Everything you wanted to know about Dynomite but were afraid to ask
Dynomite is an open source distributed key value database modeled after
Amazon’s Dynamo. Dynomite was built from the ground up to support what
would be a traditional RDBMS’s worst nightmare: low latency random read
and write access to large binary objects.
Dynomite is different from memcache style distributed hash tables in
that it handles all routing and failover internally. Clients need merely
connect to one node to have access to the entire cluster. It can provide
low read and write latency to clients even in the face of node failures
due to its relaxed consistency constraints.
During this session we will discuss how to tell whether Dynomite is a
good fit for an application. Discussions will also include dynamic
cluster configuration, how to choose an appropriate storage engine, and
how to tune a Dynomite cluster for some common scenarios.
Cliff Moon Dynomite creator and committer
Cliff Moon is a software engineer at Microsoft who gets to spend some
of his time working on open source. He was instrumental in helping
launch Powerset, a semantic search engine acquired by microsoft in
2008. He started out as a Java developer, was later reborn as a
Rubyist, and lately has discovered he is an Erlanger. He is the person
behind Dynomite, an Erlang based open source distributed key value
database modeled after Amazon’s Dynamo. When not hacking on Dynomite,
Cliff likes to cook and brew beer.