Kubernetes is often selected by organizations to meet specific requirements that can give businesses a competitive advantage. Requirements may include features involving machine learning to make real-time recommendations, to perform ‘intelligent queries’ or they may be needed for real-time, on-site data processing like that available from Edge infrastructure setups.

How do you take advantage of the processing speed at the edge or other specialized cluster setups and yet still maintain security, reliability, and speed? Is it possible to avoid snowflake clusters from propagating? And finally, can you spin up a consistent cluster with all of the tools and applications that make a complete Kubernetes platform when you need one?

“From my experience in the industry, before the cloud, when a specific requirement was requested, teams would go out and do an analysis and figure out what they actually needed to solve that problem in a bespoke way. Afterward, they would go to IT and present a bill of materials, describing all of the items needed. IT would take it from there and pull everything together, providing whatever infrastructure and middleware required. But over the last decade or so, IT has moved towards a more homogeneous and commoditized infrastructure, where everyone consumes from a menu of fixed things," says Cornelia. 

This is especially true ever since Marc Andreesen’s landmark 2011 piece, “Why Software is Eating the World” where industry started to see software as the competitive differentiator for every company. Packet though turns this concept back around with the provocative statement: ‘Make infrastructure your competitive advantage”.

Packet has vast experience designing infrastructure on bare metal that can be deployed anywhere. They are seeing an increase in Kubernetes-driven architectures for specialized requirements like machine learning or other specific compute-intensive applications like those required for edge computing use cases. In this episode of the Art of Modern Ops, Cornelia Davis, Weaveworks, CTO and Cody Hill, Packet, Field CTO explore what this type of Kubernetes infrastructure looks like in practice.

About Cody Hill, Field CTO at Packet

Cody Hill is Field CTO at Packet, where he helps Enterprises and large SaaS companies take advantage of Packet’s platform. Previously the Director of Technology at Platform9 — a venture backed managed hybrid cloud provider — Cody also served as Lead Cloud Architect at General Electric. With over 15 years experience operating production infrastructure at global scale, Cody leverages his deep knowledge of VMware, Openstack, Kubernetes and the DevOps ecosystem to accelerate customers on Packet.

Sign up to be notified of future episodes:


Listen to the entire episode: 

The Art of Modern Ops · Is Kubernetes Ready for Specialization?

Listen on your favorite podcast platform:

Apple Podcasts | Spotify | Stitcher | SoundCloud | Pocket Cast | Google Podcast | Overcast