Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Riak, Latency and Distributed Systems

Riak, Latency and Distributed Systems

Everything is distributed. Any engineer building a system that's running in production needs be aware that she is relying on a series of components, services, and processes that need to work together to deliver data to the customers. This is not easy, and the tools you pick to deploy will impact greatly your ability win users and increase revenue.

Riak is an open source, distributed database being used by companies like Comcast, AT&T, and GitHub in production. Like any database, it comes with a specfic set of tradeoff that make it suitable for a certain set of applications. In this talk, Tom will discuss Riak and the types of applications people are building on it with special attention paid to dealing with latencies in such a distributed system.

Tom Santero

June 12, 2013
Tweet

More Decks by Tom Santero

Other Decks in Technology

Transcript

  1. “...widespread underestimation of the specific difficulties of size seems one

    of the major underlying causes of the current software failure.” --EW Dijkstra Notes on Structured Programming 1969 Wednesday, June 19, 13
  2. IF response > 10ms Send 2nd Request = ~ 99.9

    : 50ms (5% increase total reqs.) Wednesday, June 19, 13
  3. Demo: sudo  tc  qdisc  add  dev  eth0  root  netem  delay

     150ms Network: for  x  in  /sys/devices/system/cpu/*;  do (while  true  ;  do  openssl  speed;  done)  & done wait CPU: Wednesday, June 19, 13
  4. Write Once Read Often N W R = = =

    5 N 1 notfound_ok: False Wednesday, June 19, 13
  5. Write Once Read Hardly N W R = = =

    3 Q Q Wednesday, June 19, 13
  6. Write Once N DW R = = = 3 2

    1 CRITICAL Wednesday, June 19, 13