Red Hat engineer talks system optimization, containers

The 2015 Red Hat Summit opened this week in Boston and attendees are expected to be knee-deep in topics such as system optimization, container-based approaches and application performance concerns.

Cameron McKenzie, editor in chief at TheServerSide, talked to a developer to learn the hot items at the conference and the general state of system optimization in the face of container-based system trends in the industry.

Jeremy Eder, a principal software engineer at Red Hat, discusses performance optimization changes as you go from virtualized systems to containerized systems. For example, crossing hypervisor boundaries creates performance issues. 

"There are plenty of workarounds," he says. "It actually takes a little bit of work, and that's kind of why it's ... hands off. You get performance almost for free in containers."

You can also hear Eder's insights about the details behind the Red Hat Enterprise Linux performance optimizations in the world of non-uniform memory access (NUMA) systems. NUMA is used in a symmetric multiprocessing system, in which multiple processors working under a single OS access each other's memory over a common path.

As for the summit, Eder hopes attendees walk away with a better understanding of Red Hat's work in the area of performance workloads in the cloud, among other things.

"We're building sort of a Rosetta stone for customers to learn how to categorize workloads ... and really make a fingerprint of each application that allows them to confidently pick an infrastructure," Eder says in the podcast.

The 2015 Red Hat Summit is an open source technology event that brings together attendees to discuss platform, virtualization, cloud and system management know-how.