The IMF's information technology team has had a busy past two years as the fund has taken on an increased workload as a result of the global recession, leading the team's head, Jean Salvati, to implement a new software system to cope with the extra demand. Writer Michelle Price

Never before has the International Monetary Fund (IMF) been required to work so hard, and so quickly. As watchdog of the global financial system, lender of ultimate last resort, and promoter of global growth and economic stability, it has had a busy two years. Whatever the economic climate, however, the fund's analysis and insight on global economic boom, gloom, and latterly doom, are driven by a vast range of risk models fuelled by swathes of complex data.

Spare a thought for the IT team that makes this research and statistical analysis possible. At the IMF, the task of ensuring the fund's technology is up to the job falls to its information technology officer Jean Salvati, whose team provides support for a range of financial and macroeconomic modelling. During the past four years, says Mr Salvati, these models - which rely heavily on resource-intensive Monte Carlo simulations - have become ever more sophisticated. "Because of the financial crisis, there is an increased emphasis on financial surveillance and this entails implementing increasingly sophisticated models for risk analysis and risk management," explains Mr Salvati.

Infrastructure strain

As demand for more complex modelling has grown, however, the fund's IT infrastructure has become strained, resulting in extended computational run-times for simulations. The fund, says Mr Salvati, has "got the most out" of the individual work-station model, whereby individual processor cores are deployed to run models over a number of hours, sometimes days. Mr Salvati resolved on a change of strategy. "We started looking for a software platform that would allow us to distribute the workloads of the models across different machines," he says. The fund has turned to Platform Computing, a specialist in grid-based super-computing, whose Platform Symphony software system will allow the institution to disseminate complex economic computation across multiple, often-underutilised computers within the IMF global network.

For the IMF economists who are highly familiar with Excel, it is crucial that the new distributed system accommodates the ubiquitous application. "For us it is very important to have a tool that would allow us to interface with Excel, and one of the advantages of Platform Symphony is that it allows us to do just that: we can easily design the software so that the user sits in front of Excel, kicks off the task in Excel and the workload of that task is distributed in the background to our high-performance servers," says Mr Salvati. In short, the user has to do nothing differently.

The second critical issue was flexibility, says Mr Salvati. Because the fund's IT team will still have to service economists using legacy software applications, it would be very resource-intensive and inflexible if the Platform Symphony software system required a grid of dedicated machines. "Platform Symphony will use whatever resources are available on the machines that are on the [existing] grid: this dynamic allocation of resources was very important," says Mr Salvati. The ability to distribute or 'load balance' the workload across available machines allows the fund to optimise resources, while accommodating existing and new applications.

Platform Symphony is also "completely agnostic", he adds, when it comes to operating systems and processor architecture. Since the fund operates a mixture of machines that run both Microsoft Windows and open source system Linux, as well as a mixture of Intel and IBM processing architectures, it is vital the fund can "mix and match" the operating systems and the types of core processing when implementing the new software platform. The software provider operates an application programming interface, which allows the IT team to interface the Platform Symphony product with a range of core processing units. "It means that when it comes to select new hardware, it's not going to be a limiting factor. If something better comes along we can select that and link it to the grid."

Early stages

Mr Salvati's team is only at the beginning of the implementation phase. Most of the risk models implemented on the new platform during the first fiscal year will relate to analysis of systemic risk, and of individual institutions. "We're going to use this first year for implementing the major models and evaluation and benchmarking," says Mr Salvati. "After that, we're going to look at the results not just from my point of view, but from the institution's point of view: does it make sense to invest more money in that type of infrastructure, do we want to invest more, and how big do we want the grid to be?"

When completed, the move to a distributed model will accelerate processing time for economic modelling tenfold. As such, adds Mr Salvati, the new architecture will be instrumental as the fund develops new modelling and simulation tools to assist its members and support the fund's growing global role.

Career history

Jean Salvati

2008 - information technology officer at IMF.

2004 - developer at CAI.

2002 - engineer at ATI Research.

PLEASE ENTER YOUR DETAILS TO WATCH THIS VIDEO

All fields are mandatory

The Banker is a service from the Financial Times. The Financial Times Ltd takes your privacy seriously.

Choose how you want us to contact you.

Invites and Offers from The Banker

Receive exclusive personalised event invitations, carefully curated offers and promotions from The Banker



For more information about how we use your data, please refer to our privacy and cookie policies.

Terms and conditions

Join our community

The Banker on Twitter