My active research areas

I have a long-standing interest in the monitoring, measurement, characterization and modelling of the Internet. High-speed network monitoring (40Gbps, 100Gbps) is used when networks move to the latest generation networking speeds. Largely, this is to ensure customers, applications and computer-systems can make the best use of the higher speeds. This led to the Nprobe/GRIDprobe projects.

More recently my interests in network-monitoring involve accurately identifying the behaviour of users in project BRASIL.

Work in measurement and in characterizing network use has led to a desire to understand the limits of such measurement and this avenue of work is embodied in a new project Doe-Net.

Effective use of photonics in computer-systems drives my research in a different direction. Why photonics? Photonics (optical) systems have different trade-off concerns in comparison with Electrical systems. An example is how photonic systems have significantly lower costs for distance, lower (potential) costs for energy-needs, lower (potential) costs for heat-production, but nothing is free and photonic systems are subject to physical contraints. The Swallow project umbrellas this work. Basically,

• I want to answer the question under what conditions does a photonic network become better than an electrical network?

• Then build one.

• Want to be involved?

I also love the NetFPGA project. Intended as programmable hardware for high-speed network prototypes it is exactly the sort of hardware to help students come to terms with how all this stuff works. Our local project pages are here.

Things in the past that I've done...

• I got deeply into coding of physical layers - particularly for photonic (optical) links and all-optical networks.

• I built (and build) network monitoring systems and infrastructures (Nprobe/GRIDprobe), included in this has been contributions to CoMo

• I built a tonne of code for a prototype network switch Fairisle II/Utsire this included developing an Admission Control algorithm called Measure and Entropic (entropy based) method. The idea was to do neat tricks that allowed high-level administration of networks so I worked with NCAM too.

• Many years ago I instrumented Ultrix (DEC's UNIX for MIPS) kernels and then looked at the NFS traffic.

Sponsors

I am very grateful to my sponsors who have supported me over the years.

My sponsors include:

	Xilinx
	AT&T Research
	JaNet
	Dante
	Endace Technology
	Intel Research
	Engineering and Physical Sciences Research Council (EPSRC)

Thank you all.

EPSRC Projects:

Current

• INTERNET - Cambridge making a bid for the Internet? no this is Intelligent Energy Aware Networks Project Website
• DOE-NET - Optimal design of performance measurement experiments for complex, large-scale networks Project Abstract

Past

• UKLIGHT Monitoring and Analysis at many Scales GR/T10510/01(P) Project website
• IPv4 and IPv6 Performance and QoS - 46PaQ GR/S93714/01(P)
• A scalable monitoring platform for the GRID (GridProbe) GR/S07759/01 Project website