DevOps has come a long way, and it will no doubt continue to shine this year. As many companies look for best practices around their digital transformation, it’s important to understand where leaders believe the industry is headed. With that in mind, the following post is a collection of DevOps senior responses to 2021 DevOps trends.
Let’s see what each of them has to say about DevOps over the next year.
1. Migrating to microservices becomes mandatory.
“The shift from single services to microservices and container architectures is a must for all companies to transform digitally. It is no longer one or more options. “Kubernetes will be used more and more, and Terraform will be the automation infrastructure of choice when organizations adopt cloudy conditions.” — Sachidananda Pattnaik, Wipro DevOps Lead Engineer
2. Mixed mode will become the deployment specification.
“2020 accelerates remote working, accelerates the migration to the cloud, and transforms DevOps from a best practice to a critical part of every business. As we move into 2021, the industry will adopt hybrids in multiple ways. First, companies will take full advantage of a mixed workforce that combines the benefits of remote work and on-site team collaboration. Second, business models will become hybrid, such as meetings that combine virtual scale with local networks. Finally, hybrid will become the deployment standard as companies modernize their stacks to take advantage of cloud-native technology, but realize that not everything can be divorced from PREM. The winners in 2021 will be the companies that adopt hybrids in their businesses, models and products.” — Stephen Chin, VP of Developer Relations at Frog Technologies
3.DataOps will thrive.
“DataOps will definitely surge in 2021 due to the spread of the epidemic from Novel Coronavirus. Consumption of digital content is rising sharply due to COVID-19 and the work-at-home situation, which requires a new level of automation for automated scaling and self-healing systems to meet growth and demand. So far, the enterprise set up the system logging, monitoring, and alarm (ELK/EFK sports, Prometheus/Grafana/Alertmanager), when the conversation is to speed up the pace, use of the available data and indicators to generate valuable insights, Learning and applying machine learning models to predict accidents or outages, developing automation to learn to improve budget planning and forecasting capabilities from data. Many people have started calling for MLOps/AIOps to address the problem.” — Nirav Chotai, Senior DevOps engineer at Rakuten
4. Resilience testing will become mainstream.
“In my opinion, the intersection of observability, performance testing, and resilience testing is going to become mainstream. With the recent Ops problems of leaders like AWS and Google, and the acceleration of digital transformation across all verticals, the market will gradually realize that the unlimited scalability offered by public or private clouds is not enough.” — Patrick Wolf, product director, Neotys
5.GitOps will become the norm.
“A ‘you build it, you own it’ development process requires tools that developers know and understand. GitOps is the name of how DevOps uses developer tools to drive operations. GitOps is a way to achieve continuous delivery. More specifically, it is an operational model for building cloud native applications that are deployed, monitored, and managed on a unified basis. It works by using Git as a real source for declarative infrastructure and applications. When a commit is pushed and approved in Git, the automated CI/CD pipeline will make changes to your infrastructure. It also uses the Diff tool to compare the actual production state to the source-managed state and alerts you if there is a divergence. The ultimate goal of GitOps is to speed up development so that your team can safely make changes and updates to complex applications running in Kubernetes.” Soumen Sarkar, Chief architect of Macy’s
6. There will be more migration to serverless.
“2021 will be a year of more serverless migration. If the container and orchestration are Generation Z. The live load on no server will be generation Z+. Pay-per-use will only be paid when you use the model. Per-use payments will be made only when you use the model. Pay-per-view and pay-as-you-use may look the same. But consider running a k8S Pod based microservice to run the same service on no server if needed.” — Shivaramakrishnan G., Site reliability engineering Manager at ADP Lifion
7. NoOps appear.
“I’d like to see more hosting and reduce our DevOps operations and reduce our customers’ operating expenses. More serverless applications, more serverless services, such as Aurora serverless, Fargate, Amazon S3, and serverless static websites. Amazon ECS/EKS (new re: Invent2020) and Cloud Management services in your data center allow you to reduce maintenance and development of your data center. Likewise, porting more cloud-native principles and capabilities to the data center.” — Alfonso Valdes, CEO of ClickIT Smart Technologies
8.BizDevOps will shine
“As organizations capture value from DevOps, there is a move toward cost optimization in terms of architecture and corporate hierarchies.
Focus on flexible, cloud-native, architecture and tools that were once reserved for “big companies” and packaged for small companies (Snowflake or Hazelcast vs. Oracle/Teradata).
FaaS is just getting started (serverless, Lambda, etc.) — operational issues are being solved and people are realizing its potential.” — Chad Prey, DevOps manager at Petco
9. Infrastructure as code (IaC) will have a higher status
“Infrastructure as Code (IaC) : The core principle of DevOps for cloud environments. Your infrastructure, that is, servers, networks, and storage devices, whether locally or in a cloud defined as code. This allows companies to automate and simplify their infrastructure. IaC also provides a simple infrastructure version control system that allows teams to roll back to the “last run configuration” in the event of a catastrophic failure. This means quick recovery and reduced downtime.” — Niraj Tripathi, senior solution architect, Volvo Cars
10. Automation and chaos engineering become very important
“Everything is automated — build, deploy, test, infrastructure and release.
Single line production with required quality gates. Faster, repeatable, customizable, and reliable automation is the key to the success of any project. Chaos engineering — a critical aspect in today’s hybrid infrastructure world. System behavior and customer experience are tightly coupled, and the earlier you test, the better the customer experience.” Nikhil Bhandari, Group Development Manager, Indian Development Centre, Gibraltar
11. Cloud-native computing approaches are becoming more standardized.
As the cloud space has become so advanced (in the last 10 years or so), containerization has become the norm and everything is so standardized, almost as it was in the mainframe era.
Of course, there will be trends and money-making opportunities. But I don’t think the next big disruptor is going to be anything. Everything is basically the same as the best practices of five years ago, just a little more reliable. I think more and more people will continue to switch from Pets to Cattle, and tools like Ansible and Puppet will be left to Packer and Cloud Init to build container hosts.
In my opinion, this is the golden age of software development. DevOps and the on-premises cloud approach have achieved a lot. Pipes, hosts, storage, load balancing… All these problems have recently been solved in five minutes. – Ben Sapp
12. Security will be a high priority.
Track uncontrolled changes in your infrastructure from a DevSecOps perspective. It’s great as an infrastructure for code, but there are too many moving parts: the code base, the status files, the actual cloud state. Things tend to drift. These changes can occur for a variety of reasons: from developers creating or updating infrastructure through the Web console without telling anyone, to uncontrolled updates on the cloud provider side. Dealing with the differences between the infrastructure and the code base can be a challenge. – CloudSkiff
Chaos engineering will become more and more important.
“In more organizations, chaos engineering will become an increasingly important (and common) consideration in DevOps planning discussions. Chaos engineering is the discipline of conducting experiments on production software systems in order to build confidence in the system’s ability to withstand turbulent and unexpected conditions.
If we consider DevOps within the framework of a traditional level 5 maturity model, then Chaos engineering would be at levels 4 or 5, which is included in the DevOps practice umbrella. Just as the traditional role of an independent test/quality assurance team is incorporated into the discipline of DevOps, so should Chaos engineering.” – Kelvin Meeks, information architect for an American technology company
14. Pay more attention to instant logs to quickly verify success or failure.
“Use logs in late deployment to verify that the publication was successful or that there were serious errors. The biggest connection people need to make is to define the manual process and then take the big leap to automation. In one-click deployment, instant logs can quickly verify success or failure, and then trigger rollback. With that comes complexity and whether something can be rolled back across service dependencies, or whether additional services need to be further tested. Imagine 100 microservices (that is, pipes, or even another 100 containers). As a project, I always celebrate a successful rollback because it has no impact on service and is successful.” Craig Schultz, Director of Platform Stability at ADESA
DevSecOps will become the default part of DevOps.
“The ‘Sec’ part of DevSecOps will increasingly become an integral part of the software development lifecycle. A truly safe ‘left shift’ will be the new normal. There are fewer dedicated safety steps in CI/CD pipelines, and automatic safety awareness and operation will be part of all pipeline steps. Start with the developer’s IDE and move on to dependencies and static code analysis. Without the appropriate software components will not be released to mediate these problems. Customers will get truly secure free software.” — Eldad Assis, DevOps Architect, Frog Technologies