Towards energy proportionality for large-scale latency-critical workloads

By David Lo, Liqun Cheng, Rama K. Govindaraju, Luiz André Barroso, Christos Kozyrakis. International Symposium on Computer Architecture, 2014

330 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Hardware and Architecture, Information Systems

View Paper: https://www.semanticscholar.org/paper/ffd7144bdc14bd1ceb7abf9324d17d1a6bb12abb

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Hardware and Architecture

In Information Systems

Citations in the corpus

IX: A Protected Dataplane Operating System for High Throughput and Low Latency (2014, background): “Our work focuses on improving operating systems for applications with aggressive networking requirements running on multi-core servers.”
Tradeoffs between power management and tail latency in warehouse-scale applications (2014, background): “proposed a system that exploits this property for a websearch benchmark, adjusting DVFS aggressiveness based on the difference between observed latency and the latency agreement [18].”
PARTIES: QoS-Aware Resource Partitioning for Multiple Interactive Services (2019, background): “The third approach focuses on eliminating interference altogether, by partitioning resources among co-scheduled services, using OS- and hardware-level isolation techniques [23, 35, 36, 42, 43, 49].”
Seer: Leveraging Big Data to Navigate the Complexity of Performance Debugging in Cloud Microservices (2019, background): “Second, utilization is not always a good proxy for tail latency and/or QoS violations [14, 28, 61, 62, 73].”
Neurosurgeon: Collaborative Intelligence Between the Cloud and Mobile Edge (2017, background): “Datacenters typically experience diurnal load patterns [32], leading to high variance in its DNN query service time.”
An Open-Source Benchmark Suite for Microservices and Their Hardware-Software Implications for Cloud & Edge Systems (2019, background): “However, services with the highest utilization are not necessarily the culprits of a QoS violation [61], taking the system much longer to identify the correct source behind the degraded performance and upsizing it.”
Heracles: Improving resource efficiency at scale (2015, background): “For instance, Google websearch servers often have an average idleness of 30% over a 24 hour period [47].”