C24 and GigaSpaces partner to create Grid based matching and reconciliation engine

C24 has partnered with GigaSpaces to create a scalable, Grid based matching and reconciliation engine for financial services.
C24's Integration Objects (IO) have been integrated with GigaSpaces' Jini implementation in order to build the matching and reconciliation engine.

Wayne Meikle, managing director of C24, told stpzone.com: "We're using each other's technology to build financial services solutions. The GigaSpaces JavaSpaces Grid technology forms the platform, and C24 Integration Objects the financial services data models and toolkit to build these solutions."
C24's IO technology provides the ability to convert FpML, MurexML, SwapsWireML, Swift, FIX and many other message standards into Java Objects.
Complex matching rules can be applied to these objects and sent to GigaSpaces' JavaSpace engine. The rules can then be executed across a number of networked machines. Initial tests on Sun's Blade servers demonstrate linear scalability and performance figures as yet unseen on traditional RDBMS based engines.
Meikle says there are many benefits of converting FpML, FIX, Swift and other messaging standards into Java Objects, the task C24's application will perform. "It's a way to make the use of the message structures efficient in terms of computer resource requirements, faster and more scaleable in throughput, and quicker to deploy solutions into production."

He believes there are many pitfalls in using XML for message transfers. He says the main problem is parsing - this is the method of processing messages down to individual elements within the message or data model. "With large amounts of XML data you have to re-parse each time for each system. This is computationally very expensive as it uses lots of processing power. The reason using Java Objects is more efficient is that you parse the data once into a Java Object form, and then it does not need to be re-parsed. It already is. So moving these Java Objects around the place is more efficient. They are also generally smaller in size," he says.

"Generally using XML is good for external inter-firm data exchanges, as you need to use it as the lowest common denominator. But intra-firm it is not. And this is where the whole JavaSpaces/Java Object benefits kick in."

Meikle says the system can be based on Sun's Blade servers, and those servers have lots of distributed processing power. "The GigaSpaces JavaSpaces implementation is a Grid technology. This technology scales linearly in throughput based on available processor power if it's designed correctly. So the ability to scale from a few to many processors in a high density hardware package allows you to build a scaleable solution from the entry level to high volumes. It's cost effective based on this, and you can quickly add more capacity."

http://www.stpzone.com/Show.aspx?ID=206NEWS