Portrait of Srikanth Kandula

Srikanth Kandula

Partner Researcher


I am a partner researcher in the Networking group at Microsoft Research, Redmond. My interests are broadly in building and analyzing networked systems. I have had the privilege of working besides many amazing interns (opens in new tab) and collaborators (opens in new tab) in both academia and industry. Our work has received awards such as a Test-of-Time award at SIGCOMM (opens in new tab), best paper awards at VLDB (opens in new tab), NS (opens in new tab)DI (opens in new tab) (twice) and SoCC (opens in new tab). I’ve been a program committee member (opens in new tab) at key venues in computer networking and databases. Technologies that have directly resulted from my research include SWAN (opens in new tab) (which carries over half of the bytes (opens in new tab) between Microsoft’s datacenters), samplers (opens in new tab) and QO pushdown rules in SCOPE (opens in new tab), a distribution advisor (opens in new tab) for Synapse and the outlier mitigation (opens in new tab) technique in SCOPE. I have also been a key collaborator on other technologies that have received wide adoption in Microsoft — VL2 (opens in new tab) (a full bisection-bandwidth datacenter network design), QuickBuild (opens in new tab) (microsoft’s distributed build system) and Sirius (opens in new tab) (which disaggregates stateful network functions into bags of programmable NICs and is in GA at Azure (opens in new tab)). I completed my PhD in Computer Science from MIT (opens in new tab)in 2008 with Prof. Dina Katabi (opens in new tab).

Google Scholar (opens in new tab), Github (opens in new tab), Twitter (opens in new tab), LinkedIn (opens in new tab)


Recent projects: SWAN (opens in new tab), Lazy approximations (opens in new tab), Cluster scheduling (opens in new tab)

Past projects: Seawall (opens in new tab), Flyways (opens in new tab), CloudCmp (opens in new tab), Netmedic (opens in new tab), Broom, EXpose (opens in new tab), VL2 (opens in new tab), FatVAP (opens in new tab), Sherlock (opens in new tab), TeXCP (opens in new tab), Flare (opens in new tab), Kill-Bots (opens in new tab)