Slashdot Mirror
CouchOne, Membase Merge, Form NoSQL Powerhouse
Julie188 writes "CouchOne and Membase, two of the most popular noSQL projects, have merged in an attempt to become an open source database powerhouse. Even the company's new name is merged: Couchbase. The founders of the new Couchbase say they will offer the ability to scale from the largest data center and distributed cloud environments all the way down to smartphones and other mobile devices. As is the standard disclaimer during merger announcements, the leaders also promise to continue their support for their open source, community versions of their programs."
instead of two projects I've never heard of, I'll only have to ignore one!
Do you even lift?
These aren't the 'roids you're looking for.
whenever I hear nosql, I instantly think of this: mongodb is web scale: http://www.youtube.com/watch?v=b2F-DItXtZs
Anybody want my mod points?
Facebook (Cassandra)
Google [Maps,Earth,Gmail,Youtube,App Engine,Code,Reader,etc] (BigTable)
Zynga, Paypal, Vodafone (Membase)
Dilbert RSS feed
NoSQL doesn't mean unreliable, and SQL doesn't mean ACID-compliant.
CouchDB (one of the products mentioned in this article) goes to some lengths to preserve data integrity. It doesn't do delayed commits the way, say, MongoDB does, and it uses an append-only file format that means each document is written to disk in a completely ACID-compliant way.
MySQL didn't have any transactional capability in early versions, and even today is quite happy to corrupt tables beyond repair if the power goes out during a write operation.
Be careful. People in masks cannot be trusted.
The article you linked is mostly an explanation of why Facebook switched from Cassandra to HBase, which they currently use.
Still a NoSQL product, just saying.
CouchDB, for one, is ACID compliant. NoSQL has nothing to do with lack of ACID compliance or performance. It is a catchall term to refer to a whole range of databases that are designed to solve different problems without using SQL as the query language.
Amazon (Dynamo)
Twitter, Digg (Cassandra)
Yahoo (HBase)
Netflix (SimpleDB)
BBC (CouchDB)
The Lotus Domino database is also NoSQL and is used in many enterprises.
These tools don't "mangle" data in the sense you seem to think they do, quite simply what these tools do is provide an "eventually consistent" database. IE a write won't necessary be propagated instantaneously to every single database. Instead it will be "eventually"(with eventually depending on the scope of the problem and the overall system architectures involved) be consistent. Now for large classes of problems(social networks, search engines, message boards like /.), this is a huge boon. By relaxing the timeline constraint for consistency then can make both reads and writes MUCH faster, scale up better, handle faults better etc. However you are right in that there are also large classes of problems for which eventual consistency isn't a very useful tool. For instance a real time database in a factory or even databases for small businesses where there is no real benefit to going to an eventually consistent system with more fluid data structures. Again, right tool, right job.
Monstar L