Meeting summary

[Sidu Ponnappa]

Hello everyone,

Here's a brief outline of the stuff that came up during the meeting:
* Eight people attending
* Oldest no-sql db among those attending is 8 months old, a Mongo db
app in the medical domain
** Total size of the dataset to be stored is still uncertain
* One open-source project similar to Hoptoad that isn't in production
yet that uses MongoDB to store error logs
* There were two projects using CouchDB
** Couch cannot do map-reduce across instances
** Sharding is handled through CouchLounge
** Couch views are very I/O intensive, so the bottleneck is I/O, not CPU
** Building the B-Tree is expensive because it is append only.
** Couch uses 'materialised views' which are built up front
** Replication is slow because it goes revision by revision
** CouchDB isn't going to work too well for situations where you're
only interested in recent data, whereas Couch will update the view for
the entire dataset

The next meeting will be on Saturday, 22nd Jan 2011. See you next year!

Best,
Sidu.