Cassandra Release Brings Speed Boost: Beginning of the End for MySQL?

by Joe Brockmeier - Apr. 14, 2010Comments (6)

Cassandra LogoThe Apache Software Foundation (ASF) has pushed out the 0.6 release of Cassandra with support for Apache Hadoop, and several improvements to the Cassandra architecture and speed. Don't let the release number fool you, Cassandra is in production use on some of the busiest sites on the Web, and may be giving MySQL a run for its money soon.

Cassandra is a NoSQL distributed database that was originally developed and open sourced by Facebook and is now used by Rackspace, Digg, Twitter, and many others. The project was picked up by the Apache Software Foundation's incubator project in 2009 and became a top-level project in February of this year. The 0.6 release is the first since graduating from ASF's Incubator.

The big news with 0.6 is the support for Hadoop and the speed improvements. Cassandra is being deployed by services dealing with thousands, maybe millions, of users making requests every second. The 0.6 release notes claim a 30% improvement in speed, which means maybe a little less Fail Whale and a little more Tweet goodness.

The Hadoop support adds the ability to run Hadoop queries against data in Cassandra. Prior to 0.6, you could insert Hadoop output to Cassandra, but this release brings native support and makes it much easier for the applications to work together.

Cassandra has been growing by leaps and bounds, and may be one of the projects where development moves faster than hype. Trying to get a grip on Cassandra, and what it can (and can't) do? Jonathan Ellis put up a nice fact vs. fiction piece last week talking about some of the misconceptions around Cassandra. He also rounds up some of the Cassandra deployments and addresses the question of Cassandra vs. MySQL saying "if not quite yet the end, then the beginning of it."

The rise of NoSQL solutions like Cassandra coupled with Oracle's acquisition of MySQL may do quite a bit to dent the growth of MySQL. It's unlikely to spell the "end" for MySQL, but it seems unlikely that MySQL is going to continue to see much growth -- particularly in large-scale deployments. But I'd be interested in hearing from admins and developers working with MySQL that feel differently.



Shailesh Patel uses OStatic to support Open Source, ask and answer questions and stay informed. What about you?



6 Comments
 

"Cassandra is being deployed by services dealing with thousands, maybe millions, of users making requests every second."


Does every application need to scale to this level? Majority of applications developed do not come close to one million requests per second.


0 Votes

I'm not an expert on hadoop and NoSQL technologys, but are really comparables?

hadoop can support a full data transaction schema? Can you get really data integrity on hadoop? How can you perform a really secure schema for the data in hadoop? For some applications maybe you can live with out this... but for many others this is not an option.


0 Votes

In the next to last sentence, should the "it seems likely that" actually read "it seems UNlikely that"? Otherwise, it seems to contradict the previous meaning. Or should the "MySQL" after this phrase actually be "Cassandra"??? Inquiring minds want to know! :)


0 Votes

@Yuri Paez


You are right about ACID compliance and data consistency/integrity. NoSQL just isn't there (yet, or maybe ever) and overall the technology isn't as mature as relational/object-relational DB technology. That said, not everyone needs the level of consistency/integrity of Oracle or DB2 (or PostgreSQL for an open-source example) and with the rise of mainstream distributed computing/clouds/insert-buzzword-here Cassandra and related highly-distributed, NoSQL platofrms fit very well (data integrity IS pretty good and consistency IS achieved...eventually...in that platofrm in most cases anyways ;-)


Thing is, MySQL isn't the be-all and end-all for those that need such transactional robustness. Cassandra was a LONG way from being the basis for corporate accounting systems/ERP but MySQL wasn't the only or even best choice there either. Look at it this way: MySQL was always touted as being open source, blazing fast, simple to use (compared to BDB or GDB and such) ISAM-based solution for read-heavy high-traffic internet apps where those very issues about data consistency were glossed over. This is EXACTLY where NoSQL is focusing! Then look at the other writing on the wall:


* PostgreSQL is much more stable and robust in terms of ACID compliance, concurrency, data consistency (they've never let users enter dates like "feb 30th" etc) and has long since closed the performance gap in apps where such things are important and writes occure more frequently, so there is a Free alternative to MySQL on that front


* MySQL is now in Oracle's hands. I don't see Oracle having a problem with keeping MySQL Free (they left BDB open when they aquired Sleepycat for example) but you can bet they will DEFINITELY not be supportive of MySQL backends/tools/extenstions that encroach on Oracle's territory. PostgreSQL and other Free SQL DBMSes do not have this millstone around them, and it would take awhile for a non-Oracle fork of MySQL to gain traction.


* Where MySQL has always said to shine Cassandra already does well and it scales better.


MySQL won't die but its reign on the internet is probably over. It will always have its place but probably in the lower end, where something simpler and (in specific situations) faster than PgSQL (and cheaper than Oracle) with "good enough" transactional support is required.


But in my personal experience using both, PostgreSQL is head and shulders above MySQL when you need a robust Free DB ;-)


0 Votes

There is a need for both and most application deployed with Cassandra also have a MySQL or other RDB along with it.


MySQL (or a newer open source version of it) and Cassandra will both continue to grow.


0 Votes

AWFUL name for it.


0 Votes
Share Your Comments

If you are a member, to have your comment attributed to you. If you are not yet a member, Join OStatic and help the Open Source community by sharing your thoughts, answering user questions and providing reviews and alternatives for projects.


Promote Open Source Knowledge by sharing your thoughts, listing Alternatives and Answering Questions!