Programmable Networks: Hardware, Software, and Applications
Computer networks are amid a transformation: from delivering packets between two locations to programmatically transforming packets while they are in transit. Our overarching research vision is to make programmability a commodity feature of networks, like connectivity today. Within this vision, I will describe 3 of our research projects---spanning hardware, systems software, and applications. Menshen provides hardware and software support for multi-tenancy on fast packet-processing devices, paving the way for future packet-processing-as-a-service offerings in the public cloud. CaT and Chipmunk leverage program synthesis to provide compiler technology for fast packet processing pipelines found on switches and network-interface cards. Deadline-ordered multicast (DOM) is a new multicast primitive that leverages synchronized clocks. We show how DOM provides a useful building block for many cloud-native distributed applications including consensus protocols and fair-access stock exchanges. Time permitting, I will briefly discuss an ongoing research agenda to build a programmable processor for in-line transformations of remote procedure calls (RPCs). We believe such RPC processors can provide a flexible and high-performance communication substrate for future distributed applications.
Anirudh Sivaraman is an assistant professor at NYU's Computer Science Department in the Courant Institute. His recent research has focused on hardware, system software, and applications for programmable networks. He also works closely with the P4 community, Barefoot Networks (now part of Intel), and Clockwork. His past research includes work on congestion control, network emulation, and network measurement. He received the MIT EECS department's Frederick C. Hennie III Teaching Award in 2012, the IETF/IRTF's Applied Networking Research Prize in 2014, the ACM SIGCOMM Best Paper Award in 2017, the ACM SIGCOMM Doctoral Dissertation Award in 2018, the Amazon Research Award in 2021, and the Google Cyber NYC Research Award in 2023. Before coming to NYU, he received a PhD from MIT in 2017, an S.M. from MIT in 2012, and a BTech from IIT Madras in 2010.