About Us | Search | FAQ | Contact Us
cost effective treasury solutions
It is difficult to imagine any busy trading room operating successfully without support from the profusion of IT systems that have been deployed in the last ten years or so. The quantity and diversity of information that must be assimilated and the speed at which markets can move, continue to drive the development of ever more sophisticated software to aid dealers, treasury managers and risk managers.
This process generates its own momentum. Improvements in technology that increase the effectiveness of trading operations have an inherent tendency to create two further, important effects. Firstly, profit margins are squeezed, as more market participants are able to offer keen prices by managing their positions and exposure to risk more accurately. Secondly, the fixed costs of running the trading or treasury operation increase dramatically, with the IT budget taking up an increasing proportion of total costs. These two factors, amongst others, create a demand for further productivity improvements, which are often satisfied by another round of technology spending.
In order to continue offering a comprehensive range of services many banks find their treasury operations trapped in this spiral pattern of spending, with no obvious alternative. Moreover, as increased reliance is placed on growing volumes of data and complex software, it becomes more and more difficult to continue the evolution of support systems in a cost effective manner. In some application areas the point has been reached where only a radical rethink can deliver sufficient benefits to make further change worthwhile. One such area is a fundamental tool for traders, treasury and risk managers; the deal capture, position keeping, P&L and risk management system.
All significant dealing rooms rely on some form of deal capture and position keeping system to manage day to day trading, profitability and risk. These systems have evolved over a number of years to provide broad instrument coverage, detailed profitability forecasting and recording and a range of risk management facilities. However, there are two significant problems associated with the vast majority of such systems, namely the high cost of ownership and their failure to offer truly global function. These two are closely linked.
All these costs make up a significant proportion of a treasury department's outgoings. But for a bank with multiple trading rooms in diverse geographical locations, this is just the start of the problem. The typical system architecture, traditional client/server, relies on high power desktop client processors to perform much of the data processing. Data is stored on local servers and delivered to the client processors over a local area network. Data transfer rates are high in this environment. For example, a P&L or risk report may require the data relating to several thousand trades. This has to be sent from the database server to the client for processing, so to achieve adequate response times a broadband communications channel between the two is required. This is fine over a LAN, but becomes impractical over greater distances, as the bandwidth required would be prohibitively expensive. Consequently, each trading room of an international banking operation requires its own, local database in order to deliver timely responses to dealers.
This effectively means replication of the entire system in each location, with the consequent cost implications relating to hardware, software licences, office space and system management overheads. On top of all this, an additional layer of software, procedures and complexity is needed if global positions or risk are to be monitored. Again, more costs, and the headache of ensuring data integrity.
Functional benefits of
The key problem is, once again, the usual client/server architecture. There are two main difficulties inherent in the improvement of a typical client/server system's functional performance. The first is that processing large volumes of data at the client (desktop) does not make use of the most powerful element of the system i.e. the central processor cluster (server). The consequence is that, as more and more complex calculations need to be handled, the performance of the whole system is degraded. This results in slower response times and, at best, frustration for users. At worst, and this is quite common, the theoretical benefits of improved function are lost, because performance deficiencies outweigh the potential business advantages.
Secondly, there are some critical business functions that simply cannot be delivered within this architecture. The most obvious example is multi-site or globally consolidated positions in real time. Because databases are required in each operational location (due to high data traffic between client and server), global views of positions, profitability or risk rely on an extra layer of software to replicate or consolidate information across the network. This is expensive, difficult to manage and in many cases fails to deliver adequate performance. A global treasury or risk manager can only maximise the business's profitability if real time access to all positions is available instantly. Only the largest banks have achieved this in anything close to real-time, and at a cost that deters many other aspiring global operators.
A new architecture
The technology employed in this unique system was chosen very carefully. Web browsers and third-party tools such as Enterprise Java Beans or other transaction servers were specifically avoided because they do not scale, are not flexible or reliable enough and cannot run the complex risk reports that all modern trading rooms demand. Lucidity has a proprietary technology, built specifically for the trading room by people who understand trading room dynamics.
On the server Lucidity has written a C++ application running under Windows 2000, with the single database implemented in SQL Server 2000. These two products are extremely cost-effective and, crucially, offer exceptional performance per dollar. The C++ core of the system is extremely fast, flexible and robust. It is also fully scalable on servers with more than one processor, and can scale across multiple servers.
The client software that each trader uses is written as a standalone Java application that is typically run just like any other desktop application. It can however run within a browser if required. By using Java at the front-end, the system can be deployed on almost any trader's desktop including PCs, Sun workstations and even hand-held devices such as iPAQ, Psion and Palm.
Multiple benefits of the thin client solution.
Global high performance
Simplified system management
A whole range of costs is
drastically reduced, some are removed entirely.
n One central set of servers supports the entire enterprise. These servers are utilized far more efficiently, supporting different time zones throughout each 24 hour period instead of mutiple servers each supporting a single time zone for a local working day.
n Database licence fees are reduced by having a single database.
n System management overheads are minimized by being centralised. This includes in-house and third party support, hardware maintenance and even office space.
n No additional software, system management or hardware is required for global consolidation; it is inherent in the system.
n Implementation of the system is simplified. Once the central system is installed adding a new user, wherever located, is a straightforward task involving little more than loading a small program from a disk or downloading it from the network.
The additional option
of ASP deployment
Why aren't all systems
designed this way?
|Home | About Us | Search | FAQ | Contact Us|