Shenango: Achieving High CPU Efficiency for Latency-sensitive Datacenter Workloads
This paper presents Shenango, a system that provides microsecond-scale tail latencies while enabling unused CPU cycles to be used by latency-sensitive batch processing applications.