Draft v1 Minutes of 2nd RAMS Management Meeting UCL CS, April 1, 1995. Actions marked + 0. Introductions Present: Name Affiliation Email Saleem Bhatti UCL s.bhatti@cs.ucl.ac.uk Jon Crowcroft UCL jon@cs.ucl.ac.uk James Cowan UCL j.cowan@cs.ucl.ac.uk Dave Romano-Critchley UCL d.critchley@cs.ucl.ac.uk Zhen Wang UCL z.wang@cs.ucl.ac.uk Tim Snape ASL rams@wdi.co.uk Huon Butterworth Tadlon huon@fchain.demon.co.uk Steve Osselton PRISM steve@prismtech.co.uk Kieran Cooney Texaco udmkxc@texaco.com [Monitoring Officer mike.falla@bcs.co.uk] 1. Agenda The agenda was agreed with no changes, except that we had had an mixup with informing the Monitoring Officer about changes to the meeting (his Compuserve mail address was on the RAMS@CS.UCL.AC.UK distribution list, and appeared not to have forwarded him mail to his BCS account). 2. Minutes of Last Meeting Actions arising: * Common start date is 11/12/95. J Crowcroft will check when the EPSRC start date for the research grant was, and the named staff (Zheng Wang and one other....to be agreed in first week of Jan 1996). In Progress * UCL will setup a maillist at UCL called rams@cs.ucl.ac.uk to which any technical correspondence can be emailed. Done. * If partners wish, they should send J Crowcroft the email addresses they wish addd to this list. These can be local distribution lists (e.g. rams@prismtech.co.uk might be one) if preferred. Done. * travel. We will liase with the M.O. about whether we can have the meetings in london, and he can separately visit partners during our separate technical meetings. Done. * UCL need to confirm this with CVs to EPSRC. In progress. * J Crowcroft will contact Jack Houldsworth to see if we can get a direct project input and output from this process. In progress There will also be an HPIP program newsletter. * UCL should provide an abstract for the second one of these Done. * DTI should make sure ASL are on the maillists for this and EPSRC should add them for their publications (IMPACT). Done 3. Presentation of Level 2 Plan [J Crowcroft] Crowcroft went through the draft level two plan and we checked off the deliverables due so far. UCL - UCL 1 informally delivered at the meeting. Prism - provided their 1st quartlery report. PRIMSM 1 and PRISM 2 will be combined under one cover, and delivered in the next week or so. Tadlon - Some delivery problem had stalled monitoring on the Texaco network, but later this week, they would expedite the measreuments and provide results and analysis. Deliverables 1 & 2 to appear in the next 2 weeks. Abbotsbury - All site deployment well in hand (see http://www.wdi.co.uk/welcome.html) - In fact, work on metrics for performance and for outage analysis (the effect of Mean Time To Repair on Web page access statistics) and even on the initial ideas for the quality document (final deliverable) are in hand. [See http://www.wdi.co.uk/quality/home.html] 4. Technical discussion of work There was a detailed discussion of three areas. i) The metrics to cover the Abbotsbury Web site and Internet provision. Crowcroft presented some ideas on Settlements, a mechanism for levelling charges between ISPs. Snape presented some ideas on a Web Ping program that they had been developing to aid with load balancing. The idea is akin to application level routing and makes use of HTTP Redirect messages to spread load - an information provider would run a set of web sites, and outsoruce their data to other web sites, and would then use the publically auditable WebPing program to gather performance statistics from different subscriber viewpoints to different providers sites. This could then be used as input to the load balancing algorithm used to dereive the redirection tables that would optimise the subscribers latency, and minimise providers web provision costs. Input from the detailed statistics from the UCL-1 document on TCP performance (e.g. retransmission and duplicate statistics, as well as congestion window information) could be correlated with these application level statistics to give a more detailed level view of the underlying ISP problems. ii) A discussion on the resource management work at Prism, and the Idl ensued. iii) Some details of the Tadlon measurement work were gone into. 5. 1st quarter reports PRISM provided their 1st quarter report. Abbostbury and Tadslon's are nearly reader. 6. Claims These are going through the partner's auditors at the moment. 7. Wrapup/Time&Date&Location of Next meeting Crowcroft liased with the M.O, and the next management meeting can be in Early June (1-14), at UCL. Any interim technical meetings should be notified to the list so that the M.O. may be able to decide whether to attend. 8. AOB. UCL to setup a mail list rams-tech@cs.ucl.ac.uk, for discussion of purely technical matters. Tim Snape proposed that when exchanging commercial in confidence network monitoring data bey electronic mail, we should use encryption - the PGP package was proposed - Tadlon, UCL and the others will set this up and see how effective (easy) it is to use. Crowcroft mentioned two events of interest - the Internet Pricing meeting at Cambridge will discuss resource control through effective charging; The Globecom 96 conference at the QE2 Conference center in London in Nov 18-22 will address Java/Web, Internet Security, IP and ATM, and Multimedia on the Internet. Details at http://www.statslab.cam.ac.uk/~frank/PRICE/ and http://www-net.cs.umass.edu/tccc/internet96 respectively 9. Actions + Common start date is 11/12/95. J Crowcroft will check when the EPSRC start date for the research grant was, and the named staff (Zheng Wang and one other....to be agreed in first week of Jan 1996). + UCL need to confirm this with CVs to EPSRC. + J Crowcroft will contact Jack Houldsworth to see if we can get a direct project input and output from this process. + Finalize Time and Date of next full management meeting in Early June at UCL