150 likes | 161 Views
Read insights from the HTASC report on Video Conferencing recommendations for HEP institutions to improve collaboration. Also, discover recommendations on Windows 2000, network security, and Object-Oriented programming experiences.
E N D
HTASC - Report to HEP-CCC David Kelsey, RALd.p.kelsey@ rl.ac.uk12 November 1999, CERN (http://home.cern.ch/~eauge/htasc/public/) D.P.Kelsey, HTASC report
HTASC #13 28th/29th October 1999, CERN Agenda included: • Minor items • Review Video Conferencing recommendations • HEPNT/Windows 2000 sub-group • Update on UK PIPVIC2 project (Alan Flavell) • Computer/Network Security (see Tobias Haas, later) • Experiences of ‘OO’ programming • Networking: QoS and Differentiated Services • Future meetings/topics D.P.Kelsey, HTASC report
HTASC Recommendations on Video Conferencing (June 99) 1. VC is an essential tool for HEP collaboration and all efforts should be made to make this available to and easy to use by all HEP personnel. 2. Every HEP Institute should be equipped with, or have adequate access to, an ISDN/Codec circuit-based VC room, or roll-about equipment. More equipped VC meeting rooms - if possible with both circuit-based (ISDN) and packet-based (Internet) technologies - are required at the major labs where the lack of rooms is already causing problems. D.P.Kelsey, HTASC report
Recommendations on Video Conferencing (2) (June 99) 3. Equipping a desktop computer for VC is now relatively cheap. All HEP staff requiring desktop VC for their work should be provided with this equipment. VC should be seen as a standard part of the HEP desktop. 4. The various VC standards and tools are still evolving rapidly. R&D work on VC, such as the Caltech/CERN development of VRVS, should continue to be supported and encouraged, to ensure that the tools are suitable for the evolving HEP use and easy to use. D.P.Kelsey, HTASC report
Video Conferencing • Request from HEP-CCC June 99 meeting • HTASC recommendations should be phrased in a way that convinces Management that VC saves money • HTASC discussed this issue again • HTASC does not wish to argue that VC saves money • Use of VC is growing (see VRVS statistics) • so many have already voted with their feet • The most important features are • it saves time • it makes collaboration more possible/efficient D.P.Kelsey, HTASC report
HEPNT • In June 99, HEP-CCC/HTASC agreed to disbanding HEPNT as a closed sub-group, but... to continue as an open forum (like HEPiX) • HEPNT met jointly with HEPiX at SLAC from 4-8 October 1999 (see http://www-project.slac.stanford.edu/hepix/home.html) • This attracted ~90 attendees. Many stayed all week • Success! - so will be repeated in April 2000 D.P.Kelsey, HTASC report
Windows 2000 • Major topic of the Oct 99 HEPNT meeting • HEP-CCC already agreed to the creation of a new Windows 2000 Coordination sub-group of HTASC (June 99) • Many have expressed interest in joining • BNL, CERN, DESY, FNAL, France, JLAB, Italy, RAL, SLAC, UK, ... • But, still searching for a Chairman/Coordinator! • The group must start work very soon. D.P.Kelsey, HTASC report
W2000 Coordination groupDraft mandate 1. To investigate and test the new features of Microsoft's Windows 2000 operating system, with particular emphasis on those issues related to the migration from Windows NT V4.0 which may need coordination across HEP. 2. To make recommendations to HTASC/HEPCCC on those areas where a coordinated migration plan is required. These plans should take into account any potential benefits for the integration of Windows 2000 and UNIX as used in HEP. 3. To share the expertise gained with other HEP Windows NT system managers by organising an open Windows 2000 migration workshop and/or by other appropriate means. 4. To collaborate with the HEPiX and HEPNT groups. D.P.Kelsey, HTASC report
Experiences of Object Oriented programming • What lessons can be learned from use of ‘OO’ by current experiments? E.g. for training requirements • See slides from Neil Geddes (RAL/BaBar) • Discussion/Conclusions • too early to tell whether OO helps maintenance • An infrastructure (people and tools) is needed to support the core developers • HTASC wishes to consider experiences of other experiments D.P.Kelsey, HTASC report
‘OO’ Training Conclusions • Training is essential • in OO design for the core software team • needs to be within experiment’s environment and at the right time. Also needs to continue. • abstract training is of much less use • HTASC members reported lack of funds - ideally to be chanelled through the experiments • there is a lack of suitably qualified trainers • Three levels of training for users • general introduction • ‘hands on’ - use of experiment’s applications • new people need colleagues to ‘hand hold’ (mentoring) D.P.Kelsey, HTASC report
Networking: Quality of Service and Differentiated Services • See slides from Olivier Martin (CERN) • QoS at CERN • See slides from Tiziana Ferrari (INFN Bologna) • TF-TANT and Diffserv D.P.Kelsey, HTASC report
QoS: Conclusions (1) • QoS is an important and fast moving topic. • It is by no means certain that this will ever become a general service over the public Internet. • HEP involvement in tests of QoS should be encouraged. • HEP does have special requirements for QoS, e.g. for remote control of experiments, in addition to the more general requirements, e.g. Video Conferencing and Voice/IP. • Experience of a Managed Bandwidth Service (4 Mbps) between Prague and CERN has shown that this improves interactive performance significantly. File transfers over this PVC, however, are in general no faster than over the public infrastructure. D.P.Kelsey, HTASC report
QoS: Conclusions (2) • While some form of QoS is likely to be very useful for some applications, we should warn that if it does become available it will cost money! The likely funding mechanisms are not well defined as yet. • It is too premature to make any technical recommendations at this time. • HTASC should look at this topic again in 12 to 18 months time. D.P.Kelsey, HTASC report
Future HTASC meetings/Topics(provisional dates/plans) • 2/3 March 2000 or 9/10 March 2000 (CERN) • HEP Directories (LDAP) • Revisit ‘OO’ (FNAL Run II + current CERN expts) • First report/plans from Windows 2000 sub-group • 8/9 June 2000 (not CERN - a Regional Centre?) • Issues for Regional Centres (e.g. for several expts) • 19/20 October 2000 (CERN) • Other possible topics: • Markup Languages, Digital Signatures/Certificates, Linux coordination, revisit Security, revisit QoS • Any other ideas? D.P.Kelsey, HTASC report
Summary • HTASC invites HEP-CCC to • note our conclusions on ‘OO’ training • note our conclusions on ‘QoS’ • suggest future topics for HTASC consideration D.P.Kelsey, HTASC report