Could Big Data damage Oracle?
By Andy Ormsby
19 Jul 2011
Category:
Business Insights
I’ve seen the future and Exadata is not the answer.
I've spent the last nine months talking to a large number of well-known startups and more established companies doing exciting things with big data. Some have big ambitions, others have some seriously large and fast growing datasets. However, they all have one thing in common. None, that I have met, plans to use Oracle. One is using Oracle but is actively working to get off, but that's a subject for another day.
That's not to say that they are avoiding SQL. Plenty of them are using MySQL and Postgres, at least for part of their infrastructure. Even here, a common theme is that these solutions are not expected to scale as the business grows and that something different, such as Cassandra or Riak, is needed. I'm not hearing that they plan to move to Oracle to get the scale or performance they are looking for.
And who can blame them?
Why? Part of the reason is software cost. Almost all you need to know about the cost of Oracle software is explained by the existence of companies such as Rocela (http://www.rocela.com), a fine group of people who have built a business around providing excellent advice to Oracle customers on how to mitigate their Oracle licence costs.
There is hardware cost too. Want to scale out your Oracle database across lots of low cost commodity servers with cheap, direct attached SATA disks? Good luck with that.
What about big enterprise customers? They will keep buying Oracle and insanely expensive shared storage to use with it because it has become a habit, they are risk averse and because they can afford to. But even here, there is change in the air. We are getting approached by some of the world’s biggest corporate IT departments with use cases that fit Cassandra well and their opening statement is often: "We're not sure what the solution is going to look like yet, but we're absolutely certain that it won't be on Oracle."
Interesting times.
blog comments powered by Disqus