SE Podcast #17 – Kyle Brandt & George Beech

Jeff & Joel are back with guests this week - joining them are Kyle Brandt and George Beech, our very own sysadmins/ops guys/[insert your own term here]

After a brief test of the emergency broadcast system, we plunge right into the podcast, including:

• Joel is late because he had to go up to AOL HQ to pick up some more floppy disks for the office

• We recently launched Blog Overflow - which was actually a pretty substantial engineering and technical background to making it happen.

• A huge discussion of the the SE infrastructure setup and why we host our own instead of relying on Amazon or another outside provider.  Hint: not only can we roll our own better, but its a lot cheaper too.

• In the department of little tweaks with big effect, check out our "network apocalypse": at one point, Stack reconfigured to separate different types of network traffic due to microbursting overloading the network - this simple upgrade greatly increased the throughput and efficiency of the existing infrastructure.

• In the good news department: Jeff talks about the expanding user cards just introduced and based on a conversation from our podcast a few weeks ago with Michael Natkin

• In the bad news department: we've decided to cancel Stack Overflow DevDays due to low ticket sales - you can read Joel's full blog post for all the details.

• Finally, we welcome our newest engineering hire - Demis!

That's it!  Tune in next week at the usual time for another episode with more guests!

Episode source