The Dark Side of Scale

The infrastructure space is bursting with activity, with cloud computing and containerization leading the charge. New software, new approaches, new tools, and new dogma. And at the forefront of nearly every conversation about all these new developments, is “scale”. But when did “scale” become synonymous with “good”?

Most of us aren’t building search engines. Most of us aren’t exclusively running distributed microservices with service meshes in Golang. The IT masses are running Rails and PHP and MySQL and commercial off-the-shelf software. For every person running “kubectl”, there are 1000 piping “tail -f” into “grep”, and a million clicking buttons in Windows Event Viewer. Most IT is not dealing with Google, Amazon, or Microsoft sized problems.

Have we let our collective obsession with scale drive the narrative around the “right” way to deal with infrastructure? Is this negatively affecting how we build systems? How does this relate to the imposter syndrome many of us feel? Have we lost sight of the individual, the skilled operator, and the kinds of problems they care about? How did we get to this point, and how do we fix it?

By the end of this talk, audience members will be able to:

  • think more critically about scale, and its trade-offs vs. simplicity ** …when architecting systems and infrastructure ** …when looking at resumes of potential colleagues ** …when consuming tech media / press
  • expand their technical repertoire to include things that explicitly don’t scale to arbitrary heights, yet still nicely solve the problem
  • better identify cyberpunk tropes :)



Deepak Giridharagopal

Deepak Giridharagopal is CTO at Portland’s own Puppet. He is a long time Puppeteer who has worked on a huge variety of projects from Puppet itself, to creating PuppetDB, to helping launch newer ...