Serverless computing will be the next new normal!

Serverless - The Gold Standard of Cloud Computing 

The word 'Serverless' does not help make good grammar, but then when it comes to technology, unwritten rules hardly matter if the concept is understood, applied and of course monetized. Technically, however, the meaning 'Serverless' is again misunderstood - 'misnomer' as some would say - because it means the fewer server or as in the extreme cases, no server at all while actually, all it means is less management of the server. Now, if you are thinking 'Serverless' is suffering from an identity crisis; you are grossly mistaken. It is the next big thing - in fact, it already is - in cloud computing and revolutionizing the new frontier of technology by changing the way developers work on a project and how organizations view the roles and responsibilities of their IT departments. We have helped, Governemnt of India & One of the Big 4- Consulting firm to accelerate the adoption of serverless computing with Benefits such as decreased infrastructure costs and improved time to market are major factors driving this trend. Adoption of micro-service architecture patterns and DevOps are further accelerating the shift toward serverless technology!

Content of the blog

·      What is Serverless?

·      Benefits and scope

·      Case Briefs

·      Limitations and Conclusion  

What is Serverless Computing?

In order to understand Serverless computing, we first need to understand what it is not. Let's go back to the early days of the web. If you lived those challenging times, you might have loved the opportunities while at the same you would have wished for more flexibility. For a web application to be developed, you required to have an expensive undertaking: A full-fledged physical hardware to run an always-on server. The smooth functioning of the server’s computing resources meant fixed and recurring costs; no matter the amount of computing work actually performed by the server. And then the extra headache of operating systems, maintaining security patches, scaling configuration. In Cloud 1.0 you acquired the liberty to move storage and networking to the cloud, but you are still required to access, manage, monitor and perform core duties through VMs - short for virtual machines. 

Serverless computing makes things smarter - leaving off the complexity in the process and bringing in a new era of quick decision making, quicker deployment, and true scalability. So, again what Serverless computing truly is? Serverless is an architecture that allows you to run applications on an as-used basis; without having you to manage the server and/or the infrastructure. In other words, a server still remains somewhere. You just no longer need to buy, manage and maintain it. What you do is outsource all server management to a third party cloud provider; who takes care of client requests and responses, monitoring operations, scheduling tasks, capacity provisioning, etc. This means developers can stop worrying about the server and its resources and entirely focus on writing codes - which they do the best. The cloud provider begins and concludes a container platform as a service and bills you based on the usage. That's it.

Benefits and scope

Lower cloud costs: Pay-as-you-go. Don't pay if your application is not running. Developers pay for the time when an application executes a certain function

Lower development costs: Developers can rent as many resources as necessary for software development. So, no wastage on resources not actually required or utilized altogether 

Ability to invent more customer-facing features: Developers can solely focus on their core coding duties. Application deployment, workload intensity, resource distribution are headaches with the cloud provider. So, reduced development responsibilities means developer can invent more customer-facing features

Reduced time to market, reduced human resource investment, agile working environment are other aligned benefits of Serverless computing brings.

Case briefs 

One of the governing body for citizen registration in India, has planned to conduct the verification drive to deport the illegal immigrants from Assam to their respective countries, With 3.29 crores of applicants expected to do the e-verification , right sizing for the infrastructure is highly challenging & the capital investments required to build the physical DC & time frame to complete the verification was highly impossible.

moving to AWS Lambda was an extremely timely-taken decision that made its AWS Costs plummeted from a staggering $30,000 to just $3,000 in just three months, along with other benefits including On-demand scaling, flexibility to rapidly integrate various data collection and analysis tools, Pay-per-use billing model is generally more cost effective than traditional pay-to-provision models, Resilient to large traffic spikes with minimal risk of data loss.

Optimized Serverless solution was provided with help of AWS Lambda, Autoscale & Aurora for scalability , ease in migration of data of crores citizens using AWS Data Migration Services seamlessly .Leveraging the scalable solution provided the entire state population was able to do the e-verification using the portal in the stipulated period of time.

Limitations and Conclusion

While Serverless is the Swiss army knife for the cloud environment, it is not without limitations. Increased invocation latency, runtime limit, function complexity, vendor lock-in are some of the major challenges that need regular improvements.

But the benefits far outweigh the limitations. Working on the framework, developers learn and grow with a relatively new technology that ultimately makes software deployment and deployment more meaningful. New software companies are quickly experiencing the benefits, while many others who were relatively skeptical about this development are now testing it as an extension to their available microservices based applications.

The best thing about Serverless is you don't need to replace your entire technology stacks. You can start with a small component, and as it makes sense; involve more into it. To conclude don't hesitate experimentation. You have Amazon Web Services Lambda, Microsoft Azure Functions, and IBM Bluemix OpenWhisk, Google Cloud Functions as Function-as-a-Service vendors. Move non-critical tasks to a preferred cloud provider to test if it works. Most probably, it will.