TrendWatch: DevOps in 2021

Today’s swiftly changing business landscape means companies only stay relevant if they deliver higher-quality products faster than their competitors. One of the ways to achieve this goal is by adopting a solid DevOps strategy. Since DevOps bridges the gap between production and development teams, improving collaboration enables enterprises to better products in less time. 

Co-founder and CTO of tech recruitment start-up, Knack.io, Leroy Ware, says the demand for DevOps has not slowed one bit over the last year. “Despite the COVID-19 pandemic and the resulting economic slowdown,” says Ware, “65% of hiring managers are looking to hire more DevOps talent, up from 59% in 2018, and 70% of managers are looking to hire engineers with cloud expertise.” 

With the DevOps market size expected to reach $12.85 billion by 2025, looking at emerging trends can help guide your strategy. We spoke to a few folks in DevOps leadership - including our own engineers - to see what they thought about emerging DevOps trends. Here’s what they had to say.

DevSecOps 

DevSecOps introduces security protocols into every layer of code, operations, and best practices to make software development safer. This DevOps trend allows teams to deal with security threats when they arise instead of responding to them later on in the production cycle.

CPO of 42Crunch, Dmitry Sotnikov 

In today’s world of cloud-native applications, serverless and microservices architectures, every component may be exposing an API of its own and thus be a part of the overall attack surface. Thus, it is essential [that] your DevOps pipeline includes both static and dynamic security audits and testing of every such component and its API. This will turn your DevOps process into DevSecOps and enable your organization not only to ship quickly but also to maintain a high level of system security.‍

President of Flexagon, Dan Goerdt

2021 will see continued focus on security, extending comprehensively across areas such as role-based access controls, secrets, and credential management, MFA, single sign-on, static code analysis, API scanning, AI-based security automation, and beyond.‍

Virtasant Engineer Team

There is no clear consensus on the definition of DevSecOps. Many times, DevSecOps, is just security. An example of DevSecOps is if you were to do Compliance-as-a-Service. Automated scanning of infrastructure for vulnerabilities is clearer DevSecOps method. There is also a simpler version: shifting remaining security from the app development process, which is analogous to what we do with QA.

Building applications and infrastructure with security in mind is a more organic approach to DevSecOps. Instead of Dev and System admins battling it out and then adding security to that, bringing Security into DevOps allows all parts of the process to work together.

Microservice Architecture 

A survey of 200 tech leaders at large US companies shows that 84% of companies have adopted microservices to accelerate business innovation. Microservice architecture is all about speed and agility. Incidentally, this is also how DevOps operates. 

Compared to a monolithic system, the microservice architectural style breaks down the system architecture into bite-sized independently deployable services that function as independent units. Since the services are small and decoupled from each other, a change in one doesn’t affect the other. 

Viachaslau Shahoika, Group Manager at iTechArt 

In 2021, microservices have become a must for companies wanting to scale. It helps businesses run multiple work streams at a time, engaging bigger engineering teams and significantly reducing time to market.

‍Virtasant Engineer Team

Microservices are better for large products vs. start-up companies. The idea behind microservices is to have better fault isolation and better business agility with better support for the DevOps. Microservices can be very good, but great architects and sophisticated DevOps teams are needed not to miss bits and pieces in the process of creating a product.”
‍
Any monolith is always going to beat microservices in time to market. With well-defined service boundaries, microservices architecture is speedy to react to changes. This response implies a faster time to market with microservices architecture.

AI in DevOps

A Gartner survey notes that by 2023, 40% of DevOps teams will update their application and monitoring tools with AI. The volume of data generated today is too huge to be handled by traditional data analysis methods. ML and AI-driven approaches can help companies analyze massive data sets and build secure infrastructure and applications. With AI, DevOps teams can improve productivity. Further, by applying AI/ML monitoring solutions to observability data, companies can uncover threats and enhance application security.

Eran Kinsbruner, Chief Evangelist for Perfecto, Perforce Software 

In 2021, organizations will uplift their investments in DevOps and adopt more intelligent solutions to boost the developers’ and testers’ productivity. AI and ML within DevOps are going to be the top priority for executives looking to integrate AI-based solutions into the development life cycle from automated code reviews using observability and AIOps tools to minimize escaped defects, respond faster to outages, and predict production issues.

Virtasant Engineer Team 

The marriage of AI and ML does not necessarily create more innovative DevOps tools. Though artificial intelligence and machine learning can observe patterns and derive conclusions better than a human operator, it is still mostly limited to the scenarios on which it’s trained.‍

However, AI can still operate outside of what it’s trained to do. There are many cases where AI self-trains based on merely a stated goal and a scoring system. So we will continue to see an uptick in AI being used in security as we see an uptick in AI used in cyberattacks.

Serverless Computing

Serverless is a DevOps trend because of its pay-as-you-go model, where you pay only for the resources you use. This ease helps organizations look for fast, cost-effective ways to enter the market. Serverless technologies allow DevOps teams to perfect their product instead of wasting time troubleshooting and maintaining builds. Since, in serverless, the configurations and business logic are stored in the same repository, it makes rollbacks and deployments also quite simple.

Chris Blackden, DevOps Engineer at the Children’s Hospital of Philadelphia

You can make a serverless application with almost no limits on runtime or dependencies, persistent storage, and incredibly flexible scaling. Now you’ve got an incredibly powerful application that runs for a fraction of what an on-prem server would cost. And that is going to be extremely attractive to companies launching new products or refactoring old ones—especially companies who are maintaining their data-centers that have reduced on-site staffing because of the pandemic.

It’s great from a DevOps perspective because you can take something that has been traditionally tough to package and manage, like setting up a Kubernetes cluster and significantly simplify its operation and maintenance. I’m not saying that Kubernetes is going anywhere, especially for large applications. But for small applications or running a proof of concept, it can be overkill compared to these newer options.‍

Virtasant Engineer Team ‍

Serverless will gain in adoption because the cost profile is superior to traditional infrastructure. Serverless also saves time by not having to maintain the infrastructure and support multiple programming languages, which means it’s accessible to a lot of developers.

GitOps

Adopting Kubernetes and cloud-native technologies comes with a lot of benefits. However, they also come with a certain degree of complexity. GitOps simplifies the process and allows development teams to make deployments reliably and quickly. GitOps acts as an extension of IaC by using Git repo as a golden source for changes and automatically synchronizes changes to the live system.

An important feature of the GitOps model is declarative infrastructure and automating it using Git repositories. In the GitOps model, the git repository is widely considered the single source of truth. Updates are made to the repository through various “pull” requests and automatically put into production. Any divergence from this truth is highlighted and quickly restored to the original true state.

Principal Architect, Soumen Sarkar 

The reason GitOps is rising is due to Kubernetes, where code and configuration are sourced from Git. This is possible since Kubernetes is a next-generation infrastructure, and the modern trend is to have Git as the firewall where human contribution stops and automation takes over. GitOps succeeds with Kubernetes since this open-source project is friendly to git-based production operation flow where no changes are made directly—everything goes via Git.

Virtasant Engineer Team 

GitOps is undoubtedly the way to go. However, proceed with caution. But, this is not a one-size-fits-all solution and doesn’t always work with large organizations. Why? Because so much compliance is built-in, that the model gets reduced to regular CICD. The smaller the organization, the easier it is to implement. If you’re building infrastructure in 2021 and you use Kubernetes, GitOps is probably the de-facto solution for you.

The experts have weighed in to shed some light - not only on what we’ll see more of in DevOps but also on the value in these DevOps trends as it applies to your business. Adopting suitable DevOps methodologies helps companies become more agile, go to market faster, achieve more significant customer or client satisfaction and most importantly - continue to innovate.

‍