Amar Phanishayee

About

Hi! I am a Senior Principal Researcher at MSR in Redmond. I currently lead Project Fiddle at MSR.

The goal of my research is to enable the creation of high-performance and efficient systems for large-scale data-intensive computing. To this end, my work follows two broad approaches:

1. Radically rethinking cluster architecture

Energy-efficient clusters for data intensive computing – FAWN
Freespace optics for datacenter networks – ProjecToR

2. Build robust clustered systems and protocols

First to analyse and solve TCP throughput collapse and latency insensitivity for clustered systems – TCP Incast
Distributed KV systems – FAWN-KV, Flex-KV (and the underlying Ouroborous protocol)
Reliable and scalable multicast – Ricochet, Plato
Compositional reasoning for systematically testing distributed systems
Efficient distributed deep learning – Fiddle

I have also occasionally looked at addressing issues in geo-distributed systems:

High-perf multi-hop wireless transfers using opportunistic caching – Ditto
Scalable one-hop overlay routing – NuRON
Systems/abstractions for connected devices – Bolt (storage) and Beam (programming framework and runtime)

As a graduate student at Carnegie Mellon University I worked with Dave Andersen and some stellar collaborators on FAWN, Incast, and Ditto.