Chukonu is an ancient Chinese weapon: a repeating crossbow. It shoots out multiple arrows when fires.
We aim to design a similar system for networking in datacenters: simple, parallel, and fast.
Data centers exists because of a standalone server/rack can no longer meet the requirements of modern day applications: web search, ad recommendation, online commerce, machine learning, etc. Different from traditional networks, data center networks enjoy high bandwidth, low latency, and minimal packet loss. These features, however, are not fully utilized today, because application developers are usually unfamiliar with datacenter environment and/or networking stack and its tuning. Therefore, we are motivated to design a networking stack that is highly efficient in data center environments, application-aware, and transparent to the developers.
Our ultimate goal is to lessen the burden of developers, and fully realize the potential of data center networks.
Traffic Prediction for Apache Hadoop
Appeared in IEEE INFOCOM '14 [pdf][ppt]
Also in IEEE/ACM ToN '15 [pdf]
Generic Traffic Prediction for Parallel Computing
Appeared in IEEE ICDCS '15 [pdf][ppt]
Joint Routing & Scheduling for Coflows
Appeared in IEEE INFOCOM '15 [pdf][ppt]
Identifying & Scheduling Coflows in the Dark
To Appear in ACM SIGCOMM '16 [pdf][ppt]