More thoughts on derivatives risk and grid computing April 25, 2008
Posted by James Webster in : finance, development , trackbackLike ‘em or loath ‘em, derivatives are here to stay… probably. For more on that discussion, check out Economist.com’s well-balanced article ‘Wither the derivative?’.
What the credit crunch and sub-prime crisis have highlighted is the ever increasing need for effective risk management of derivative contracts. My day job has involved enhancing a scenario analysis system to stress-test hundreds of exotic derivative positions; ultimately this means processing a huge number of quantitative finance calculations. Naturally a grid computing environment is involved.
So a few thoughts about that. I previously wondered about using Amazon EC2 to cheaply scale up on demand. It looks like DataSynapse, one of the leading grid vendors on the Street has had the same idea: see Today’s Grid forecast: “Partly Cloudy”.
GridGain, an open-source grid computing framework for Java has been coming along nicely. It would make a suitable alternative to Hadoop in my recommended open source derivatives risk platform stack. They are also thinking about EC2 on their roadmap:
We are working on seamless integration between external clouds like Amazon EC2 for on-demand discovery within GridGain. This feature will allow automatically or programmatically “boot-up” external cloud when load picks up and “power it down” when load subsides.
Nice. Where is the GridGain equivalent for .Net? Alchemi.Net looks fairly dormant. Also, Terracotta have almost released a new version of their ‘clustered JVM’ product; combined with its support for Ehcache is it an effective replacement for Oracle (nee Tangosol) Coherence?
Comments»
James — You might also want to check out GigaSpaces’ Amazon EC2 offering at http://www.gigaspaces.com/ec2
Geva Perry
GigaSpaces