K
Kelly Sommers
kellabyte.com · Personal blog · on HN since 2011
Kelly Sommers writes about distributed systems, databases, protocols like REST and AMQP, and performance work on systems like Redis and LevelDB.
What HN saysCommenters cite her posts on why network partitions are not rare and call her REST explanation the best one.
21HN people shared it
6linked it in comments
15stories submitted
107story points
11 years agolast shared
2011people per year2026
Most-shared posts
- 1The “network partitions are rare” fallacy3 people · 5 points · first shared 2013 · discussion on HN
- 2Clarifying AMQP2 people · 37 points · first shared 2012 · 5 comments on HN
- 3Clarifying REST2 people · 9 points · first shared 2011 · discussion on HN
- 4What fuels my passion for technology & writing code1 people · 30 points · first shared 2011 · 5 comments on HN
- 5Continuous Client: Our multi-device dream but how do we build it?1 people · 5 points · first shared 2011 · discussion on HN
- 6Improving the protocol parsing in Redis1 people · 4 points · first shared 2014 · discussion on HN
- 7Using a bloom filter to reduce expensive operations like disk IO1 people · 4 points · first shared 2013 · discussion on HN
- 8Using bitmap indexes in databases1 people · 3 points · first shared 2013 · discussion on HN
- 9Create benchmarks and results that have value1 people · 2 points · first shared 2014 · discussion on HN
- 10When Metro design falls off the tracks1 people · 2 points · first shared 2011 · discussion on HN
- 11Clarifying the Saga pattern1 people · 2 points · first shared 2012 · discussion on HN
- 12The 99th percentile matters1 people · 2 points · first shared 2014 · 1 comments on HN
- 13Scalable Eventually Consistent Counters1 people · 1 points · first shared 2014 · discussion on HN
- 14Creating boundless experiences with Continuous Client and Transmedia | kellabyte1 people · 1 points · first shared 2012 · discussion on HN
- 15An alternative to paxos the raft consensus algorithm1 people · first shared 2015
What HN people wrote when they linked it
striking · 2015-05-27
"In my experience, all distributed consensus algorithms are either:
"1: Paxos,
"2: Paxos with extra unnecessary cruft, or
"3: broken. "
— Mike Burrows (possibly)
Edit: I think the reason this exists is best explained in this blog post (http://kellabyte.com/2013/05/09/an-alternative-to-paxos-the-raft-conse… It's not that this provides more than Paxos, the goal is to write something that's simpler and easier to implement. The article here doesn't actually explain this, and doesn't compare it meaningfully with Paxos, making it difficult to reason about.
shanselman · 2014-02-08
Kellabyte has the best explanation: http://kellabyte.com/2011/09/04/clarifying-rest/
messel · 2011-10-17
Here are the ones in my reader, admittedly I haven't visited often enough over the past couple of months:
http://www.igvita.com/
http://ejohn.org/
http://codingrelic.geekhold.com/ (Denny covers assembly and networking issues often in great detail)
http://buhrmi.tumblr.com/
http://elegantcode.com/
http://72lions.com/
http://kellabyte.com/ (mobile-ish dev)
http://lambda-the-ultimate.org/
http://ihumanable.com/blog
http://4loc.wordpress.com/
http://www.maxogden.com/
http://till.klampaeckel.de/blog/
http://www.quirkey.com/blog
http://www.mikealrogers.com/
http://cocoawithlove.com/
yummyfajitas · 2015-04-15
Partitions are not rare. A partition simply means that one node is not available to another, which does not have to be caused by a network failure. For example, a stop the world GC is equivalent to a network failure - the node is unreacheable at this time. So is a deployment with downtime (e.g. a 10 second restart per-node).
Also, while you can't access the partitioned area, you might have nodes outside that area. The question is whether the non-partitioned nodes become useless to you or not.
http://kellabyte.com/2013/11/04/the-network-partitions-are-rare-fallac…
https://webcache.googleusercontent.com/search?q=cache:D3htezwrrNIJ:kel…
itsdrewmiller · 2014-07-15
It's a little weird to leave leveldb out of the discussion here:
http://kellabyte.com/2013/06/06/leveldb-was-designed-for-mobile-device…
jbellis · 2013-11-25
Ah, the myth of the sufficiently redundant network.
http://pl.atyp.us/wordpress/index.php/2010/10/when-partitions-attack/
http://kellabyte.com/2013/11/04/the-network-partitions-are-rare-fallac…