Ken Birman, Robbert van Renesse, and Hakim Weatherspoon are working with industry to develop new fault-tolerance and scalability options for the world's fastest network routers. This effort is exploring ways of migrating some of their prior work on distributed computing into the router itself, so that applications such as BGP can scale better and handle component failures gracefully, and new kinds of functionality like overlay networks or VOIP can be executed right on the router fault-tolerantly. A goal is to compete with telephony levels of availability: the router must be “continuously on” even when handling a crashed node. Weatherspoon is also looking at packet loss dynamics on the National LambdaRail, an optical network spanning the country. Andrew Myers and Emin Gun Sirer developed Trickles (NSDI 2005, ACM TOCS 2008), a stateless network protocol stack. Traditional operating system interfaces and network protocol implementations force system state to be kept on both sides of a connection. Such state ties the connection to an endpoint, impedes transparent failover, permits denial-of-service attacks, and limits scalability. Trickles is a novel TCP-like transport protocol and a new interface to replace sockets that together enable all state to be kept on one endpoint, allowing the other endpoint, typically the server, to operate without any per-connection state.

Cornell faculty has done extensive work in the Peer-to-Peer networking area, ranging from file sharing to media streaming to network monitoring. Van Renesse and Birman developed the highly scalable Astrolabe network monitoring system (IPTPS 2002, ACM TOCS 2003), now used at a major e-retailer. Emin Gun Sirer has created a large number of P2P systems, including Antfarm (NSDI 2009), a content distribution system based on managed swarms, Octant (NSDI 2007), a system for geolocation of Internet hosts, and Corona, an Internet-scale Publish-Subscribe system (NSDI 2006). Van Renesse developed Fireflies, a Byzantine-tolerant P2P overlay network (Eurosys 2006). Weatherspoon designed and implemented the Antiquity system, a secure P2P storage facility (Eurosys 2007). Birman's group is working on Web2.0 collaboration. This effort is looking at the challenges of using Web 2.0 technologies (mashups) in support of demanding collaboration applications, such as one sees in the military or in hospitals. All sorts of hard security, privacy and storage issues arise, and they're studying how best to solve them and how to scale the solutions up for really wide adoption. As part of this, they've built a mashup technology of their own, Live Objects (ECOOP 2008, Middleware 2008), but the hope is to end up with technology that would also apply to Google's Wave, Microsoft's Silverlight or other mashup solutions. Have a look at the demo.

Cornell is particularly well-known for its foundational and practical work on fault-tolerant distributed systems. Fred Schneider's oft-referenced State Machine Replication tutorial is standard fare in systems courses around the world (ACM Computing Surveys 1990). Van Renesse and Schneider formalized and analyzed the Chain Replication paradigm (OSDI 2004). Ken Birman's ISIS system (SOSP 1985, SOSP 1987) has been extensively used in industry for building fault-tolerant systems. Birman and Van Renesse subsequently built fault-tolerant middlewares include Horus (Comm. ACM 1996) and Ensemble (SOSP 1999). Currently, Van Renesse is investigating various aspects of tolerating Byzantine failures. For example, Bosco (DISC 2008) is a Byzantine consensus protocol that decides in one round under favorable conditions. Van Renesse and Schneider are investigating building robust distributed systems based on stepwise refinement. Nysiad (NSDI 2008) is a system that implements a new technique for transforming, through stepwise refinement, a scalable distributed system or network protocol tolerant only of crash failures into one that tolerates arbitrary failures, including such failures as freeloading and malicious attacks.

A relatively new effort at Cornell is Energy Aware Systems. In the area of low-power sensor networking, Van Renesse has looked a power-aware epidemic protocols (SRDS 2002). Hakim Weatherspoon and his students are looking at designing datacenter storage systems that are frugal with energy use. The KyotoFS file system (HotOS 2007) is a log-structured file system. Using multiple disks, only the disk that stores the head of the log has to be spinning most of the time, leading to significant energy savings.