(Message inbox:45) Return-Path: Received: from waffle.cs.ucl.ac.uk by bells.cs.ucl.ac.uk with local SMTP id ; Thu, 2 Mar 1995 09:49:33 +0000 To: mmn@cs.ucl.ac.uk cc: J.Crowcroft@cs.ucl.ac.uk Subject: Draft Minutes of Meeting of 1/3/95 Date: Thu, 02 Mar 95 09:49:26 +0000 Message-ID: <1036.794137766@cs.ucl.ac.uk> From: Jon Crowcroft comments to me ------------------------------------ Draft Minutes of BT URI MMN Technical Meeting of 1/3/95 Jon Crowcroft (draft of 2/3/95) 1. Attendees Jon Crowcroft UCL j.crowcroft@cs.ucl.ac.uk James Cowan UCL j.cowan@cs.ucl.ac.uk Ian Marshall BT marshall@drake.bt.co.uk Andrew Grace BT apgrace@jungle.bt.co.uk Morris Sloman IC m.sloman@doc.ic.ac.uk David Hutchison Lancaster dh@comp.lancs.ac.uk Francisco J. Garcia Lancaster frankie@comp.lancs.ac.uk Frank Ball Lancaster frank@comp.lancs.ac.uk Steven Simpson lancaster ss@comp.lancs.ac.uk Avril Smith Oxford Brookes avril@dsrg.brookes.ac.uk Xu Hui Chung Oxford Brookes hui@dsrg.brookes.ac.uk Nat Pryce IC np2@doc.ic.ac.uk Damian Marriott IC dam@doc.ic..ac.uk Iain Phillips Loughborough i.w.phillips@lut.ac.uk Derek McAuley Cambridge dm@cl.cam.ac.uk Geoff Tagg Oxford Brookes agt@dsrg.brookes.ac.uk Mick Mulvey BT mulvey_m@bt-web.bt.co.uk Julian Hill BT jhill@drake.bt.co.uk 2. Introduction Due to travel problems, the BT contingent arrived late. The partners discussed a) UCL's experience with the PNO pilot [a report has been posted to the MMN list and separately to BT]. b) SuperJANET access problems. 3. The agenda was agreed. Each of the partners presented a brief report of the work so far in the WP they led: WP2 Policy based Traffic Management UCL Jon Crowcroft gave a rundown of ATM practical work at UCL, followed by a description of the Policy work - both the case study and the language design work/ WP1 Configuration management. IC Morris Sloman described IC work on configuration, and on policy with UCL. There was some discussion on toolkits, languages and so forth. WP3 End system QoS management. Lanc Lancaster gave 4 presentations, on i) QoS standards work ii) Session Acceptance Models iii) the "Service Factory" model iv) Steven Simpson (BT/EPSRC Case PhD student) work on SMDS performance monitoring/analysis. WP4 Traffic monitoring and Measurement. LUT & OBU Iain Phillips gave an overview of the monitoring work that LUT had been carrying out in agreement with BT. Some very nice graphs of SMDS delay distributions were presented. [Jon Crowcroft cited the "End to End Packet Delay" paper by Jean Bolot of INRIA from ACM SIGCOMM '93 as worth a look to further this work]. WP5 Security. Cam Derek McAuley reported on the secure work. They had gone into some very fundamental thinking about the nature of identity and proof, before moving on to applying it in signaling. 4. Divergence Crowcroft for the record, raised the question of divergence from workplans. BT agreed that as long as this was in each case in line wth discussion with the BT technical person, and was reported at the management meetings, this was exactly how they wanted the project to proceed. There was no need perceived to report this to JISC, since their role in evaluating the technical work was only at the selection of the project at the start. 5. Problems for BT to raise with JISC Many of the partners had had problems getting access to the network or information about the network in many ways. 1/ ATM access at the level of managing traffic and having detailed information about cell behaviour had been continually thwarted. Cambridge, IC and UCL have all had various partial sucesses with national access, and better success at international (PNO)_ access. However, whenever a coherent plan had been presented to UKERNA (via the HS-Net and LLTAG's workign groups on ATM), it seemed that UKERNA would adopt the proposed strategy for themselves. For example, in the first part of the project, the GDC (used to be Netcomm) DV2 switches that carried the SuperJANET video pilot were incapable of carrying variable bitrate ATM traffic abouve very modest rates so we were unable to use VBR video or other experiemtnal traffic - in fact ,the 140Mbps PDH net was solely used for a 2Mbps CBR video network since anything else disrupted this. In the next phase, during the SDH rollout, we proposed using the spare PDH tribs at UCL, IC and Cambridge for a private ATM net using Fore ASX200 switches we had got from EPSRC funding. This idea was so sensible, that UKERNA went and bought their own switches and fielded them, leaving no spare tribs. In the meantime, the design flaw in the DV2s has probably been corrected, so a feasiable (though unsatisfactory) line of pursuit is to try using these with a 10Mbps VP network between UCL, Cambridge and IC, hoping that we do not disturb the video network. None of this is made easier by the UKERNA management, or by the fact that the pilot service is managed by Edinburgh University Computing Service, who, although extremely competent, have other priorities and are somewhat distant from the partners. If we do not resolve this and have a workable high speed multiservice network (as opposed to two single service nets, the DV2/PDH net supporting 2Mbps vidoe and the SDH/ASX200 net supporting IP) in the very near future, we shall have to reconsider our research directions (basically, keep all ATM work local). Similar problems getting information had been reported by Lancaster and LUT on the SMDS. On a very constructive note, BT reported that there are two networks inside BT labs, a "Captive SMDS" network, and the "futures testbed" (probably with SuperJANET and PNO access). If we put together detailed experiments and need a controlled environment to compare with the open setup on SuperJANET, then we can ask BT to use these networks. 6. Year 2 (and 3) Work Plans It was announced that the MOU is now signed (though not sealed and delivered by a couple of partners). Several partners reported that even with this, it would still be helpful in staff recruitment and retention to have 2 or 3 year rolling funding contracts. On the technical front, it was agreed that within the scope of this project, we should not propose andy major changes in areas (e.g. charging or call routing), but ratehr should build in the sketch of the year 2 workplan, and extend the existign packages and directions given by BT in year 1. 7. Time, Date and Place of Next Meeting Project Management Meeting Tuesday 28th March. 10 am. Room 433, Huxley Building, IC Techncial meeting dates tba. 8. AOB It was noted that several of us are speaking at Networkshop (e.g. Crowcroft on March 29), and aso at the EPSRC workshop on 17/18 May. Funding opportunities: for those who havn't applied for a ROPA the last time, there is another opportunity due date 15 March. The EPSRC MNA call was announced. Proposals due by 1 July. Everyone is no doubt busy til march 15 on Framework IV (Telematics and/or ACTS proposals). UCL are involved in 18! Ritchie reported that there had been 51 1 page proposals in the Applications Projects second round. 9. Actions Arising 1/ UCL/ALL To obtain slides from presentations online 2/ BT to raise questions with JISC at management meeting on 4/3/95 3/ ALL to propose to BT any experiements they wish to carry out in conrolled SMDS or ATM environment 4/ BT to investigate whether a 2 or 3 year rolling contract could be placed at the end of year 1. 5/ BT/UCL to make sure JISC rep (Ron Rogerson) is at Management Meeting. 6/ All to provide technical input for year 2/3 workplans (1 side A4 suffice) before the management meeting 7/ Sloman will contact people at IONA with a proposal for a deal over Orbix for the relveant partners. BT will investigate support for this.