Capital markets have become significantly more complicated over the past few years. Rapid cloud migration continues to challenge traditional exchange protocols and trading behaviors. As a result, network connectivity requirements between data centers and dispersed endpoints have been rerouted and redefined in today’s high-frequency trading ecosystem.
Financial exchange best practices can serve as a blueprint for all hyperscale data centers. Like it or not, behind every transaction is a labyrinth of algorithms and networking infrastructure technology that, despite the dispersed cloud computing ecosystem, converge in one place: the data center.
It’s complicated, and data center managers have a lot on their minds. Very few financial institutions have the inhouse resources, talent, and bandwidth to execute an effective data center strategy. The resource requirements to keep ahead of the maze of technical and logistical options to execute low-latency and high-performance trading are difficult to achieve and even harder to manage in meeting high-intensity, post-pandemic customer demands.
Cloud migration requires choosing a data center colocation strategy. The decision boils down to a do-it-yourself (DIY) versus managed service approach.
The sands are shifting
Colocation is when an organization locates its computer equipment in a data center managed by a third party. Financial markets, exchanges, and trading venues make space available within their data centers for financial institutions and managed service providers to colocate.
Originally, firms colocated to benefit from ultralow-latency trading. Now, drivers have expanded to include achieving deterministic latency; access to accurately time-stamped market data; and connectivity to other venues, exchanges, and cloud services.
Low latency remains vital for algorithmic trading. Measured in microseconds, and even nanoseconds, latency is the time it takes for a piece of data on a network to travel from one endpoint to another. Reducing latency is crucial for traders who need to respond rapidly to changing market conditions and breaking news. Many factors affect latency, especially hardware location and network connections. Trade execution speed is critical in maximizing profit and loss, and a competitive advantage comes from having the best communication links to hardware in the best locations.
The main advantage in colocating at a trading venue’s data center is to reduce overall latency impact and to gain direct access and proximity to the venue’s matching engine, trading network, and data.
In contrast, a DIY approach is when companies decides for themselves to arrange the installation, configuration, connectivity, and monitoring of their data center equipment within the space they lease from the exchange inhouse. There are several things to consider when going DIY, including how much space and power is required, network connections, bandwidth, latency, market data feeds, staffing, and whether there is a need to connect to additional third-party data or services.
Connectivity, market data, power, and space costs can quickly add up, along with capital expenses and unpredictable fees for equipment, maintenance, and support staff.
How can data center professionals make the right network management decisions to maximize business continuity, reduce time to market, lower costs, and decrease risks?
Managed service advantage
There are myriad factors to consider when moving a data center from one jurisdiction to another. It’s paramount to work with an expert provider that knows the local rules and regulations. In the financial markets, outsourced providers give access to a suite of services immediately in a new location including order routing, market data and the procurement, installation and management of trading infrastructure. Managed service providers (MSPs) can make the process of moving locations less stressful and more efficient, with a single point of contact and 24/7 boots-on-the-ground access to both technical and capital markets expertise.
Exchanges are partnering with cloud providers to migrate their offerings but remain tethered to capital market data centers around the world. Nasdaq has announced a multi-year partnership with AWS using its on-premise private cloud infrastructure. AWS’ Outposts product family directly connects into its core network to deliver ultra-low-latency edge computing capabilities from its primary data center in Carteret, NJ. Similarly, the Chicago Mercantile Exchange (CME) has announced a 10-year partnership with Google to move their trading platform to the cloud.
Mass migration of data centers in capital markets requires that trading performance keep pace with rapid technical advancements. Trading success requires accessing extremely powerful servers, with the best data lines and connections close to where the trade is physically taking place. Processing close to the source of the input data provides the lowest possible impact to the network infrastructure due to bandwidth demands and lowest latency access between input and response – and speed matters. When stocks are trading, microseconds can mean the loss or gain of millions of dollars.
A recent independent report by industry analyst firm Celent, Colocation of Financial Markets Trading Infrastructure, identifies the pros and cons of in-house DIY management versus a managed service model. The report found that MSPs offer beneficial value-added services for capital markets clients.
Advantages include cost savings, trade efficiency and simplified access to data and network infrastructure support. These capabilities enable trading firms to focus on their core business competencies. Celent analysts interviewed trading firms and their technology providers and found that the key decision criteria when engaging an MSP includes:
- Consultation and expert advice on the ideal configuration of hardware, network connectivity, location, data feeds and network bandwidth.
- Agility and flexibility to take advantage of ever-changing investment opportunities by rapidly and easily deploying trading strategies in new markets.
- Access to high-end network services, leveraging high-speed solutions, including ultra-low latency, in-data center Layer 1 connectivity to link to trading venues, new customers and other service providers.
- Operational efficiency and future proofing with access to the latest technology, and highly experienced staff in all global jurisdictions who help to navigate cultural, linguistic and regulatory obstacles.
The future is today
Firms need to keep their data center infrastructure ahead of the technology adoption curve. Financial exchanges are achieving this by partnering with colocation service providers that have experience in delivering vast amounts of raw market data, supporting multicast requirements and scalable solutions to accommodate ever-expanding bandwidth.
As traders diversify their portfolios, their market data needs can place excessive network capacity pressures on exchange infrastructures. Competitive advantage comes with provider partnerships that can accommodate these requirements and handle high-volume data bursts. The heat is on, and we’re seeing far more high-impact events due to market volatility. The factors driving high and low data spikes vary widely, from political and economic events to bad actors creating cyber chaos with more sophisticated breaches and disruptions.
In a survey of global market CIOs, Frost & Sullivan noted that nearly one third will be moving components and workloads to MSPs by 2025. And as exchange connectivity requirements evolve into even more complexity, partnering helps cut through the administrative burdens and costs.Buyers, sellers, brokers, analysts, market makers, regulators and stakeholders need connectivity, colocation and the ability to access market data instantaneously and reliably across hybrid, multi-cloud environments. Streamlining toward productivity gains is essential.
As the trend of data center migrations continues to accelerate in the volatile world of trading, financial exchange and data center managers will need to leverage single source relationships to obtain the multiple feeds, continuity and economies of scale they need to serve the growing community of worldwide around-the-clock traders.