Description:
Princeton Docket # 15-3153-1
Researchers at Princeton University, Department of Computer Science, along with researchers at ETH Zürich have proposed a solution to speed up the convergence time following a network equipment failure.
Analyses show that an increase in network delay of a few hundred milliseconds can translate into millions of dollars of losses, therefore high network availability is especially important for those companies whose revenue directly depend upon network connectivity. Most network devices deployed today converges slowly upon link or node failures. On forwarding tables found in the off-the-shelf, ordinary routers, the convergence time of a particular destination can be as high as 2.5 minutes. The core of the issue lies in the time it takes for an IP router to update its forwarding table or FIB. Most routers use a “flat” FIB design which is highly inefficient for convergence as it requires the router to update each entry individually upon failure. One solution is to use a “hierarchical” FIB design for top of the line routers. However, routers with hierarchical FIB tend to be very expensive as they require dedicated hardware support. Also, to be the most useful, hierarchical FIB should be available on all edge routers, which are in the order of hundreds in large networks.
This innovation describes a network design which employs existing routers and low-cost SDN switches to provide any router with a hierarchical FIB. With this invention, any router can converge within 150ms, irrespective of the number of destinations it carries. It also proposes a novel online algorithm for control plane organization and data plane organization using any routing protocol. This design can be used with any legacy router to boost its capabilities. There is no need to reconfigure the legacy router or purchase a whole new expensive fast-converging router. It represents a cost effective approach to dramatically improve network connectivity.
Applications
• IP Router Performance Boost
-Reduce disruption period following Network equipment failure
Advantages
• Fast convergence
• Easily deployed
• Cost-efficient
• Compatible with any router
• Can utilize any routing protocol
The Inventors
Laurent Vanbever is an Assistant Professor at ETH Zürich where leads the Networked Systems Group (NSG) since January 2015. Before that, Laurent was a Postdoctoral Research Associate at Princeton University where he collaborated with Prof. Jennifer Rexford during two years. He obtained his PhD degree in Computer Science from the University of Louvain in Belgium in 2012. Laurent has won several awards for his research including: the ACM SIGCOMM Doctoral Dissertation Award (runner-up); the University of Louvain Best Thesis Award; the SIGCOMM 2015 best paper award; the ICNP 2013 best paper award; and two IETF Applied Networking Research Prizes for his works on inter-domain routing.
Michael Alan Chang is an undergraduate computer science major at Princeton University. At Princeton, Michael works with Professor Jennifer Rexford primarily on software defined networking research. Before researching in software defined networking, Michael was involved in network security research with Princeton Professor Prateek Mittal. Outside of the academic community, Michael has interned at Big Switch Networks and HP Networking, where he worked on network fabric analytics and led an effort to utilize machine learning on network traffic.
Thomas Holterbach received his Master degree in computer science from the University of Strasbourg, France, in 2014. After his graduation, he interned for half a year at Internet Initiative Japan, one of the largest Internet service providers in Japan, where he performed Internet measurements using large-scale measurement platforms. He is now working as an assistant researcher at ETH Zurich with Professor Laurent Vanbever on Software-defined Networks.
Intellectual Property & Development status
Patent protection is pending.
Princeton is currently seeking commercial partners for the further development and commercialization of this opportunity.
Contact:
Michael R. Tyerech
Princeton University Office of Technology Licensing
• (609) 258-6762• tyerech@princeton.edu
Xin (Shane) Peng
Princeton University Office of Technology Licensing
• (609) 258-5579• xinp@princeton.edu