Date | February 24, 2017 |
---|---|
Speaker | Christos Kozyrakis Stanford University |
Title | No Tradeoff: Reconciling Low Latency and High Efficiency |
Abstract | Low tail latency is essential for data-intensive, online applications, such as web-search, social networking, and software-as-a-service. Unfortunately, the simplest way to guarantee low latency is to disable any techniques that lead to high utilization in the public and private clouds that host these applications, including sharing cloud resources between multiple applications. Even worse, when targeting microsecond scale latencies, it is often necessary to resort to non-mainstream hardware and software approaches that further increase operational costs. In this talk, we will discuss the technical challenges behind the low-latency/high-efficiency tradeoff, present a set of hardware and software solutions that can eliminate the tradeoff, and review open questions on this area. |
Bio | Christos Kozyrakis is an Associate Professor of Electrical Engineering & Computer Science at Stanford University. He works on architecture, runtime management, system software, and programming models for systems ranging from cellphones to warehouse-scale datacenters. His past research includes energy-efficient data-parallel architectures, transactional memory technology, and practical hardware support for robust security abstractions. His current research focuses on resource efficient cloud computing, energy efficient compute and memory systems for emerging workloads, and scalable operating systems. Christos received a BS degree from the University of Crete and a PhD degree from UC Berkeley, both in Computer Science. He is a Fellow of the ACM and the IEEE. |
Resources |