Infinity’s attempt over the last couple of years to transform itself through acquisitions into a dominant one-stop trading and risk management vendor has left a lot of people wondering how it intends to knit all its recent purchases into a single integrated suite of products. Last month, the company unveiled a new integration strategy that relies on middleware supplied by recently acquired Mint Systems as the glue.

Mint’s middleware will serve as the system’s backbone, providing a rule-based messaging technology that will allow disparate Infinity and third-party trading, risk-management and back-office systems to communicate with each other and with outside settlement messaging technologies.

Under the system, an executed trade using the Infinity system would be automatically routed to the back office and the relevant custodian bank or counterparty via a network such as SWIFT. Both the trading and back-office systems would communicate directly with the risk-management technology to ensure that timely data are available for risk assessments.

The middleware will work in collaboration with Infinity’s Network Trade Model, a shared-data model that transfers trade and related information between applications. The Network Trade Model links myriad trading programs on the basis of asset class, instrument and trading volumes. The model not only allows risk managers, traders and quantitative analysts to make precise search inquiries of the data, but also to access the same information at the same time. Foreign exchange swap traders and bond traders, for example, will be able to access identical pricing or yield-curve information for U.S. Treasuries simultaneously.

The Infinity MasterFiles, another key data management function, provides a reference data server to store and administer all reference data centrally, including counterparty names and hierarchies, calendars, and instrument-static data.

Infinity has restructured its mix of trading technology offerings into three product lines: Infinity Forex, focused on high-volume foreign exchange and money-market trading; Opus, focused on interest rate derivatives and structured products; and Infinity Derivatives, which is designed to sit on top of the Infinity Data Model and Fin++ Class Library trading system.

Infinity says it is emphasizing open, out-of-the-box solutions that can be customized by the client. Its Enterprise-Wide Risk Management (ERM) system features the Infinity Platform, a tool kit that builds custom risk-management applications using third-party developers or Infinity’s global professional services group. The platform consists of the Infinity Data Model, housing instrument data, and the Fin++ Class Library, a risk application development tool.

Risk managers can use the ERM system to integrate and deploy third-party models that the risk manager has customized for the firm’s specific trading needs. A risk manager can also integrate third-party upgrades to extend a credit risk model to new asset classes, such as those related to the euro. The Fin++ Class Library streamlines the process of generating new code to launch new models.

Infinity’s Devon System will provide trade-entry and processing functionality for a wide range of standard instruments, including interest rates, futures, foreign exchange and equities. Infinity Derivatives Back Office, built on top of the Infinity Data Model and Fin++ Class Library, provides straight-through processing primarily for interest rate derivatives and their hedging instruments.

Finally, Infinity has recognized the popularity of distributing risk information firm-wide and to regulators via the Internet by building into new technology the capability to use a variety of Internet languages, including Microsoft’s ActiveX and XML. ActiveX defines all documents as objects, which improves a system’s search-and-retrieve capabilities. XML offers greater flexibility in writing Internet code for particular financial instruments or reports. Panorama Net+ already uses ActiveX for Internet-based reporting. This year, Infinity will focus on building XML to exchange data between components, and thin-client interfaces on Intranets and the Internet.

New partners

These technological realignments have been matched by a series of marketing initiatives designed to leverage relationships with third-party vendors. Infinity has extended its trading technology to handle equity derivatives through third-party products such as Equimax, which is integrated with Infinity Derivatives, and RioEquity, which is based on Panorama.

It has also formed a partnership with RiskMetrics that will give the latter access to Infinity’s global distribution network, and will give Infinity the opportunity to attract the client base of RiskMetrics’ 415 and CreditManager.

Under the agreement, Infinity clients now have the option of starting out with a cheaper risk solution and scaling up to Panorama. “We would like to be able to sell a system to institutions that are not prepared to invest in a full-fledged, production-quality risk management system at first, and then upgrade to a production system such as Panorama,” says Brian Robins, senior vice president of global marketing at Infinity.

Mirror, mirror on the wall, which is the fastest risk management system of them all? It’s a topic that has preoccupied both Infinity’s Panorama and MKIRisk’s Carma systems.

The competition started last year, when the analytic processing speed of Carma was tested by American Management Systems’ independent consultants. The analytics generated between 1.43 million and 3.2 million valuations per minute on one machine with two processors.

Infinity responded to the challenge by running its own tests at the Dell Computer Corp. Application Solution Centre in Limerick, Ireland. According to recently released figures, Infinity’s analytics executed 8.4 million valuations per minute during a Monte Carlo value-at-risk calculation using 2,000 scenarios—five times faster than other risk management systems, according to Infinity officials.

It’s important to note, however, that Panorama was tested using considerably greater computing power—11 workstations, including 10 dual-processors and one mono-processor workstation. Infinity’s analytics conducted portfolio valuations consisting of 261,000 trades in nine minutes; generated a historical VAR with 250 simulations in 20 minutes and 16 seconds; and produced a Monte Carlo VAR using 2,000 simulations in 61 minutes.

“These results show that Panorama has the core database and calculation speed to meet real-world requirements and, crucially, that performance can be maintained no matter how large the relevant portfolios become,” says Ian Green, Infinity’s senior vice president in charge of Panorama.

MKIRisk officials concede the scalability of Infinity’s analytics, which is useful to firms needing to expand computing power for increased trading activity. However, the same officials contend that scalability can also be a liability. “The issue is that while Panorama may be scalable, users would have to scale up by buying a few hundred machines containing about 1,000 processors running in parallel to match our speed on only one machine with only two processors,” says Jeff Mantel, director of product management at MKIRisk. “The monstrous operational risk this number of machines running in parallel would spawn repulses most banks. This is why the vast majority of Panorama sales are to banks other than tier-one.”

A key factor that distinguishes MKIRisk’s Carma analytics, contend company officials, is that the technology was originally built as a risk-management system, which means Carma is more robust than typical trading-based risk technology. Carma also houses the Vector Processing Architecture, which is unique to Carma and allows all trades to be processed with only one run through the executable. Infinity, on the other hand, says Panorama is ideally suited to run the large number of simulations needed for portfolios that profile credit exposure, and the cost of scaling to many machines is modest, given today’s prices for PCs.

CastleNet, the New York-based risk technology firm spun off from broker-dealer giant Tullett & Tokyo, has announced plans to offer market data and real-time analytics over the Internet.

The BEAST, launched last year, is an Internet-based server that includes a suite of financial calculators for interest rate derivatives such as swaps, caps and swaptions. The server, which operates on a Windows NT platform, allows users to exchange analytics and data via a browser such as Microsoft Explorer or Netscape. Financial institutions can use the single server to simplify systems that may be riddled with a maze of spreadsheets and databases across the firm.

By offering TheBEAST.com, CastleNet has made the BEAST analytics and data available on a web site. Investment managers simply log on the site; the site then grabs the selected portfolio information and runs the analytics. TheBEAST.com contains analytics tailored to investment managers, which means that instrument pricing and risk calculations can incorporate longer-term horizons. The system calculates pricing and hedging parameters for most financial instruments.

The system’s analytic capabilities are based on BEAST Financial Components, developed last year. This is a collection of analytics ranging from international yield calculations to valuation and term-structure models for derivatives, using Microsoft’s Component Object Model (COM)—a library that enables developers and traders to deploy analytics in any model that supports COM. CastleNet, which is expanding its technology to handle ActiveX and XML, wants to allow developers to assemble and disassemble applications quickly, as well as wire models together on different trading technologies.

TheBEAST.com joins other firms such as Bankers Trust, State Street and Measurisk.com in targeting Internet-based analytics at institutional investors such as hedge funds. Still, CastleNet has yet to sign on any clients besides Tullett & Tokyo and a $25-million venture with Dow Jones.

One of the greatest benefits of TheBEAST.com, a thin-client, 32-bit system, is that it is remotely accessible from any laptop. Carl Carrie, president of CastleNet, says the system can calculate Bermuda swaptions in seconds, as opposed to minutes with other systems. TheBEAST.com can also perform simple yield calculations as well as calculations for exotic structures. “There are no competing electronic communications networks that can do this,” he adds, noting that while Bloomberg may have greater capabilities in the domestic paper arena, CastleNet’s product is much stronger in the global fixed-income market.

The market data used in these calculations are currently obtained from Tullett & Tokyo and GovPX, but the software company is in discussions with a number of the primary market data sources. CastleNet is also preparing to release a Java version in the near future. Carrie shrugs off concerns that Java lacks the speed and efficiency to deliver real-time data. “It is pure Java, but it performs startlingly fast,” says Carrie, whose firm has been working on the project for two years.

CastleNet is offering a 30-day free trial. For more information, see www.TheBEAST.com.

Was this information valuable? Subscribe to Derivatives Strategy by clicking here!