*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Tess Malone, Communications Officer
Summary Sentence: Alex Daglis gives a lecture.
Full Summary: No summary paragraph submitted.
Alex DaglisTITLE: Network-Centric Computing for Online Services
ABSTRACT:
Modern datacenters provide an abundance of online services to billions of daily users. Each service comprises several software layers deployed on thousands of servers, which communicate over the network to collaboratively construct a response to each incoming user request. In addition to frequent inter-server communication, a large class of services involves minuscule computation per request and microsecond-scale response latency requirements per participating server. Such execution profiles establish networking as a first-order performance determinant and motivate a vertical system design rethink, taking a network-centric approach. In this talk, I will introduce a holistic system redesign targeting the most challenging latency-sensitive online services, including (i) a specialized lightweight network stack; (ii) scalable on-chip integration of the network interface logic; and (iii) new network operations with richer, end-to-end semantics that can be efficiently executed on smart network endpoints without CPU interaction. I will highlight the role and demonstrate the effect of each of these three key features with systems built throughout my dissertation.
BIO:
Bio:
Alexandros (Alex) Daglis is a sixth-year Ph.D. student at EPFL, advised by Professor Babak Falsafi and Professor Edouard Bugnion. His research interests lie in rack-scale computing and datacenter architectures. Alex advocates tighter integration and co-design of network and compute resources as a necessary approach to tackling the performance overheads associated with inter-node communication in scale-out architectures. He has been a founding member of Scale-Out NUMA, an architecture, programming model, and communication protocol for low-latency, distributed in-memory processing. Scale-Out NUMA has been prototyped and awarded a U.S. patent. As an intern at HP Labs, Alex worked on the design of The Machine’s unique memory subsystem.
