided by Azure for IoT purposes:

There are other modules, along with add-ons such as developer tools, monitoring tools, Bing Maps application programming interfaces (APIs), and so on.

I hope this provides you with brief but useful insight into Azure. Let's dive into Azure Cosmos DB now.

Azure Cosmos DB is baked with scaling and distribution. Mobility and cloud are the current major sources of computing use and that creates the challenge of high availability with less latency.

Azure Cosmos DB provides an easy way to build and distribute your application across an Azure data center without doing manual work and configuration. Azure automatically replicates your data across a number of data centers available in the Azure network.

We will be covering turnkey distribution in the next chapter in detail.

Let's look over multilevel data models.

Azure Cosmos DB supports multiple data models, including document stores, key values, graphs, and tables. Azure offers consistency and high availability across all data models, regardless of their nature, to store the data.

At the time of writing this book, Azure supports APIs for the following models:

The Microsoft team is working to support other data models, if required, and to enhance the existing ones.

We will study the data models in detail in Chapter 3, Cosmos DB Multi-model Databases.

Let's look over the scaling feature of Cosmos DB.

Azure Cosmos DB introduces a global turnkey feature; using this, you can deploy and scale your application with literally the click of a button. Azure handles every key technical detail such as infrastructure management, replication, traffic diversion, and so on.

Almost every database as a service offers this scaling, so what's new? Azure offers a software level agreement for latency and deployment time, and if it takes more time than promised, you will get your money back!

Let's look over the latency responsiveness of Cosmos DB.

According to the Azure documentation:

This is a huge promise. Cosmos DB offers end-to-end latency reading under 10 ms for 1 KB of payload, and write under 15 ms in the same region of Azure. This is very impressive and helpful.

With this high response rate and low latency, let's see how Cosmos DB manages consistency.

Most of the distributed databases offer consistency models, mainly strong and eventual consistency. Azure offers three more consistency levels. They are:

Let's look over the main promise of Azure Cosmos DB, which is its money-back guarantee.

Azure Cosmos DB promises that your data gets there fast, or they send your money back. Azure Cosmos DB provides service level agreements for availability, latency, throughput, and consistency.

An SLA, as of now, is offered only by Microsoft. You can refer to the complete documentation here: https://azure.microsoft.com/en-us/support/legal/sla/cosmos-db/v1_0/.

Let's look over the few use cases of Azure Cosmos DB.

IoT consists of various devices that emit data every fraction of a second. This places a need for high ingestion from multiple sources. Azure IoT hub can be used to receive the data from various IoT devices.

Once the data is ingested, you can use Azure HDInsight to process and derive real-time insight. You can then store this data in multiple sources, such as tables, to view in SaaS-based applications, DocumentDB to process the data for graphs and processing, and so on.

Let's look over the next use case.

Microsoft uses Azure Cosmos DB in the Microsoft e-commerce platforms Windows Store and Xbox Live. Azure Cosmos DB can be used to store product data, search the products, and perform the session, which is very important for e-commerce applications.

Azure Cosmos DB can also be used for event sourcing using its own change feed functionality. Change feed provides the amazing event generation mechanism to handle micro service communication.

You can refer to the following diagram about micro service communication:

Change feed is a concept introduced by RethinkDB; Packt authored a book on this called Mastering RethinkDB.

Let's look at the next use case, which is my favorite one, gaming.

Games are highly mission critical and should be very much available at all times. That requires fast and highly available databases. Azure has already proved its worth in games such as Halo 5, The Walking Dead, and so on. Azure Cosmos DB performance can be used to handle millions of requests and to show real-time data feeds and social engagement.

Azure Cosmos DB provides an API that can be directly used in games, and with the help of database functions such as Hadoop, could be used for real-time analysis and feeds.

Let's look over the web and mobile use case.

Azure provides various sets of tools such as key value stores, traffic manager, gateways, and of course Cosmos DB. These can be used in building social media applications, news feeds, and mobile applications using various frameworks.

Azure machine learning and Hadoop can be used as well to generate personalization reports, feeds, and recommendations, which is an amazing part of any modern application. Since most of the data is stored in JSON, it can be communicated and parsed across various scenarios such as the frontend, mobile, backend, and database.