2

I'm using couchdb since a while. I love that, but I was always working with a stand alone configuration.

Now I'm trying to set up a serious system made up of a "federation" of coucdbs. (i.e. a cluster). I understood that the most proper tool to use is coudhb-lounge but now I've to understand if such architecture is sustainable. In particular in terms of amount of data.

I tryed to get as much information from the following sources One way replication with CouchDB, What is the CouchDB replication protocol? Is it like Git?, CouchDB Replication Protocol , but I still feel I need some support.

So, as a first set of question, I'm trying to understand:

  1. I searched a bit around and from this post and this post I understood that it works over HTTP. Is it really efficient for huge amounts of data?
  2. is it possible to replicate 10Gb of data each day without encountering any serious load issue?
  3. In case it is slow, how much erlang programming would be needed to change couchdb replication protocol? Would I have to seriously hack the code or couchdb envisage also to use different protocols?

thanks

Improve this question
2
  • 2
    There is a project called BigCouch that provides clustering that was initially missing from CouchDB. bigcouch.cloudant.com They are moving the clustering work from BigCouch back into the original Apache CouchDB. From my setup of a BigCouch cluster, the data seems to move between nodes via Erlang. Might want to check out the product, or at least how they're moving data via the code: github.com/cloudant/bigcouch Commented Apr 16, 2013 at 17:46
  • thanks! I'll try that and let you know. Commented Apr 16, 2013 at 18:33

1 Answer 1

Reset to default
2

There isn't a "sync protocol", the sync is using the regular API. There is a loose description of the algorithm here, along with some comments on performance:

https://github.com/couchbaselabs/TouchDB-iOS/wiki/Replication-Algorithm

As to your situation, if you need a large replication setup consider also CouchBase. CouchBase was designed by the same people who designed CouchDb, and for their second take they designed it with focus on replication performance, iirc.

Improve this answer
Sign up to request clarification or add additional context in comments.

Comments

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.