The Cloudcast - Cloud Computing

Discussing Service Mesh Architectures

February 27, 2019
The Cloudcast - Cloud Computing
Discussing Service Mesh Architectures
Chapters
The Cloudcast - Cloud Computing
Discussing Service Mesh Architectures
Feb 27, 2019
Aaron Delp & Brian Gracely
Aaron and Brian talk with Armon Dadgar (@armon, Founder/CTO @HashiCorp) about the problems service mesh can solve, the underlying technologies, control plane vs. data plane considerations, and who is making decisions about service meshes within an IT organization.
Show Notes

Show: 387

Description: Aaron and Brian talk with Armon Dadgar (@armon, Founder/CTO @HashiCorp) about the problems service mesh can solve, the underlying technologies, control plane vs. data plane considerations, and who is making decisions about service meshes within an IT organization.

Show Sponsor Links:

Cloud News of the Week

Show Interview Links:

Show Notes:

Topic 1 - Welcome to the show. It’s been a couple years since HashiCorp has been on the show, so give us an update on the company - big round of funding ($100M) in November.

Topic 2 - A couple months ago we saw you in a video called “What is a Service Mesh?”. It was intended to be a “let’s make this simple” and you realize that a Service Mesh could be a lot of things - L4-L7 routing, Proxy, Encryption, Authentication, Application patterns. Is a Service Mesh solving a new problem, or is it pulling together lots of things that have existed at L4-L7 and application stacks in the past? 

Topic 3 - “Service Mesh” has become a pretty crowded and fragmented market over the last couple years. HashiCorp Consul has been around since 2014 (was originally “Service Discovery”) and now there’s Linkerd, Istio, Envoy and a bunch of variations. As you talk to people in the market, how are they evaluating the options out there?  

Topic 4 - Consul has evolved from Service Discovery to Service Mesh, and seems to have come from more of an authentication and security perspective (some others tends to be more routing-centric). Are there use-cases when one Service Mesh is a better fit than others, or should we expect that all/most of them will more or less converged on features over the next 12-24 months? 

Topic 5 - Can you give us some examples of how companies are using Service Meshes today (parts or all of the capabilities) and what teams are usually driving the adoption (infra/ops, security, app-dev, etc.)?

Feedback?