Monte Carlo on EC2 August 8, 2008
Posted by James Webster in : finance, development , trackbackVia TheServerSide.com I came across Max Gorbunov’s (Grid Dynamics)Â report on the scalability of GridGain on Amazon EC2. Having previously thought about whether hedge funds might find EC2 a quick way to scale up batch processing of derivatives risk reports, I found the following EC2 limitation discovered by Max and his team interesting:
Our second problem turned to be the default maximum number of running instances per user – 20. Since we were going to run grids much larger than 20 nodes, we needed to override the default limit. It took several steps and a few more days to negociate with Amazon EC2, but eventually, we were granted the right to run up to 550 nodes. It seems that the business process of requesting a large amount of nodes is still not very well-defined by Amazon.
So there is still an opportunity for another cloud computing player to focus on massive scalability only. IBM perhaps?
Grid Dynamics are also doing some interesting work in their Convergence project which aims to bring together compute and data grids. Although most vendors make the claim that their products act equally well in both roles most products seem to be strong in just one only… and most investment banks have installed both a cache (i.e. data grid) and a grid.
Comments»
James~
I would like to suggest, in terms of massive scalability, Joyent. They are the official host of Ruby on Rails and their hardware load balancing allowed LinkedIn’s Bumper Sticker app for Facebook to scale at 1 billion page views per month.
You can’t get much better than that!