Published by Gbaf News
Posted on July 18, 2014
4 min readLast updated: January 22, 2026
Add as preferred source on Google
By Nikita Ivanov, CTO and founder of GridGain
The financial services industry has some of the highest computing demands of all sectors in business. While speed, scalability and real-time analytics are important considerations in any industry, in financial services speed literally translates directly into revenue, and latency into loss.
As a result, the industry was among the first to adopt in-memory computing (IMC), a technology which increases the speed at which data is processed by orders of magnitude. With disk-based processing — the de facto method — you take your data to the computation. IMC, on the other hand, brings the computation to your data, yielding superior performance. My firm, for example, worked with one top global financial institution to achieve processing at a billion transactions per second (tps). By comparison, disk-based processing is time-consuming and resource-intensive; yielding performance that at best reaches a few million tps.
While there are no limits to IMC’s applications, financial services companies are currently leveraging the technology in three major areas:
Risk analysis automation
In-Memory Computing Provides Extreme Computing For Hyperscale Demands Of Financial Services
In financial services, each decision must be weighed against risk, and strategies must be continually realigned to mitigate for an endless variety of factors. For example, a Central American labor strike might delay shipments in a way that affects a particular investment’s earnings. As the questions that must be asked in order to determine how a scenario like this will ultimately play out are incredibly sophisticated, the degree to which identifying and analyzing risk can be automated largely determines a company’s agility in the face of uncertainty.
Automation of a process this complex, however, requires computing power that is many times faster and more scalable than what can be achieved with traditional methods. IMC provides advanced data processing capabilities required for obtaining answers quickly enough to act on them. By enabling the automation of this process, IMC enables firms to eliminate latency associated with human analysis.
Real-time Trading
When trading decisions must be made instantaneously, traditional computing methods fall short in their ability to process large amounts of data and perform transactions within seconds. For example, Forex trading entails variations so small that firms must trade in huge volumes to reap even a modest return. Therefore, any degree of latency — even milliseconds — can lead to loss of tens of thousands of dollars.
Staying competitive in trading (including support, modeling, testing, back-testing and re-pricing) — requires highly optimized hardware and software that operates at levels that may currently only be achieved through IMC.
Consider, for example, a major international investment bank that needed to react much more quickly to market events. Without the ability to very rapidly process massive amounts of data, it was safer to simply not trade than to risk trading the wrong way. This meant that a market event could potentially derail the day’s strategy and bring trading to a halt. Using IMC, they were able to recalculate the market change mathematical models and backtest against the previous 60-90 days to determine possible effects and iterate strategies. Instead of freezing, IMC enables the company to instantly analyze the impact of the event, iterate the strategy in real-time and continue trading.
Hyperscale post-trading processing
An additional element that traditionally puts the brakes on trading is the requirement that firms close their books on the day’s trade before starting the next. Of course, in an era of 24/7 global trading, waiting for hours to process the data isn’t an option. IMC delivers the performance necessary to manage the extremely large amount of data at the speed that this entails, enabling uninterrupted global trading.
For financial services, the question is not whether to leverage IMC, but rather how to best leverage it. It is worth keeping in mind that as IMC runs on commodity hardware, these upgrades do not require a significant upgrade or change in infrastructure. For example, a leading global investment bank used IMC to deploy a 3x higher resolution model without replacing any existing hardware — a competitive edge that ultimately enabled them to double the size of their business.
Now, with the falling cost of RAM, and the availability of platforms with enterprise-grade security features, we can expect to see even more innovative uses of IMC in the near future. And, while the technology is being rapidly adopted in other sectors, we can expect the financial services industry to continue leading the way in the use of what is considered to be one of the most powerful computing technologies currently available.
Explore more articles in the Technology category
