Implementing Ultra Low Latency Data Center Services with Programmable Logic

Abstract

Data centers require many low-level network services to implement high-level applications. Key-Value Store (KVS) is a critical service that associates values with keys and allows machines to share these associations over a network. Mostexisting KVS systems run in software and scale out by running parallel processes on multiple microprocessor cores to increase throughput. In this paper, we take an alternate approach by implementing an ultra-low-latency KVS in Field Programmable Gate Array (FPGA) logic. As with a software-based KVS, lookup transactions are sent over Ethernet to the machine that stores the value associated with that key. We find that the implementation in logic, however, scales up to provide much higher search throughput with much lower latency and power consumption than other implementations in software. As with other KVS systems like redis, memcached, and Aerospike, high-level applications store, replace, delete, and search keys using a standard Application Programming Interface (API). Our API hashes long keys into statistically unique identifiers and maps variable-length messages into a finite set of fixed-size values. These keys and values are then formatted into a compact, binary Open Compute Storage Message (OCSM) format and transported in User Data Protocol (UDP)/Internet Protocol (IP) Ethernet frames over 10 Gigabit/second Ethernet (10GE) or faster network links. When transporting OCSM over 10 GE and by processing theperforming the key/value search in FPGA logic, a fiber-to-fiber lookup latency of under a half microsecond (0.5 a#x03BC;S) was achieved. Using four 10 GE interfaces, a single instance of the FPGA-accelerated search core achieves a throughput of 150 Million Searches Per Second1 (MSPS). Compared to traditional software, the FPGA KVS is 88 times faster while using 21x less power than socket I/O. Compared to software optimized with DPDK kernel bypass, the FPGA KVS was measured to process messages 10x faster while using 13x less energy.