AFTAs 2017: Best New Technology Introduced Over the Last 12 Months: Infrastructure—Metamako

Metamako, founded in 2013, is a provider of low-latency network devices designed to allow capital markets firms to simplify their network stacks, increase their network visibility, and support high-performance processing. The firm’s clients typically use its devices to gain a competitive advantage (in terms of speed and low latency) over other market participants, with market data replication performed in four nanoseconds and multiplexing in 69 nanoseconds.

A key area of differentiation for capital markets firms is their ability to connect to exchanges quickly and with minimal and predictable/consistent latency, a process known as determinism. And while pure speed and performance are basic requirements for most firms’ trading technologies, the need for programmability and configurability has become crucial in recent years, a trend that led to Metamako unveiling its E-Series devices, designed to run three FPGAs on a single platform, while maintaining the lowest possible latency. According to Metamako, the advantages of adding FPGAs to devices supporting trading activities include: greater speed, as there is no overhead when accessing memory; determinism or the ability to consistently and accurately predict latency; and performance by way of processing large data volumes at high speeds, by way of parallel processing. 

Metamako has been working with FPGA technology for a number of years. However, its E-Series can now run up to three FPGAs on a single device, providing users with increased processing power and super-fast connections directly to exchanges. And while the biggest limiter of data transmission speeds is cable length, having three FPGAs in close proximity on a single device dramatically reduces transmission times for trading firms looking to minimize latency. 
Metamako’s E-Series initiative allows user-firms to run multiple performance-sensitive applications on a single device—for example, trading algorithms, matching engines, pre-trade risk checks and market data processing can be run simultaneously at ultra-high speeds—while all three FPGAs can be configured to operate in concert, offering combined processing power to single trading strategies, while also supporting the running of larger, more complex applications.