Laboratory for IP Next Generation Orienteering (LINGO)
Proposal for R&E under the HP Internet Philanthropic Initiative
Theme #2
Internet-Uncharted Territory
1. Participants
Dr Soren-Aksel Sorensen (PI)
http://www.cs.ucl.ac.uk/staff/s.sorensen
Department of Computer Science, UCL
Gower St London WC1E 6BT
tel +44 171 380 7296 fax +44 171 387 1397
Professor J Crowcroft (co PI)
Professor Peter Kirstein
Professor Steve Wilbur (HoD)
Dr Steve Hailes
Dr Angela Sasse
Dr Vicky Hardman
Dr Graham Roberts
2. Summary
Since the creation of the ARPANet in the early seventies, when the department was the first ARPANet node outside the United States, it has played a leading role in the development of Internet technology. The expertise build up over the years has been disseminated through our undergraduate and postgraduate programme. The postgraduate programme include both an advanced MSc in Data Communications Networks and Distributed Systems (DCNDS) and research students. The proposal is to update the postgraduate activities by creating a HP Measurement and Modelling Laboratory aimed at exploring the next generation Internet Technology. The laboratory has two objectives. It will give Msc students the opportunity, through a mixture of real and simulated experiments, to study most aspects of network operation and management. At the same time it should provide a forum for research students and staff in which they can explore emerging Internet technologies and find facilities enabling them to perform state of the art experiments.
3. Background
UCL is funded by DARPA, the EC (in the ACTS Prospect, Telematics Merci and ICE-TEL, and Esprit HIPPARCH project) and BT (under the URI and Pearl) to investigate next generation Internet and Internet and ATM through the entire range of applications, down to transmission systems, from routing, QoS support through scheduling, signalling and measurements. The department has also had excellent relation with industry and has often been engaged to solve problems related to performance and network architecture. The relationship is reflected in a number of extensive donations received over the years
We have been teaching the advanced DCNDS MSc for 12 years, and we contribute 2 modules and project supervision on a Telecoms MSc which has been running for 3 years now. These courses produce students with a very high level of expertise who are much in demand. The course includes group projects during which students perform work at a level suitable for publication. A significant proportion of our PhD output has also been in the networking area with special emphasis on the Internet environment.
The research facilities at UCL for multimedia and networking are the best in the UK, thanks partly to the industrial donations. However, although research groups may give access to these facilities, this will be limited to those who undertake relevant projects. Taught Msc students and research students who are not associated closely with existing research currently have less than state of the art systems on which to carry out their work. This limits the possibility for staff to create modern courseware and for research students to keep their work at the forefront of research.
The DCNDS course benefited from a generous donation from HP when the course when it was set up; indeed the original donation of HP workstations for this purpose let to it being set up in the "Hewlett Packard Laboratory". That equipment, however, has long been obsolete; recent HP donations to UCL have focused on specific research projects. The current Call seems to be broader in scope; hence this current response to it.
4. The Proposed System
The facilities would be based around groups of advanced DCNDS MSc students and research students working on networking - naturally with the main theme of advanced Internet technology. The plan is to create a fully integrated environment within which students can experiment with existing and new protocols and test them on a set of real or simulated networks of machines with extensive monitoring and measurement capabilities. The simulation and measurement facility should cater for links, switches and end system performance monitoring. The facilities would be invaluable for staged learning and development. They would allow trace logging, and visualisation of the static and dynamic properties of protocols and networks environments. This would allow faster development, and better understanding of problems related to scaling - a prime area of network research, and particularly Internet research, today.
The proposed Hewlett Packard Network Performance Laboratory would centre around a system running OpenView with the NetMetrix Domain Manager and Reporter together with the Extensible SNMP Agent. The system would run on a HP9000 platform under HPUX 10.2, provided an upgrade path is available from our existing HP?? facilities. Alternatively, Solaris versions of the system could operate on existing SPARC Ultra 1 platforms - but we would prefer to have an all-HP system. This configuration would merge smoothly into more advanced research facilities about to be acquired as part of a donation from HP. The system would be supported by 5 Ethernet LANProbes.
The measurement facility would be complemented by a distributed simulation system for network teaching and research based on the ns simulator (http://www-mash.cs.berkeley.edu/ns/) and the Berkeley Network Animator (nam), both of which are being used very successfully in the department. These systems are being developed in an ARPA project on advanced Internet technology, with which we are also concerned in the CARIN programme. A networked resource based on this software (a web site with common case simulations available via Java applets) and an authoring system both running on hi-end multi-Pentium servers would be set up. User access to these facilities would be through 5 HP Vectra multimedia workstations. The workstations would be running the UCL Multimedia applications, and both IPv4, and IPv6 with RSVP etc. We would ensure that these systems would have access to the LEANET (a fibre-optic network between Cambridge U, UCL and British Telecom being put in for advanced network research) and the ARPA CAIRN IPv6/ATM facilities (through the Arpa-provided UK-US ATM link) for wide area experiments, The measurement system would be integrated with the simulation system, which would use the same source code. The wide-area network system would be connected to this laboratory, to allow detailed statistics of system performance to be gathered on real links or logged to be used as trace files for driving simulations.
The laboratory would allow easy set-up of real world experiments, and symbiosis between work on existing network systems (Internet, SuperJANET, LEANET, CAIRN test beds etc.) and simulations.
5. Timetable and Milestones
This would be resourced from people in the UCL systems group and from the DARPA, Meccano and Merci projects.
This is part of our normal teaching efforts
This entails the largest effort and will be part of the D50 and possibly BV20b group project work of the students themselves (self and mutual group assessment). Academic staff and some our research staff will be heavily involved in this stage.
This would consist itself of an MSc DCNDS project.
We will continue this via our collaborations in the IETF, with NLANR, as well as participation in the TERENA measurement groups.
Routing between the real and simulated protocol entities - this needs a router that takes IP packets from the real and virtual world, and an address assignment system and so on - this could be used to investigate re-addressing and routing problems in the real and simulated world (and IPv4/v6 inter-working) without needing a large real network (but a small real network)
In line with the plans of Section 4, we propose the following HP equipment to set up the new Laboratory:
Summary of Resources Proposed
Please note that Item 8 could be omitted by our using one of our existing Solaris systems.