Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News KubeCon + CloudNativeCon Europe 2020 Keynote Summary

KubeCon + CloudNativeCon Europe 2020 Keynote Summary

This item in japanese

The annual KubeCon + CloudNativeCon Europe event ran as a virtual conference in August this year, five months later than planned due to the COVID-19 pandemic. Key takeaways included: the cloud native community strives to be inclusive, and it is possible for everyone to contribute in their own way; there is an increasing importance in understanding end-user adoption patterns of technologies; it is essential to recognize and discuss the topics of security, observability, and energy efficiency as more organizations move to the cloud; and the community should always remember (and empathise with) the people, systems, and businesses involved in adopting cloud technologies.

Priyanka Sharma, general manager of the Cloud Native Computing Foundation (CNCF), opened the event by encouraging people from all walks of life to get involved with the CNCF and Kubernetes communities. According to the new State of Cloud Native Development Report there are 6.5 million “cloud native” developers in the world, 1.8 million more than in Q2 2019. The clear call to action by Sharma was to “get involved”; the community strives to be inclusive, and welcomes contributions from engineers, documentation writers, program managers, and people with a range of skills.

Next on the virtual stage was Cheryl Hung, VP ecosystem at the CNCF, who promoted the importance of establishing and examining the "real world" view of cloud native technologies. She discussed the new CNCF end-user Technology Radar that is being created by a group of over 140 companies who are meeting regularly. According to the CNCF announcement blog post, the goal of the CNCF Technology Radar is to share what “tools are actively being used by end users, the tools they would recommend, and their patterns of usage”. The first edition of the Technology Radar focused on continuous delivery.

CNCF End-User Tech Radar

Liz Rice, vice president of open source engineering at Aqua Security and chair of the CNCF Technical Oversight Committee (TOC), presented an overview of the CNCF TOC. She discussed the pros and cons of a project’s participation within the CNCF from the perspective of end users, vendors, contributors, and maintainers. As captured in a CNCF tweet that quoted Rice’s talk, the TOC admits and oversees all projects in the CNCF and has a mandate to “facilitate driving neutral consensus” for projects:

Foundations like #CNCF create a framework for trust so that people can come together & build great tech, even if they work for competing organizations.

Another core theme of this KubeCon was the importance of security within the cloud space. Shane Lawrence, senior security infrastructure engineer at Shopify, and Kris Nóva, chief open source advocate at Sysdig, presented a keynote titled: “Open Source Intrusion Detection for Containers at Shopify”. Use of the CNCF-hosted Falco Project was discussed in this talk and throughout the conference, both in other presentations and the Slack-based “hallway track”.

Falco intercepts and parses system calls at runtime using “deep kernel tracing built on the Linux kernel, eBPF, and ptrace”. Kubernetes and container metainformation can also be captured, and engineers can describe security rules that protect against unknown or unwanted behavior, such as coin mining.

In the keynote, “How to Love K8s and Not Wreck the Planet”, Holly Cummins, worldwide IBM garage developer lead, opened by stating that data centres alone consume 3% of the world's energy. As an increasing amount of this energy is being used by workloads running on Kubernetes, she asked “is k8s helping, or making things worse?” Cummins argued for identifying and eliminating unused “zombie workloads”, aiming for increased utilization, and embracing “disposable infrastructure” that can easily be torn down and rebuilt using best practices like GitOps.

De-zombify cloud workloads

Derek Argueta, senior software engineer at Tesla and previously a site reliability engineer at Pinterest, presented a keynote titled “Building a Service Mesh From Scratch - The Pinterest Story.” As shared at a previous EnvoyCon, two years ago Pinterest began rolling out the Envoy Proxy as an ingress load balancer for its dynamic web traffic. Since this time they have expanded Envoy's footprint to include service mesh service-to-service communication while supporting Kubernetes and non-Kubernetes deployments using a homegrown control plane.

Key takeaways from this talk included the need for platform and tooling teams to always focus on solving business problems (“avoiding tech simply for the sake of tech”), plan for incremental progress and deliver value at each step, and strive to get early buy-in from other teams that will be using the technologies.

PInterest Envoy adoption

In the keynote, “My Stint as a Chameleon”, Constance Caramanolis, principal software engineer at Splunk and co-chair of the conference, discussed the different roles she has undertaken over the years when working with cloud native technologies. Wearing “multiple hats” and working in roles and departments such as software engineer, support, customer success, sales engineer, and product management allowed her to develop empathy and understand that “success is more than technology; it is about successful adoption.”

Several attendees on Twitter, such as KellyAnn Fitzpatrick, industry analyst at RedMonk, liked both Caramanolis’ call to “write more documentation” and the suggestion that “reference documentation != user documentation.”

Vicki Cheung, staff software engineer at Lyft and co-chair of the event, presented a keynote focusing on “Observing Kubernetes Without Losing Your Mind”. A lot of teams find operating and monitoring larger scale distributed systems like Kubernetes daunting. Cheung suggested that end-users focus on four main tools: event streaming, such as streaming Kubernetes events to queryable storage; metrics, for example capturing data from all components of the platform and processing these and displaying them on dashboards appropriate for the task; logs, which included both streaming logs to a centralized collector and using the CLI-driven “kubectl logs -f X” for quick identification of problems; and inspecting the system, via out-of-tree solutions like “kubectl-debug”.

Many of the presentation slide decks can be downloaded via the Sched pages on the KubeCon website, and the videos will be made available to the public in due course.

Several community leaders and analysts have summarized their key takeaways from the event, including Katie Gamanji, Rich Burroughs, Janakiram MSV, and KellyAnn Fitzpatrick.

The next KubeCon + CloudNativeCon 2020 will be the North American edition that will also run as a virtual event, November 17 - 20.

Rate this Article