Links
Want to work with me at the $DAYJOB?
I am the CEO of OmniTI where I do all sorts of stuff I find absolutely fascinating.
Writing
I spent a few years writing down my thoughts about how one should approach problems. If you're looking for a how-to guide, a cookbook, or a reference this book is not for you. If you want to learn by challenging the way you think, pick up a copy.
Hacking
It is rumored that I write code sometimes. I often don't believe this myself, so I use this to follow what it is that I'm working on:
You need to upgrade your Flash Player
Categories
Recent Entries
- Plan Wagon
- Monday, May 10 2010
- pg_controldata from SQL
- Thursday, April 29 2010
- from pcap to postgres: forensic performance analysis
- Friday, April 2 2010
- pg_amqp slides from PgEast2010
- Tuesday, March 30 2010
- Thoughts on the cloud
- Tuesday, March 23 2010
- Managing risk by deploying incessantly
- Thursday, March 18 2010
- Squeezing precise numbers into fixed width types... or faking it
- Friday, March 5 2010
- What does it mean to be a patriot?
- Friday, February 26 2010
- Death by icicle.
- Friday, February 12 2010
- Sexified fonts for the web.
- Monday, January 11 2010
Thursday, June 11. 2009 at 18:41 (Reply)
Excellent explanation of erasure, thanks. Do you see erasure codes not being just for storage, but for dynamic data like in-memory databases, caches, and messaging? Services that also suffer from the dropsies.
Thursday, June 11. 2009 at 23:11 (Link) (Reply)
I don't. But I have to admit that I haven't thought about it so much. Erasure codes are usually used to counter unreliable storage. There is also a non-zero cost of reconstructing data given a set of erasure codes. The things you describe are very performance centric and paying for a erasure code reconstruction would ruin them... Perhaps if processors had erasure code-centric instructions that would help that. Your question is definitely something something I need to digest mentally. Thanks for the feedback. I like it when my brain hurts.