Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News Cheryl Hung on Trends in Cloud Native and DevOps for 2021

Cheryl Hung on Trends in Cloud Native and DevOps for 2021

This item in japanese

In a recent keynote for The DEVOPS Conference, Cheryl Hung, VP ecosystem for the Cloud Native Computing Foundation (CNCF), shared her top 10 predictions for cloud native in the upcoming year. This includes improvements in cross cloud support, growth in GitOps and chaos engineering practices, and an increase in the adoption of FinOps.

The 10 predictions for cloud native in 2021 from Hung's keynote

The 10 predictions for cloud native in 2021 from Hung's keynote (source: Twitter)


Hung notes that 83% of respondents in the 2020 CNCF survey are running Kubernetes within production. This heavy usage of Kubernetes is mirrored in the catalog of CNCF projects. For example, Flux is a tool to enable the GitOps methodology with Kubernetes. This approach focuses on declaratively describing the entire desired state within git. As Hung notes, with GitOps "you get a perfect audit log through your entire infrastructure and it's easy for you to roll back in case there are any problems."

Two recent CNCF projects, Litmus and Chaos Mesh, focus on enabling chaos engineering practices within Kubernetes environments. Hung explains that "I think [chaos engineering] is actually a very sensible way to handle infrastructure problems and I'm a little bit surprised that it's not already more widespread."

FinOps is a recent movement attempting to bring financial accountability to provide more transparency on cloud spending. Hung describes it as "a group of practitioners who are trying to find the best practices and tooling to really understand and optimize where cloud is being used and how to reduce that from an organization's point of view."

InfoQ sat down with Hung to discuss these predictions and trends in more detail.

InfoQ: Why do you see GitOps growing significantly over the next year?

Cheryl Hung:

  1. In June 2020 CNCF End User Technology Radar, Flux was placed into Adopt by the CNCF End User Community, indicating a high level of confidence and maturity even though it was only a sandbox project at the time.
  2. In November 2020 the GitOps Working Group was founded by Amazon, Codefresh, GitHub, Microsoft, and Weaveworks.

GitOps has reached a point of community consensus and the benefits are clear, which is why I foresee growth.

InfoQ: In a tweet you mention that you didn't include observability in the list as you see it as "already mature". Similar to observability, chaos engineering has a strong ecosystem of tooling and practices already available. What changes do you see in the upcoming year around chaos engineering and how organizations are practicing it?

Hung: Observability is a far earlier need than chaos engineering - all cloud native deployment requires observability, whereas only a proportion will get to the level of complexity where chaos engineering practices become worthwhile. Litmus and Chaos Mesh joined CNCF in July 2020 and November 2020 respectively and there's a list of adopters on and

InfoQ: Multi-cloud is a topic that tends to be divisive. Many argue, such as Corey Quinn, chief cloud economist with The Duckbill Group, that it is "the worst practice to be avoided by default". What do you see as the growth drivers behind it "becoming more real" in 2021?

Hung: This is tied to storage and data. Running stateless applications in a multi-cloud environment is becoming more feasible (and despite Corey Quinn, does have benefits in certain scenarios) but extracting data across clouds is still extremely difficult and expensive. I see storage vendors focusing on enabling multi-cloud, but I don't know whether this potential will be realised hence "becoming more real" rather than "becoming real".

InfoQ: Regarding your prediction of "end user driven open source" you've been quoted as saying "end users are not passive, they're becoming leaders in open source projects". What key changes do you see being driven in cloud native technology by these new active leaders?

Hung: Argo, Backstage, Envoy, and Jaeger are examples of open source projects donated by end users (Intuit, Spotify, Lyft, and Jaeger respectively). They cover the range of cloud native topics but it's more an ecosystem change that I see - instead of end users relying on vendors such as Red Hat to implement upstream changes, end users realise the value of donating their own projects and building their own communities.

This is also in line with CNCF End User Technology Radar, which enables end users to collectively express an opinion on cloud native technologies. This forces vendors and project maintainers to consider how end users view the choices available to them when choosing a tool, and why they would consider one product over another.

InfoQ: What current trends do you see falling by the wayside over the next year?

Hung: Serverless hasn't taken off to the same extent as containers. It will continue to be used but for specific niche cases, but not as a broad development paradigm.

Rate this Article