Marketcetera 1.5 Released; Tests Demonstrate Speed, Performance On Par With Proprietary Products

by Ostatic Staff - May. 17, 2009

While news of Marketcetera's 1.5 release of its open source trading platform will certainly appeal to those working with financial-services specific software deployments, there's a neat little gem in this story that will make any open source software enthusiast smile. Marketcetera, the open source pioneer in the automated stock trading platform arena, has always had the advantage of fast deployments, infinite extensibility thanks to its built-in scripting engine and open nature, and impressive scalability -- and its 1.5 release builds on that foundation.

But claiming that a software platform is fast, scalable, and makes better use of the hardware it runs on is one thing -- proving it is another. Truth be told, different hardware configurations can yield slightly different results, and most companies don't have the time or resources to test software under controlled circumstances for any length of time.

Most companies, with the possible exception of Intel, with its labs in various locations for just this sort of purpose. Aside from the new features and enhancements Marketcetera users will experience in the 1.5 release, they can also rest assured that in testing conducted at Intel, Marketcetera performed on par with its proprietary competition. Thanks to community innovation and contribution, functionality and performance are continually increasing.

Of course, "testing" is also a rather vague term in this respect. I had the opportunity to ask Marketcetera's Vice President of Engineering, Yoram Talmor what the testing entailed and what the results mean for brokerages deploying Marketcetera's software. Saying that Marketcetera performed on par with proprietary products is satisfying and reassuring to brokerages using (or considering) the platform and open source software supporters alike -- but it's infinitely more impressive once one hears what was measured.

Talmor says that the main focus of the testing at Intel Labs was to measure the latency, throughput, and scalability of the platform's critical path -- meaning, essentially, how quickly and efficiently the platform handled the incoming market data as it moved through various modules in the platform, then sent it on to the order server through the embedded communication bus, and finally completed the transaction by routing the orders to the broker. It sounds very involved -- it is very involved -- but this is an industry where speed and accuracy are critical. Marketcetera's open trading platform held its own here, but it gets really interesting when one starts looking at scalability models.

Talmor says that once the performance characteristics of the critical path (the end-to-end flow of information described above) were determined, scalability could be worked out by creating many different paths, with different combinations of modules and functions. The measurements showed that the data flow overhead was minimal.

What that ultimately means is that the process is fast and efficient, and lends itself well to increased hardware utilization and scalability -- making the platform ideal for brokerages with diverse hardware configurations and a wide range of users and transaction volumes.

One of the modules Talmor specifically mentioned when he described the Intel Lab testing is greatly enhanced in the 1.5 release -- the Marketcetera desktop client app, Strategy Studio. This allows traders to deploy and manage strategies directly on remote servers, as well as manage multiple agents on multiple servers. This affords traders using the Marketcetera platform a better level of control along with the obvious increase in scalability. Strategy Studio also supports real-time intraday position and profit and loss tracking, and integrates a number of depth-of-book market feeds, including NASDAQ Level 2 and TotalView.

Marketcetera's platform is intriguing because it is making its way -- and proving itself -- in an industry that is one of the last bastions of solid proprietary platform usage. It has already definitively demonstrated the major advantages of an open source platform, and with surveys indicating almost half of the buy-side market considering changing automated trading systems within the next year, the company is in a strong position to face the future.