As published in HPCWire
Gaining an Edge: Wall Street Firms Expand Productivity in “Edge HPC” Arena
High Performance Computing – or as it is increasingly called, High Productivity Computing – continues to drive innovation and profits for a wide range
of industries. Today, technological evolution is driving HPC in new directions with non-traditional “Edge HPC” applications. Wall Street firms
are increasingly seeking to improve productivity with both traditional and Edge HPC.
Traditional HPC users are scientists, engineers, or analysts who run applications that are computationally demanding enough to require a more scalable
architecture than a standalone PC or workstation. Commercial HPC usage has encompassed automotive engineering, pharmaceutical design, oil exploration,
Hollywood special effects, and (of course) financial analytics. For decades, major banks have conducted economic simulations and portfolio analysis
to make optimal investment decisions, minimizing risk or exposure for their investment vehicles. Thanks to dramatic improvements in price/performance
brought about by commoditization, boutique investment firms are among the new group of entry-level users, and the finance industry is growing its
overall share among HPC users.
While these traditional high-performance applications are continuing to evolve, the HPC industry is undergoing a fundamental expansion as well. A number
of factors are combining to fuel the adoption of HPC technologies in non-traditional application areas. Tabor Research refers to this group of
applications as Edge HPC.
Driving the Edge
One important change driving Edge HPC adoption is the continual explosion of data, types of data, and sources of data. Data is generated continuously
from bar codes, RFID tags, video broadcasts, cell phone conversations, financial transactions, surveillance videos, CT scans, and seismic activity
monitoring. Of course, merely recording this data is not enough – in fact, recording it all can be impractical or impossible – and the resulting
challenge is to extract advantageous decisions from these mountains of data. Meanwhile, communications networks have gotten more complicated and
more intertwined with computation.
In addition to the above drivers, incremental HPC usage has an important enabler: the rise of scalable solutions based on industry-standard components.
HPC systems used to typically involve RISC processors and proprietary UNIX operating systems. As systems migrated from RISC to x86-based processors
from Intel and AMD, and the operating environment shifted away from UNIX, the systems became easier for more new users to adopt. Today Linux is
the predominant HPC operating system, but Windows is becoming more prevalent, especially in the commercial sector, and again this is opening the
doors to HPC for a broader set of users.
Edge HPC Applications in Finance
One way in which the finance industry is being transformed is in the ability to track and analyze continuous streams of trading data for currencies,
stocks, bonds, and commodities worldwide. Although the volume of information is daunting, there are tremendous rewards available to a bank that
can, for example, find a temporary arbitrage amidst a combination of derivatives, or gain seconds of advantage in determining a pricing trend.
Real-time trading analysis is a stellar example of a company increasing productivity through complex event processing.
Process optimization is another significant Edge HPC category for finance. Setting prices for portfolios is subject to risk analysis (a traditional
HPC application) and free market forces, but the process by which prices are set – for example, by combinatorial auction – can be optimized through
the use of HPC.
Gaining an Edge
For both traditional and Edge HPC applications, it all comes down to competitive advantage. Financial analytics and portfolio analysis applications
help Wall Street firms lower risk for their projected returns, and real-time pricing analysis and business process optimization help them make
decisions and execute faster.
The finance industry is therefore continuing to use HPC to improve productivity, not only in traditional applications, but also in the emerging Edge
HPC space. Through the increased adoption of high-productivity computing, both large and small Wall Street firms can continue to seek a compitive