How Cloud Is Democratizing RDMA Use in the Financial Services Industry
Imagine David and Goliath were to meet yet again for their famous faceoff—except this time, each had the same equipment at their disposal. That’s what’s happening now in the financial services sector, thanks to the strides that public cloud providers have made.
Migrating to the cloud has much been talked about in the financial services industry recently, but one of the biggest nudges the industry as a whole has gotten to take the plunge comes from NASDAQ. Earlier this year, the stock exchange announced its latest, herculean venture: moving all of its 28 public markets to the cloud by 2030 in order to scale and introduce new features.
Financial services firms—whether you work in the trading industry or not—take note: If you haven’t already done so, it’s time to migrate to the cloud.
By moving to the cloud, as NASDAQ has realized, companies—whether the size of a Goliath or a David—get unparalleled access to a multitude of innovations entering the market every day, as well as older ones that weren’t as easily available. One such innovation that the financial services industry should be excited about? The availability of Remote Direct Memory Access (RDMA) via cloud.
What Is RDMA?
RDMA is a set of underlying technologies used in high performance computing (HPC) environments that allows computers to send data across networks at exceptionally high data rates with low latency by bypassing normal network communications data processing to load data directly into a remote application. In the high-frequency trading field where every second counts, RDMA’s high performance and low latency makes it extremely popular among the world’s top trading firms.
Traditionally used only by hedge funds and large investment banks and other rich Goliaths, RDMA once required specialized software and hardware—and a massive price tag to integrate it. Firms had to spend time setting up an enormous, private data center, paying licensing fees, and wait months to install expensive upgrades only to use parts of the incredible features this powerful program offers. This limited its use to firms that could justify those costs.
How Has Cloud Changed RDMA?
Once an exclusive capability only available to the largest firms, RDMA is much more accessible to financial services companies of all types and sizes thanks to the advancements of public cloud service providers, like Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform (GCP).
Unlike in RDMA’s earlier days, firms now only need to pay as they go. If a firm only needs to use RDMA two hours once a week, it only needs to pay to rent the services for those two hours. No longer do they have to buy a costly monolithic data center, pay to install it, wait for the installation to finish and pay licensing fees, only to use some of its features before eventually having to replace it when the system can no longer accommodate more upgrades. Instead, they can have access to just the features they need—however often they need them—immediately.
By offering RDMA, cloud providers are essentially democratizing this once-exclusive capability.
How Can It Be Applied Across the Field?
While RDMA’s most obvious use case has been in high-frequency trading, other financial services firms shouldn’t sleep on this newly accessible technology. RDMA can be used in a variety of scenarios, especially where heavy data loads and transfer across networks between discrete systems are key performance characteristics. Some examples include:
For investment firms, risk modeling has… well, a lot of risks. Traditionally, this involves compute-intensive analysis of large volumes of historic market data, portfolio holdings and other product static data—already a complex and intensive operation.
To improve investment decision-making and provide better outcomes for clients, wealth and asset managers are seeking to add more and more data sources and analytic techniques to the mix. From geographic information, weather and satellite imagery to real-time supply-chain analysis and shipping data, the variety, complexity and volume of inputs are growing exponentially. By combining RDMA with artificial intelligence (AI) platforms, investment firms finally have the computing ability to act on this vast stream of data, using today’s news to predict tomorrow’s biggest stories, and accompanying market swings, with great accuracy.
Identifying indicators of risk, however, is only the first part of the problem. Wealth and asset managers also need the ability to consume the outputs of this analysis in real time, so they can optimize thousands of portfolios continuously in reaction to the changing outlook. Investment managers can expand the risk picture and act on indicators to adjust global market data assumptions; expected returns; and regional, sector or currency forecasts. Asset managers looking to provide their clients with the best service possible can also leverage RDMA to maximize the response times of portfolio optimization engines. By automating this process and operating at scale across the entire client base, wealth managers can reduce volatility and drive superior performance for their clients’ portfolios.
In the world of payments, there is a need to process enormous volumes of transactions in the shortest possible time to cope with the vast flows passing across payment networks and the massive parallel systems that are used to spread the load across many processing nodes. Because of the huge quantities of transactions being processed, even a small reduction in the time taken to process any single transaction on a single node can significantly reduce the size of the parallel systems required to process the total or peak volume. Similar to the original trading use case, the time latency and duration of the processing can make a huge difference. By using RDMA at the points of communication across networks, significant end-to-end savings could be realized in reduced infrastructure needs.
If your financial services firm wants to gain access to powerful features that RDMA—and other programs like it—offers, you must migrate to the cloud, and quickly. You need a sophisticated technology partner to help you move to the cloud and help leverage some of these newly accessible tools, so you can take full advantage of its benefits.