DevOps: The Journey So Far and What Lies Ahead!

If you have been in the IT industry for over 10 years, I am sure you have seen the evolution and massive transformation DevOps brought in as organizations continue to shift from optimizing for cost to optimizing for speed, and that shift is exponentially increasing its pace to adopt DevOps. Today, when I see life before and after DevOps, I can easily see that some terminologies has been changed in our daily life primarily because of DevOps adoption.

a) Manual => Automated

b) Physical Datacenter => Virtual Private Cloud

c) Outages => High Availability/Zero downtime

d) Enterprise/Web archives => Containers

and the list goes on and on…

DevOps, which started as a buzzword, is now becoming a standard for every organization in order to meet the demands of time to market and release better products to stay ahead of the competition, and that’s the reason big names like Google, Netflix, Amazon, and Facebook are heavily investing in it and have experienced the value coming out of it.

So, what exactly has DevOps changed?

No More Working in Silos

DevOps has improved the software development culture and mindset. Welcoming new changes, blameless culture, transparency, accountability, embracing failures, right collaborations, and communication between different teams are some of the keys organizations have enabled to successfully unlock DevOps culture.

Time to Market

DevOps enables organizations to develop and deploy software faster and more efficiently enabled by an end-to-end automated and integrated process using CI/CD pipelines. Continuous Delivery allows developers to continuously roll out tested code that is always in a production-ready state and can be released to production based on business approval. As soon as a new feature or story is complete, the code is immediately available for deployment to test environment, UAT, or production.

DevOps-as-Code

Over the last few years, there has been a tremendous development on how automation has been done. Pipeline-as-Code to automate CI/CD pipelines, Config-as-Code to manage configurations and orchestration tasks, Infrastructure-as-Code to automate environment provisioning are gaining momentum. Languages like Groovy and Python where the core OOPS concepts are required are being used for most of the automation. Also, unit test cases are written for every automation to validate their code similar to how application development tests their code. This has enabled and encourages many application developers to understand DevOps workflows holistically, gain expertise, and contribute, which was earlier mere a black box for them.

Software Killed Hardware

Gone were the days when sysadmins were heavily involved in receiving new hardware, and then setting and configuring new servers with each server having a custom configuration. Today the servers, network, firewalls, load balancers, and everything else is virtual, living somewhere at Amazon or Google or Microsoft. Today, you write software to provision, manage and decommission infrastructure. Upgrading resources to a new server, adding identical servers, securing infrastructure are all driven by software.

Containers and Microservices to Maximize Deployment Velocity

Microservices have enabled developers to have the freedom to make changes to one service, create a Docker image, and deploy it independently without impacting other services in the system. If there is an issue in any service, it can easily be isolated to one single service so that a fast rollback can be made easily. This speed of deployment with minimal risk is the primary reason why organizations like Netflix and Amazon have adopted microservices-based architectures, ensuring they are eliminating as many bottlenecks as possible to release the application to end users. Platform-as-a-Service tools like Amazon ECS, Google Kubernetes, and Redhat OpenShift have helped enterprises to adapt to microservices architecture and migrating their existing applications to microservices and dockerized in production.

Cloud Encourages the Birth of Many Startup Businesses

Cloud computing provides an added advantage to the start-up business. Businesses previously required heavy time and money for housing, powering, and cooling infrastructure. With cloud, there are limited upfront capital costs as it provides the ability to match revenue with expenses since you pay only for the resources you use. For a festive season or other cases of peak traffic, you can easily scale up and down your infrastructure.

DevOps Predictions

With technology evolving at a rapid pace, DevOps will continue to gain momentum and break barriers. Here are high-level predictions what lies ahead in DevOps world.

Cloud Migration

Cloud adoption will continue to evolve. New startup businesses are already adopting cloud for hosting their applications. There has been a substantial increase in the big enterprise giants who migrated their physical data center to the cloud and this trend will continue to gain momentum for their survival.

Continuous Deployment Is So Close, Yet So Far

While continuous integration and continuous delivery practices have enabled organizations to release new features to the market in the most efficient way, there are hardly any buyers who want to adopt continuous deployment. Product-based companies like Amazon and Netflix are making frequent strides with continuous deployment, but financial firms are still focused on having a robust application and infrastructure with performance and security being their primary areas, with the desire to release features to market using a manual trigger.

Serverless Computing

Renowned training company A Cloud Guru is running their application on serverless architecture. There are no infrastructure costs they have to pay, as they pay based on the number of visits they get to their course content. This allows them to offer their courses cheaply, which also provides them an added competitive advantage as compared to their competitors. AWS Lambda, Google Cloud Functions, Azure Functions are all examples of Function-as-a-Service platforms that support serverless architecture. The only concerns many IT leaders shared are their fears over vendor lock-in. Choosing a cross-vendor programming language and adopting standardized services over fully managed services provided OOTB by cloud provider are two ways to avoid vendor lock-in fear when adopting serverless architecture.

DevSecOps

Security is going to play a major role as more and more applications will be migrated to cloud. Whether it’s an application, VM, container or an entire network, one needs to understand the entire process and lifecycle and ensure there is no corner left for any vulnerability for an outsider to hack into your application or your system and this can be achieved only when you integrate all flavors of security testing in your DevOps process.

SRE for Service Management

In general, an SRE team’s responsibility is to ensure the service is available all the time, and that application health and monitoring and emergency response that has been done by Ops team. However, this is changing and organizations are looking for engineers who can code as well and take care of ops. For example, Google has put a cap of 50% on the overall ops work for all SREs and in the remaining 50% of the time SREs are actually doing development. They found this model has many advantages as SREs are directly modifying code, building and supporting the system, and bridging the gap between the product development teams and SREs during cross training of new features release.

Cognitive DevOps

Cognitive DevOps will excel in developing an automated system that should be capable of resolving problems and providing solutions without any human intervention. It uses machine learning algorithms that will help deal with the real-time challenges faced in DevOps by gathering and analyzing data across different environments which will eventually lead to smooth and error-free releases. IT operations analytics, network performance analytics, security analytics, application performance management, digital performance management, and algorithmic IT operations are some of the key areas vendors are targeting to implement cognitive operations in the journey to move from DevOps to NoOps.

I would like to summarize this blog by stating that technology is changing at a very rapid pace and DevOps is fueling the demand with its workflows, tools, and practices. With so much already achieved in the last decade and so much to achieve in years to come, the DevOps journey ahead will be exciting and full of surprises.