In distributed storage systems and hyperconverged environments, storage is scattered across a bunch of hosts all over the network. So does performance suffer if a workload’s data is across the network somewhere?

Should workloads have local data to perform as well as they ought? In other words, does data locality matter?

The Datanauts beam aboard storage expert Howard Marks to dive into the issue of data locality. Howard is a well known writer, speaker, consultant and Chief Scientist at DeepStorage. He’s also co-host of the Greybeards on Storage podcast.

Howard and the Datanauts discuss the basics of data locality and storage networking; explore the myriad bottlenecks and latency sources that can affect an application’s interaction with a storage system; and delve into the issues around hyperconverged infrastructure (HCI), data locality, and replication and resiliency.

Sponsor: Datadog

This week’s show is sponsored by Datadog. Built by engineers, for engineers, Datadog is a SaaS-based monitoring platform for cloud-scale infrastructure and applications. Sign up for a free trial and get a free T-shirt at datadog.com/datanauts.

Show Links:

DeepStorage

HCI And Data Locality – DeepStorage

Evaluating Nutanix’ original & unique implementation of Data Locality – Josh Odgers

Looking for the silver bullet that’s not there? My (and Datrium) response to the Data Locality debacle – Andre Leibovici