DevOps Paradox Hosts Section CTO Daniel Bartholomew to Discuss Running Containers at the Edge
August 15, 2022
What is the edge? It’s a complex question to answer, and the answer often depends on who you ask. As the podcast hosts note, the term “edge” is ambiguous and hard to grasp. If you talk to someone at a hyperscale data center, for example, they may define the edge as the network perimeter of the region you’re operating in. Put another way, the edge is the limitation of their network boundary.
But for many workloads, choosing a single hyperscale facility is a tradeoff between performance and ease of management. Imagine this all too common scenario: a company has developed a low-latency application and deployed it to a single hyperscale region, only to find that they were never able to get the performance they expected or needed. So, the logical next step is to rearchitect that application so it can run in multiple data centers. Unfortunately, even after rearchitecting the app and running it in multiple hyperscale locations, this company still has not solved its performance challenges.
Our co-founder and CTO Dan Bartholomew recently joined the DevOps Paradox podcast to talk about this frustrating scenario and how Section is helping companies run containers at the edge. In this episode, Dan discusses how Section thinks about the edge as a compute continuum. He explains how Section views the edge as a large network of location that can run your application, where a hosting system is responsible for the placement of the application.
In an ideal world, this means that under certain conditions it might be best to run your application in a single hyperscale facility. In other circumstances, however, it might be most appropriate to have your application run in multiple hyperscale regions. Or, there may even be other times when your application should be running in tens or even hundreds of locations simultaneously.
But accomplishing this degree of flexible, distributed application hosting is extremely difficult. As Dan explains, Section considers all potential compute locations – from hyperscalers all the way down to the ISP – to determine where your applications should run. This is how we answer the question, “what is the edge?” It’s our view that distributed compute offers the potential for broad benefits in terms of not only improved performance and ease of management but also increased availability and resilience, better scalability, decreased cost, and more.