The New Stack Context: Kubernetes 2020, by the Numbers
The New Stack has just released an updated eBook on Kubernetes, “The State of the Kubernetes Ecosystem,” and so this week on The New Stack Context podcast, we’ve invited TNS analyst Lawrence Hecht to discuss some of the analysis he did for this volume. We covered Kubernetes adoption in the cloud, storage and networking concerns and the changing DevOps culture around cloud native computing. At the end of the podcast, we also discuss what to expect from next week’s KubeCon + CloudNativeCon Europe virtual conference.
This 80-page ebook, a complete revision of our 2017 version, also explains how Kubernetes is the underlying architecture for enterprise data centers, cloud services and a hybrid approach — as well as at the edge. In this book, Hecht’s findings provide many glimpses into the community, including:
- The most used cloud native storage solutions are from cloud providers. Almost 50% of respondents said that they use AWS, and nearly another 20% use Google and Azure each. Open source software solutions like Rook or Pure Storage are used by far fewer.
- Customers of traditional storage companies were significantly more likely to complain of storage challenges. For example, 46% of the customers of traditional storage vendor Pure Storage had challenges handling container-related storage, compared to just 27% for the average Kubernetes user.
- Traditional HAProxy/NGINX are still clear leaders in the network proxy space, with Envoy and Traefik as runners up.
- The most used service meshes are Consul and Istio (just under 20% each), though many customers may have already been using Consul for other purposes.
- Also in the service networking space, Netflix’s Zuul quietly seems to be gaining users with nearly 5% — right under LinkerD — while at the same time generating fewer headaches for admins: only 23% of Zuul users complained about service mesh challenges, as compared to 27% for all respondents. Is Zuul a “dark horse” to watch here?
- Respondents with more than five clusters continued to see a decline in challenges related to scaling of deployment, but companies with five or fewer clusters were faring worse — particularly around scaling deployments, development culture and logging.
- Smaller companies were not as far along in realigning their organizational structure to meet the demands of cloud native computing. Although many organizations formed DevOps or SRE teams, the developer’s role in these organizations continues to be pulled towards handling aspects of the infrastructure layer.
- Among container management offerings from cloud providers, Amazon’s Elastic Kubernetes Service just edges out Google Kubernetes Engine. Both EKS and GKE have been tried by about 30% of survey respondents, while only about 17% have used Azure.
For more insights, please check out the “The State of the Kubernetes Ecosystem.”