|Previous page||Table of Contents||Next page|
1. The Information Technology Syndicate develops computing policy for the University and the Colleges and supervises the University Computing Service; this report covers the activities of the Syndicate and the Service for the year from August 2000 to July 2001. The printed version of the Statistical Annex gives only top level summaries of the use of services but breakdowns to the level of individual institutions are available on the World Wide Web at http://www.cam.ac.uk/CS/ITSyndicate/AnnRep/stats00.01.html/. Please note that the statistical information is provided only to give a general picture of the width and volume of Service activities and makes no pretensions to being fully comprehensive.
2. The IT Syndicate again functioned smoothly and effectively under the chairmanship of Professor Longair, while its Technical Committee, chaired by Dr Heath and now including a representative from the Management Information Services Division of the University Offices, continued to provide extremely valuable input to the discussion of any business with a technical dimension. Issues discussed by the Syndicate and dealt with in more detail below include the EastNet regional network, the future development of the CUDN, reducing the impact of hackers by blocking certain types of traffic, a Code of Practice for Computing Service staff dealing with personal information, and plans to avoid running out of IP addresses.
3. Computer staff employment issues continued to be a major concern. At the February Syndicate meeting, the Director of Personnel, who was present with two of his senior staff, outlined the progress made since the previous summer (including agreement in principle to drop the strict age relationship for starting salaries) and future possibilities (including plans to establish a transparent reward structure based on well thought out criteria). Dr Sayers was subsequently asked to contribute to a general review of academic-related staff recruitment and retention by producing proposals for changes to Computer Officer contracts and salary scales, and most of the June meeting was given over to a detailed discussion of his draft paper, which was then modified accordingly and sent off to the Personnel Committee.
4. By the end of the year, the academic side of the Computer Laboratory was ready to move both itself and that name to West Cambridge, while the Computing Service remained on the New Museums Site. This physical separation of the Laboratory and the Service was preceded by the approval of a Report which formally established the Computing Service as an independent institution under the supervision of the General Board with effect from 1 January 2001, including appropriate changes to various regulations and the establishment of two new Appointments Committees for the officers in the Service.
5. During the Lent and Easter terms, the General Board re-allocated the New Museums Site accommodation being vacated by the Computer Laboratory. The Computing Service share is less than was requested (particularly in the Arup Building, where the space is to be shared with the Language Centre) but will meet all short term needs. By the end of the year, in addition to the new (and mostly HEFCE Teaching Infrastructure funded) undergraduate classrooms on Arup Floor 2, the Estate Management and Building Services (EMBS) were preparing to refurbish the Hardware Support workshop and the Service's newly acquired parts of the Austin Building, including the relocation of the Common Room from Arup Floor 2. Thought was also being given to the security and access implications of needing to make the new classrooms (and adjacent Language Centre facilities) available to students around the clock.
6. The facilities and services provided by the Computing Service include:
Services such as networks and electronic mail, which can only be provided on a University-wide basis;
Centrally-managed services such as the Public Workstation Facility (PWF) and the Central Unix Service (CUS), which are available to any member of the University who needs them but provide facilities of a kind which individual institutions often supply in-house to their own staff and students;
A wide range of support services for institutional and individually-owned facilities, from local area networks and personal computers to powerful multi-user computers and scientific workstations;
Advice and consultancy services which include a Help Desk for the centrally-managed facilities, the Techlink support scheme for local computer staff, and a wide range of training courses.
7. The Computing Service is made up of five divisions; two of these manage central facilities and provide specialized technical support of various kinds while the other three provide direct support to users, either individually or by institution.
The Network division is responsible for the University Data Network (CUDN), the Granta Backbone Network (GBN), support for institutional networking, the PWF and the Managed Cluster Service;
The Systems and Unix division is responsible for the CUS, the Thor teaching facility, electronic mail and news services, Unix support to individual institutions, and the Pelican archive service;
The Technical User Support division is responsible for the Help Desk and the Techlink scheme for local computer support staff, training courses, general technical user support, and the Hardware Maintenance and Videoconferencing services.
The User Services division is responsible for information provision, user administration, Software Sales, hardware purchasing advice, and the Printing and Photography and Illustration services;
The Institution Liaison division maintains contact concerning institutional IT strategies and related matters with Heads of Department, other senior managers, and local computer support staff.
|CUS and Thor||144||30||174||4.6%||174||235|
|Other centrally-run facilities||228||10||238||6.3%||28||266||248|
|Other charged support services||290||0||290||7.6%||474||765||722|
|Advice and consultancy||639||49||687||18.1%||687||666|
|Directorate, liaison, and staff eqpt||253||113||367||9.6%||367||398|
|Administration and overheads||182||154||336||8.8%||336||327|
8. The table above summarizes the estimated cost of various types of service provided by the Computing Service, distinguishing central funding (UEF and equipment grant) from costs met by charging users (with combined 2000 figures in italics). The cost recovery services account for £1.57m (29% of the total) compared with £1.58m in 2000 and £1.32m in 1999.
Some of the major developments during the year are discussed immediately below, followed by paragraphs about each of the individual services.
9. The EastNet regional network, which resulted from a successful bid for funding under the second phase of the HEFCE Metropolitan Area Network Initiative, links together seven universities (Cambridge, Anglia Polytechnic, Cranfield, East Anglia, Essex, Hertfordshire, and Luton) plus Homerton and Writtle College (Chelmsford). EastNet has its primary JANET connection at Cambridge (in the Computing Service) and Cambridge was also the lead site in its implementation, although ownership may be transferred to the Association of Universities in the Eastern Region. Following a procurement process during Summer 2000, ntl were awarded a contract to provide the EastNet network services, while a separate procurement for network interface equipment resulted in the use of Cisco routers supplied by Logical UK.
10. Although thrashing out the details of the contract with ntl cost the head of the Network Division much time and effort, their implementation work was completed in early April (only slightly behind schedule) and, after the installation of the routers, EastNet came into service during May. An invitation from the HEFCE to bid against a £860,000 allocation for enhancing EastNET was used both to uprate the routers and to improve the reliability of the power supply for the equipment at Cambridge by installing an Uninterruptible Power Supply backed up by a diesel generator. A contract was negotiated with UKERNA to manage EastNet; the associated costs will be covered by counter payments in respect of UKERNA's use of EastNet to meet its contractual requirements to provide JANET connections to HEIs and FEIs in the region.
11. By the end of January, Worldcom had completed the Cambridge SuperJANET 4 node, installing new fibre cables from the Science Park to the Computing Service, where they were connected to the UKERNA Backbone Access Router. The SuperJANET 4 connection provides a total capacity of 2.5 Gbps for all the institutions in East Anglia; at 1 Gbps for both in-coming and outgoing traffic, the Cambridge share of this is almost two orders of magnitude more than the previous (outgoing only) limit of 12 Mbps under SuperJANET 3.
12. In mid-August 2000, CUDN performance was badly degraded for several days by an intermittent fault which caused major loss of connectivity and was made more difficult to investigate by the attempts of the network's automatic reconfiguration software to respond to the constantly changing situation. The fault caused great inconvenience to Finance Division staff trying to process payroll and other vital functions using procedures which depended critically on good data communications between offices at the Old Schools and computers at the old Royal Greenwich Observatory building (although it was not responsible for most of the long delays then being encountered by other users of the newly live CAPSA project). It was the first significant CUDN failure since February 1999 and during the interim both JANET failures and planned electrical shutdowns had caused far more total loss of connectivity.
13. Following the fault, steps were taken to improve reliability of the links between the Old Schools and the RGO. A more general conclusion, reinforced by similar but much shorter failures in the Michaelmas Term, was that the intermittent faults were related to overloading on the central ATM switch. Existing plans to simplify the design of the CUDN by ceasing all use of the ATM protocol and to convert the backbone network to use Gigabit (1000 Mbps) ethernet were therefore accelerated and by July the backbone network was using IP only, with Gigabit ethernet in use on all main trunk links to the routing nodes and also on most backup ones.
14. However, CUDN development never stands still and in May the University submitted a successful £1.2m bid for its further enhancement using money from the Science Research Investment Fund (with a 25% contribution from the University). To enable the network infrastructure to support the next generation of e-science projects, the central switches and routers will be replaced by higher performance equipment which can use the full capacity of the SuperJANET 4 link and deliver Gigabit speed connections to departments. Using high-performance Cisco routers with built-in redundancy should also improve network reliability.
15. Protecting the CUDN from hackers is a never-ending battle over constantly shifting ground. During the year, attacks became noticeably more automated, and a successful probe would immediately use its target system to attack other machines on the CUDN (or in one case to launch 'denial of service' attacks against external targets, effectively paralyzing the Cambridge connection to JANET). Investigating and recovering from these incidents can be very prodigal of staff time. For example, when a single insecure machine was compromised in December and used to probe the entire network for a service normally blocked at the JANET-CUDN network, the result was five more compromises, all in different institutions. Users at each institution had to reset their passwords and the overall cost of the incident was at least thirty person-days of Computing Service staff time, plus that of the local computer support staff.
16. Incoming traffic was already being blocked at the JANET-CUDN gateway (with a few exceptions for host destinations with valid needs) on several ports not much used by ordinary users but popular with hackers. However, between October and December, so many probes were observed on the ftp (file transfer protocol) port that, despite the fact that it would affect many users, it was decided to block from January 22 all incoming requests for ftp transfers in either direction. The blocking was widely advertised to users during the Christmas vacation, staff were primed to provide advice on alternative procedures and a few exceptions were made for user systems with a large need and a good security record. In the event, users adapted to the ftp blocking much more readily than was first feared and its introduction caused very much less overall disruption than if the probing had been allowed to continue unchecked.
17. For the rest of the year, external probing continued at alarming rates but the blocking kept it at bay apart from a few small and well-contained security breaches. From May, web servers running Microsoft IIS were attacked by the Code Red worm, which became more of a threat with each month's new variant. For much of late June and July, major denial of service attacks sent vast amounts of unwanted transatlantic traffic to certain Computing Service and other cam. ac.uk addresses. Attacks on this scale caused notice-able SuperJANET degradation and after consultation between the local and JANET Computer Emergency Response Teams countermeasures were installed at the US end of the transatlantic links to limit their effect.
18. Previous reports have described Computing Service plans to comply with this Act by the time the transitional exemption for existing computing activities expires in October 2001. In December, the Service produced a Code of Practice for its staff which sets out procedures for dealing with various kinds of request involving personal data. The Code takes into account not only the Data Protection Act 1998 but also other relevant recent legislation including the Human Rights Act, the Regulation of Investigatory Powers Act, and the Telecommunications Regulations 2000. After approval by the IT Syndicate, the Code of Practice was made available on the web for the benefit of University institutions and Colleges facing similar problems.
19. If no changes are made to the way in which Cambridge IP (v4, 32-bit) addresses are currently allocated, they may well run out in two years or so (i.e. well before IP v6 with 64-bit addresses comes into widespread use). Following the Technical Committee's consideration of how to use the existing total Cambridge allocation more effectively (IP v4 addresses being currently in such short supply world-wide that asking for more is unlikely to succeed), the Computing Service has agreed to produce an action plan. Possible strategies include using Network Address Translation for new student users, freeing up sparsely allocated address ranges and creating a strategic reserve from addresses currently allocated to the Computer Laboratory. The practicability of connecting College rooms via an Internet Service Provider (ISP) rather than the CUDN while maintaining student access to appropriate cam.ac.uk facilities will also be investigated.
20. Network division staff again handled all operational matters for the GBN Management Committee. During the year, as part of EastNet, Homerton finally obtained its long planned GBN connection. At West Cambridge, the re-routing of GBN cables to allow the access road to be widened was completed, also the GBN node in the new Computer Laboratory building. Following the completion of the Unilever Centre, cables were installed on the new GBN route into Chemistry and the re-instatement of diverted services was put in hand. Building work at New Hall necessitated a move of the GBN cabinet and cables to the opposite end of the College, although staff managed to limit downtime on the CUDN connection to a few minutes. Finally, discussions which are in progress about the re-routing of GBN ducts and cables because of separate building projects at the Sidgwick Site and at Gonville and Caius Harvey Court aim to limit overall disruption by agreeing in advance one new route across both sites.
21. Almost all of the network now consists of ethernet connections, either at 10 Mbps or less (slow) or at 100 Mbps or more (fast); these institutional links are connected via eight area routers to the Gigabit ethernet CUDN backbone network (see above). Uninterruptible Power Supplies were installed at two of the area routers (North and West) and when their performance has been assessed, more may follow. During the year, the number of fast connections increased from 48 to 76 while that of slow ones fell from 94 to 73 (for fuller details see the Annex). In general, each of these connections will support one or more local area networks (LANs), thereby interconnecting equipment in a whole building or even a whole Department.
22. Around Easter, the first Gigabit ethernet connections to individual institutions were installed and both new connections and upgrades to this speed now appear as standard items on the CUDN price list. Virtual Private Network tunnelling products were trialled, with the aim of allowing the same secure authenticated access to CUDN services via other Internet Service Providers, as if the client were within cam.ac.uk. Total registrations for the Magpie modem service were again just over 2,000 and the existing lines and modems handled the load without congestion; most subscribers now use V.90 modems at 56 Kbps, and so notice was given that the older and slower V.34 and SLIP service will be withdrawn from April 2002.
23. While the ethernet connections data shows how the CUDN now reaches into every corner of the University and Colleges, the true size of the network is most obvious from the total of Internet addresses. This was 42,900 at the end of 2000-01 (up from 36,500, 29,500 and 18,900 in previous years), with 28,000 addresses in University institutions and 14,900 in the Colleges. The University ones were used not only for personal computers and workstations but also for many servers and other equipment while the College addresses included over 7,500 active connections in student study bedrooms. The July introduction of devolved IP address registration will allow College computer staff to deal with the Michaelmas 2001 rush without the previous need to involve Computing Service staff.
24. During the year, Network Support staff provided advice to individual institutions on over 210 occasions (40 more than in 1999-2000); typical topics included wiring up buildings, establishing LANs, and obtaining CUDN connections. The Network Installation team was just as busy as in previous years and carried out or supervised work which included 30 new fast ethernet installations or upgrades, new network equipment installations at 11 sites, optical fibre installation work at 25, copper wiring at over 30, and testing of existing cables at 5.
25. Individual institutions were again charged a share of the cost of incoming transatlantic SuperJANET traffic. Total traffic was 60% more than in 1999-2000, but the overall charge to Cambridge had been fixed in advance at £113,651 (down from £132,580) and so the top charge to any individual institution (all of the top six and eight of the top ten being Colleges) was again only £6,300. The same charging mechanism will apply in 2001-02, with Cambridge as a whole paying £119,968, but major changes are expected from the following year. The Computing Service continued to supply Colleges on request with snapshot monitoring information to identify individual cases of large use, but remained unable (for reasons outlined in last year's report) to produce reliable figures for the total traffic to individual IP addresses.
26. Since in the previous year saturation of the outgoing 8 Mbps link to JANET had caused considerable concern, an increase to 12 Mbps from the start of the Michaelmas Term was purchased from UKERNA. This, combined with the use of control mechanisms to limit the impact of outgoing recreational traffic at busy times, avoided any serious problems until the greatly increased bandwidth provided by SuperJANET 4 and EastNet came on stream (see above). The control mechanisms were then removed, but the immediate effect was that increased recreational use caused outgoing traffic volumes to match or slightly exceed incoming ones, so some of the controls were reinstated.
27. As a result of the advance integration of additional equipment, the Hermes mailstore system coped well with the usual sharp rise in demand at the start of Michaelmas 2000. The total volume of mail delivered by Hermes increased by 60% during the year (from 320 GB to 520 GB) after more than doubling in the previous one, while the number of messages increased by 15% (from 49.1 M to 56.6 M) after a 50% rise in the previous year. The number of users of Hermes rose by 1,800, from 26,100 to 27,900, although some of these were existing e-mail users transferring over from CUS. For much of the year, staff illness and other problems limited software work to routine maintenance and local improvements to deal with increases in load but towards the end work was resumed on essential long term software developments. The first to appear is the Web interface to Hermes which, after needing to be coded specially to achieve the required scale of operation, should go into pilot service during Michaelmas 2001.
28. For Michaelmas 2000, the data management facilities on the Computing Service web servers were improved and the pages about the Service were re-designed to direct users more effectively to the relevant electronic and paper documentation; this significantly increased the number of accesses. New material installed on the University server included the Press Office's Media Guide to Expertise (using material collected via an online database) and the 2002 Undergraduate Prospectus, including departmental contact information. By the end of the year Information Group staff had also completed the task of lower-casing all the urls on the server, to simplify access for both users and maintainers. The Technical Librarians dealt with 5,644 enquiries to Webmaster (32% up on the 4,267 in 1999-2000), in addition to maintaining the staff library and setting up user libraries for new Managed Cluster Service installations.
29. The server hosted pages for over 180 University societies (150 in 1999-2000), while 42 (25) bodies used the managed web server service for institutions which can author pages but do not wish to run their own server. A pilot web-based version of the University map was released to general approval after Systems and Unix staff had solved the technical problems of presenting a complex map on relatively low-resolution web browser screens while including features such as zoom and locate. The Information Group incorporated the map data (derived from the same base as by the University Press's paper version) and will be liaising with the Press about future updates. Staff also continued to provide input to the University Web Site Steering Committee, and to organize useful and well attended meetings of local webmasters.
30. The Computing Service University-wide web search engine uses commercial Ultraseek software to produce a single index of the contents of all the more-or-less 'official' servers within cam.ac.uk. This allows outside requests to locate material without the need to know or guess in advance which server may have it, also to make complex searches, and to search within individual servers. The following table compares use in July in 2001 and July 2000:
|July 2001||July 2000|
|Servers, documents indexed||321, 401 K||269, 267 K|
|Total searches, data sent,||111 K, 2.2 GB||90 K, 1.8 GB|
|% requests from Cambridge, rest of UK, rest of world||25%, 15%, 60%||22%, 15%, 63%|
|Originating computers: University, rest of world||4.5 K, 25.5 K||4.4 K, 21.1 K|
|July 2001||July 2000|
|Total requests received, total data sent||5.7 M, 29 GB||3.7 M, 17 GB|
|% requests from Cambridge, rest of UK, rest of world||33%, 8%, 59%||42%, 10%,48%|
|Originating computers: University, rest of world||12.3 K, 150.6 K||10.7 K, 118.0 K|
|July 2001||July 2000|
|Total requests received, total data volume||122 M, 830 GB||61.8 M, 465 GB|
|% hits (requests), % hits (data volume)||46.4%, 18.0%||48.3%, 19.3%|
|Client systems, client domains||8,559, 161||6,652, 150|
|July 2001||July 2000|
|News server connections: monthly total, peak rate||282 K,102||232 K,164|
|Total, local newsgroups||14.47K, 233||13.7K, 229|
|Monthly volume, number of individual postings||18.5 GB, 8.9 M||19.1 GB, 9.2 M|
31. During the year, there were 66.5m incoming requests (compared with 49.1m in 1999-2000) to the main Computing Service website (http://www.cam.ac.uk/). They came from over 190 country-level domains (180), with a daily peak of 304,000 (216,000). The past expenditure of time and money on monitoring and fallback arrangements paid good dividends over the Christmas break when staff were alerted to and readily circumvented hardware failures which would otherwise have caused a major outage of information services. The table above compares main server activity during July 2001 with July 2000.
32. The local and national Web caches reduce external network use and improve performance by automatically maintaining copies of recently referenced documents. During Michaelmas 2000 there was serious overload on the local cache server, and although some improvement was obtained from changes to the server software, new hardware with additional capacity was clearly needed. The new system uses a cluster of powerful PCs, which not only facilitates the provision of spare components and fallback capacity but also makes the system easier to build and to develop. After a phased introduction starting in late April, the system is already handling a greater load than its predecessor with much improved performance. The table above summarizes activity for the new cache in July 2001 and for the old in July 2000. Because one cannot distinguish individual clients in institutions which route requests to the cache via firewalls or local web caches, the client systems figures will be underestimates.
33. The Computing Service's USENET news server gives Cambridge users access to thousands of local and world-wide news groups, thereby providing them with the opportunity to disseminate information on, and to engage in the discussion of, a very wide range of topics. The table above compares the levels of activity during July 2001 and July 2000.
34. The Public Workstation Facility and the Managed Cluster Service (see below) are supported by the staff of the Small Systems Group and the Operations Group, who planned and implemented two major changes during the year. First, during the Easter vacation, six of the central PWF servers were replaced by new and more powerful hardware and four were upgraded to use Novell NetWare 5.1 with the newer version of Directory Services. This worked well in general, but an unfortunate side effect caused the loss of statistical data on PWF use from 18 April until the end of July. The following table therefore compares data for the first 8.5 months of 2000-01 with (in italics) full year figures for 1999-2000. (N.B. In general, undergraduates are registered as College users and postgraduates as University ones.)
|Public Workstation Facility|
35. During the year, staff continued to roll out a new basic Windows NT image each quarter, managing the provision of multiple variants to different sites by using Zenworks Workstation objects. All PWF Macintoshes were given a new log-in program and upgraded to OS 9.1. When escaping Microsoft NETBIOS traffic caused broadcast storm problems on the CUDN, the PWF Club members installed Netware 4 tunnel blocks to contain it within each institution. Finally, after much planning and testing, the second major change took place just after the end of the year, when a one-day shutdown allowed all PWF and MCS sites to change simultaneously to a new image which uses Windows 2000 Pro rather than Windows NT.
36. The Central Unix Service continued to be widely used by researchers throughout the University, although a significant number of mail-only users moved to Hermes. The related Thor facility supports Unix-based taught courses (primarily in Physics and Computer Science, although a Geography course was included at very short notice). The table below presents summary statistical information for their use in 2000-01, with that for 1999-2000 in italics:
|Central Unix Service|
37. Following the success of last year's pilot project to provide Linux on the Computer Science PWF cluster, Unix Support staff put in over eight man-months of work (at all levels from operating system modifications to applications tuning) to create PWF Linux. This will run alongside Windows and MacOS on dual-boot machines in PWF clusters (and MCS clusters which opt to include it) and can also be used on a dedicated machine to provide students with remote access (using SSH) during vacations to Unix-based PWF teaching material.
38. During summer 2000, substantial developments were carried out to the user database. As part of the pre-registration of 7,000 new postgraduates and undergraduates for the start of Michaelmas 2000, a new web-based facility was introduced to allow those who could establish their identity by supplying appropriate personal information remotely to be sent details of their e-mail address in advance. Although only limited password resets were required this year, surprisingly many users were out of the country and had to prove their identity over the telephone to get their new password; arrangements have therefore been made to allow such people to nominate in advance someone with authority to act on their behalf if necessary.
39. Thanks to increased network security activity (see above), the total number of incidents requiring investigation was 733 (45% up on the 506 in 1999-2000). Many of the non-security incidents involved unwanted e-mail (which often appeared to come from a Cambridge user but had actually been forged by an external user); another common type, especially in Michaelmas, was internal probing caused by inexperienced students with incorrectly configured machines.
40. The new disc farm based Pelican data archive service continued to operate efficiently, with considerably faster retrieval times than the tape based system it had replaced. By the end of the year, the total data stored was 200 GB (20% up on last year's 166 GB) for 920 different users (949) and it was clear that more discs should be added very soon. Although Pelican was used to some extent by more than eighty University and College institutions, the largest users by volume continued to be the Computing Service (for backup of programs and data) and the University Library (for various projects including the storage of scanned photographic data).
41. The Computing Service continued to provide the High Performance Computing Facility (which comes under a separate committee of management) with the services of one of its senior programmers who has a particular interest in advanced computing systems.
42. Software Sales supplies members of the University with competitively priced software through CHEST (Combined HE Software Team) and other deals and administers University-wide support agreements for Sun and Silicon Graphics workstations. After much consideration of how best to reconcile user demands for faster ordering and delivery with the need to stay within the terms of the different licences and to ensure that users are aware of the University Software Policy, the web software product catalogue was supplemented by a web Ordering System. This allows users who have already signed up to the general terms and conditions to create and print order applications on-line, although the need to accompany them with a purchase order (currently only available on paper) then slows things down to the speed of the internal mail. The table below summarizes sales for 2000-01, with 1999-2000 in italics.
|Graphics and drawing, statistics||2,226||2,626|
|Office packages, operating systems/utilities||6,510||7,702|
|Programming languages, mathematics/libraries||675||698|
|Word processing, spreadsheets, databases||1,826||2,307|
43. The Managed Cluster Service (MCS) enables institutions to reduce the running costs of their general-access computer rooms by paying the Computing Service to install and manage both the Novell NetWare servers and software for the PCs and Macintoshes, based on the standard PWF image (see above) with optional additional applications. The 17 MCS clusters at the start of the year (ten in Departments and seven in Colleges), had increased by three (at Chemistry (Unilever Building), Selwyn, and Sidney Sussex) by January 2001 and by five more (at the new Computer Laboratory, the Betty and Gordon Moore Library, Clare, Emmanuel, and Trinity Hall) by September 2001. To reduce the total number of servers and hence the work involved in installing new images, these five and all future clusters will use (and share where appropriate) servers which are supplied by and located in the Computing Service.
44. The Engineering Group again provided institutions and individuals with either a full cover contract service or repairs charged on a time and materials basis and also carried our warranty repairs on behalf of certain manufacturers for their University customers. The tables summarize repair work carried out during 2000-01, with figures for 1999-2000 in italics.
|Equipment repaired by Type|
|IBM PC &c.||147||125||138||152|
|Equipment repaired by Institution|
45. The existing Xerox Docutech and Xerox and Agfa colour machines continued to work well. During the year, the Docutech service produced a total of 0.91m impressions for the Computing Service (up from 0.75m in 1999-2000), 0.81m for the Computer Laboratory (0.72m) and 0.60m for external customers (0.52m), while the colour copy totals were: Service 7,000 (12,000); Laboratory 1,400 (1,000); external customers 7,200 (13,500). The lease on the Xerox equipment expires in December 2001 and plans are in hand to obtain replacements more suited to the reduced requirement following the departure of the Computer Laboratory.
46. Although the traditional photographic activities had a rather reduced throughput of 7,300 rolls of E6 film (compared to 8,800 in 1999-2000), 133,000 colour prints (160,000), 3,300 duplicate slides (5,000) and 3,400 slides made up from artwork supplied by customers (3,700), the number of posters produced doubled from 1,000 to 2,000 and a new service to scan 35mm film into digital images on CD processed almost 300 rolls. Staff took photographs for sale at General Admission and other Congregations, both inside the Senate-House and on the Lawn, and also carried out specialist photographic work for Departments and Colleges on request. Equipment purchases included a 60**L375** wide inkjet printer to keep pace with user requirements for larger posters, a more modern display board system for hiring out, and new digital minilab film processing equipment which will be in service by Michaelmas 2001.
47. The staff of this service spent much time during the year using their considerable expertise in both ISDN and IP (room and PC-based) videoconferencing to advise those University Departments (such as Education, Engineering, Mathematics, and the Veterinary School) which wanted to move on from extensive use of our facilities to buying their own equipment. As a result, despite an increasing overall interest in videoconferencing throughout the University, total use of the Computing Service suite for ISDN videoconferencing was only 132 hours in seventy-two bookings (compared with 170 hours in eighty-seven bookings in 1999-2000.) To keep abreast of new developments, staff purchased IP desktop cameras for trial purposes and spent two weeks in July, in association with the Cambridge-MIT Project (CMI), on successful trials of an IP/ISDN codec from Tandberg. CMI subsequently purchased four of these codecs, for the Computing Service facility and for new videoconferencing facilities in CMI, the Centre for Applied Research in Educational Technology (CARET), and the Judge Institute of Management Studies, while the Service team will manage the installation, training, and support for all four.
48. The Technical User Support division (TUS) manages and largely staffs the advice and consultancy services, including the Help Desk and Techlink (see below) and also looks after a wide range of application software on the PWF and other Computing Service facilities. At the request of the Windows Integration Group, a bookable Test Bed Facility of PC server and workstation equipment was set up to double both as a facility for learning about NT under the guidance of TUS staff and an environment in which local computer support staff may safely rehearse planned changes to their institutional systems. Although there were no major virus problems until the summer worms, staff continued to distribute McAffee VirusScan and other anti-virus software and to remind users of the need to install it and keep it up to date.
49. At the start of the year, the joint Computing Service/Library Electronic Reference Library system held ten databases, although PsycLit was soon replaced by the larger and more frequently updated PsycInfo. Additions during the year were MLA International Bibliography (covering language, literature, linguistics, and folklore) and British Nursing Index, Cumulative Index to Nursing and Allied Health Literature, and Health Management Information Consortium (covering between them nursing, midwifery, biomedicine, and health care and management). Medline was again by far the most widely and heavily used of the individual databases.
50. TUS staff give general user consultations on a wide range of topics and provide a member of the support team which co-ordinates computing developments and technical liaison with the Computing Service for the whole of the Sidgwick Site, where staff from the Literary and Linguistic Computing Centre (LLCC) also provide a specialist service for the humanities. Because of changes in recording practice, it is no longer possible to give detailed figures for staff time spent on all kinds of advice and consultancy work but the following table summarizes Help Desk calls for 2000-01, with 1999-2000 figures in italics:
More calls were initiated by telephone (45%) than in person (28%) or by e-mail (also 28%). Over 85% of calls were closed within 1 hour and over 90% within twenty-four hours while 2% of calls had to be referred to local computer staff when Help Desk staff could not provide a ready answer to a question about a non-Computing Service facility.
51. In return for asking users to direct queries about non-Computing Service facilities to local computer support staff, the Service provides those staff with 'fast track' backup and general support through the Techlink scheme, which had over 300 members by the end of the year (up from 260 a year earlier). Techlinks initiated 14.4% (13.4% in 1999-2000) of all Help Desk calls but, because their questions tend to be more difficult, again accounted for 21.8% of the total time spent answering calls. For Michaelmas 2000, the Service's various IT support mailing lists were rationalized into just two (Departmental and College), which are now widely used not only by the Service but also by support staff seeking (and usually finding) all kinds of advice and information from their peers. The Techlink seminars given by Service staff were regularly updated to reflect current concerns and proved so popular that both the room and the day had to be changed to accommodate everyone who wanted to come to them.
52. The Computing Service training courses on central services and popular applications software help reduce the advice and consultancy load and enable better use to be made of computer equipment throughout the University. Some Michaelmas Term courses were again heavily oversubscribed and additional sessions had to be arranged, although demand for the most basic courses fell off. It was clear nevertheless that some new students need IT literacy training, although additional publicity channels may be needed to reach them all. Demand for courses was high again in the Easter Term as students took their last chance to benefit from free training; the additional sessions included a two-afternoon Introduction to Unix given at New Addenbrooke's. Exit surveys continued to report customer satisfaction rates of 90% or 95%. The table summarizes course information for 2000-01, with 1999-2000 in italics.
|Spreadsheets and Databases||35||34||1,091||1,064|
|Word and Doc. Processing||42||41||931||826|
53. After the resignation of a key member of the training team, who had inter alia given all the courses for assistant staff, it was fortunately possible to recruit a well-qualified replacement at the first attempt. Well attended and received imported courses from external providers included an SPLUS introduction and two courses for webmasters from Netskills, while in July Dr Hazel, author of the widely used Exim e-mail software, gave a very successful two-day workshop to 86 participants (over double the original plan), including twenty from overseas. Computing Service trainers also provided a growing number of customised courses for specific groups, ranging from an Excel course for administrative staff needing to input RAE informa-tion through Powerpoint for staff from English and Windows 2000 for Jesus College to Access, SPSS, and Excel courses for delegates attending a conference at Anaesthesia.
54. Total loans and sales from the self-teaching courseware library went up by 25% and a total of 212 assistant staff from University institutions attended 31 Personnel Division funded course sessions (compared to 176 and 23 in 1999-2000), although the attendance at similar courses for College staff was only 50 at six sessions (compared to 167 at eighteen).
55. Major projects carried out during the year included the development of PWF Linux (see above) and the completion of the JTAP funded work to develop the friendly probing software (as described in last year's report). After much development and test-ing of this largely unsung but complex and critical infrastructure component, the introduction of a new print spooler was achieved with only minor glitches. The week-long hands-on course on Linux System Administration was given twice and the ssh CD, which allows Unix users to log on securely from remote locations, was updated. Direct institutional support, including help to recover from security incidents, was provided to Chemistry, Lucy Cavendish, Management Information Services, Biochemistry, and twenty other Departments or Colleges.
56. Using information and feedback obtained from regular contacts with major suppliers, Computing Service staff again advised individual users and institutions, either in person or via web pages, about economic sources of personal computer systems to meet their particular needs. The Service web pages were linked to the new site maintained by the University Purchasing Office, with which a close liaison has developed. The fifth annual November Demonstration Day again attracted hundreds of University users into the Small Examination Hall, where a wide range of hardware and software vendors were exhibiting their wares.
57. This division liaises on computing matters at institution level with both senior management and computer support staff, and its two staff are able to call on assistance from the Directorate, Network Support, and Technical User Services as required. The help given to Departments facing TQA assessments in-cluded advising on the preparation of IT-related base materials, attending Learning Resources meetings, and arranging for assessors to visit the central PWF classrooms. Staff also spent much time on the difficult IT resource and support issues arising out of the transfer of Homerton's teaching to the School of Education and on understanding and implementing the access control system for the Raised Faculty Building.
58. In a new development, Institution Liaison staff were asked by some institutions not only to advise on computer staff gradings and individual appointments but also to help with the appraisal of existing staff; this raised some rather more general concerns about support and personal development opportunities for computer staff which were shared with the Personnel Division. Meanwhile staff continued to organize both well-attended termly meetings open to all Departmental and/or College computer staff and smaller liaison meetings at which staff from the Computing Service and from the Library or Management Information Services could discuss issues of mutual interest. They also fitted in almost 150 visits by request to University institutions and 100 to Colleges, assisted with staff selection on nearly thirty occasions and helped introduce several new staff into the idiosyncrasies of computing in Cambridge.
|M. S. LONGAIR Chairman||R. GLEN||S. KEARSEY|
|S. J. BARTON||K. GLOVER||I. LESLIE|
|T. A. CARPENTER||R. HANKA||B. K. OMOTANI|
|D. E. DETMER||T. R. HARRISON||M. D. SAYERS|
|I. M. LEM. DUQUESNAY||M. F. HEATH||P. B. WILSON|
|P. K. FOX|
The following tables give details of various activities of the University Computing Service for the year 1 August 2000 to 31 July 2001, with figures for the previous year in italics. Those tables which break down information on an institutional basis all follow a similar pattern:
|(a)||Overall summary (University by Faculty group, Colleges, and External groups)|
|(b)||Cambridge University Faculty and similar groups (by Faculty or Department, except that the Faculty of Clinical Medicine is not divided.)|
|(c)||Colleges (by College)|
In this shortened version of the Statistical Annex all the (b) and (c) tables are omitted.
Tables 1(a) to 1(c) show Cambridge University Data Network connections by institution. With no asynchronous or X.25 lines left, the remaining connections, which are overwhelmingly ethernet in type, have been divided into slow and fast as follows:
|Slow ethernet||All 10 Mbps except for 8 at less than 2 Mbps|
|Fast ethernet||All 100 Mbps ethernet except for one 155 Mbps ATM and 4 1000 Mbps ethernet|
Figures are for the end of 2000-01; end of 1999-2000 figures are in italics.
Table 1(A) - Data Network Connections
|Slow ethernet||Fast ethernet|
|Arts and Humanities||2||6||4||3|
|Humanities and Social Sciences||10||9||4||4|
|Other GB Institutions||6||7||6||3|
Table 2(A) - Transatlantic Traffic By Institution
|Arts and Humanities||149||(0.9%)||54||(0.5%)|
|Humanities and Social Sciences||232||(1.4%)||151||(1.4%)|
|Other GB Institutions||1,439||(8.5%)||969||(9.2%)|
|Research Council institutions||636||(3.8%)||444||(4.2%)|
Tables 3(a) to 3(c) show numbers of Hermes users and messages by institution.
|Users||Includes users who access Hermes directly and those who have their mail forwarded to another system such as the CUS.|
|Messages||Includes both messages sent and messages received, which are typically in the ratio 1:3.|
Figures are for 2000-01; 1999-2000 figures are in italics.
Table 3(A) - Hermes
|Arts and Humanities||1,803||(6.5%)||1,674||(6.4%)||3,125||(5.5%)||2,335||(4.8%)|
|Humanities and Social Sciences||3,959||(14.2%)||3,673||(14.1%)||4,884||(8.6%)||3,848||(7.8%)|
|Other GB Institutions||906||(3.2%)||711||(2.7%)||1,713||(3.0%)||1,083||(2.2%)|
Tables 4(a) to 4(c) show numbers of users and sessions by institution. Note that the figures for Computing Service users include temporary identifiers issued to users on training courses.
|Users||Number of individuals using the system at least once during the year|
|Sessions||Number of times logged-on during the year|
Table 5 gives details, extracted from Tables 4(a) and 4(b), on the use of the PWF from certain University locations while Table 6 gives similar details of the use of the PWF from those Colleges which have either Managed Clusters or tunnelled links as members of the PWFClub.
In all the Tables, the figures are for 2000-01, with 1999-2000 figures in italics. Unfortunately, for reasons given in the main Report (para. 34), the 2000-01 figures only include data for the period from 1 August 2000 to 17 April 2001.
Table 4(A) - Public Workstation Facility
|Arts and Humanities||528||(5.6%)||446||(5.0%)||36,281||(5.7%)||53,456||(6.5%)|
|Humanities and Social Sciences||758||(8.1%)||657||(7.4%)||58,819||(9.3%)||82,642||(10.1%)|
|Other GB Institutions||811||(8.6%)||988||(11.2%)||14,567||(2.3%)||81,189||(9.9%)|
Table 5 - University Public Workstation Facility Sessions By Location
|Computing Service locations|
|New Museums Site (Phoenix User Area and Balfour Macintosh Room)||1,869||3,293||25,024||95,566|
|New Museums Site (Phoenix Teaching Room)||1,380||2,016||13,148||75,657|
|Old Music School||2,021||2,322||15,981||49,118|
|Sidgwick Site (Oriental Studies Basement)||585||786||13,462||54,038|
|Managed Cluster locations|
|Clinical Veterinary Medicine||338||324||37,518||55,420|
|Archaeology and Anthropology||89||63||4,368||2,355|
|Social and Political Sciences/African Studies||94||41||7,335||853|
|Arts and Humanities (English, Philosophy, MML Graduate Centre, and Divinity)||642||13,246|
|PWF Club locations|
Table 6 - Public Workstation Facility Use From Colleges
|Managed Cluster locations|
|Gonville and Caius||610||97||54,730||4,203|
|PWF Club locations|
The Central Unix Service is a general user facility based on a multi-processor configuration. The Thor teaching facility is used for certain undergraduate and taught postgraduate courses.
Usage figures are given in percentage terms only and refer to the use of the Solaris 2 operating system. Note that the figures for the Computing Service include all systems, applications, and operational support activities, together with Unix-based training courses. Under Solaris 2 a certain minimal level of usage may be automatically recorded by the system against a user who has not actually logged on during the year; an active user has therefore been defined as one whose recorded use exceeds an appropriate threshold.
Tables 7(a) to 7(c) are for the Central Unix Service and show the relative use giving:
|CPU||CPU time consumed as a percentage of total use|
|Users||Number of individuals making some active use of the system during the year|
Figures are for 2000-01; 1999-2000 figures are in italics.
Tables 8(a) to 8(c) give similar figures for the use of Thor.
Table 7(A) - Central Unix Service
|Arts and Humanities||3.7%||5.1%||265||(7.5%)||326||(7.8%)|
|Humanities and Social Sciences||7.4%||11.5%||446||(12.6%)||552||(13.2%)|
|Other GB Institutions||18.9%||19.5%||308||(8.7%)||312||(7.4%)|
Table 8(A) - Thor
|Arts and Humanities||0.0%||0.4%||1||(0.2%)||2||(0.4%)|
|Other GB Institutions||5.2%||5.1%||17||(3.6%)||22||(4.8%)|
Tables 9(a) to 9(c) show the use of the Pelican data archive service by institutions at the end of 2000-01 with end 1999-2000 figures in italics.
|Stored||Amount of data stored (in MB)|
|Users||Number of individuals with stored data|
Table 9(A) - Pelican Data Archive Service Use
|Arts and Humanities||19,811||(9.9%)||6,104||(3.7%)||60||(6.5%)||50||(5.3%)|
|Humanities and Social Sciences||10,039||(5.0%)||9,539||(5.7%)||113||(12.3%)||112||(11.8%)|
|Other GB Institutions||101,279||(50.6%)||71,037||(42.8%)||110||(12.0%)||112||(11.8%)|
Table 10 shows totals of software sales items supplied by month. Tables 11(a) to 11(c) show software sales items supplied by institution. General software (excluding media) and SunSpectrum/Silicon Graphics items are shown separately. Table 12 lists software and other items distributed by the Computing Service. An item may be any of a full copy of a software package, an upgrade, or a maintenance contract.
Figures are for 2000-01; 1999-2000 figures are in italics.
Table 10 - Software Items Supplied By Month
|TOTAL FOR YEAR||13,730||(100.0%)||15,100||(100.0%)||486||(100.0%)||556||(100.0%)|
Table 11(A) - Software Items Supplied By Institution
|Arts and Humanities||1,080||(7.9%)||467||(3.1%)||11||(2.3%)||11||(2.0%)|
|Humanities and Social Sciences||427||(3.1%)||1,225||(8.1%)||1||(0.2%)|
|Other GB Institutions||2,410||(17.6%)||1,937||(12.8%)||62||(12.8%)||135||(24.3%)|
|Research Council institutions||734||(5.3%)||95||(0.6%)|
Table 12 - Software Sales By Product
|Apple ARA Server||11|
|Apple at Ease||3|
|Apple Network Assistant||3|
|Interarchy (was Anarchie + Finger)||1||105|
|Microsoft Services for Netware||2||6|
|Microsoft Services for Unix||4|
|Novell Netware (per server)||53||48|
|Bartholomew's Map Data||2||8|
|Corel Print Office||24||222|
|GRAPHICS AND DRAWING||1,493||1,507|
|Apple Language Kits||26|
|Apple System Software||250|
|Microsoft Exchange Server||72||25|
|Microsoft ISA Server||2|
|Microsoft Plus for Windows||98||2|
|Microsoft SQL Server||72||154|
|MS Systems Management Server||7||21|
|MS Windows (3.x/Wkgp/95/98/Me)||128||1,011|
|Microsoft Windows NT||1,214||2,076|
|Microsoft Visual Basic||40||45|
|Microsoft Visual C++||69||118|
|Microsoft Visual FoxPro Pro||8||26|
|Microsoft Visual InterDev||2|
|Microsoft Visual J++||2||3|
|Microsoft Visual SourceSafe||2|
|Microsoft Visual Studio||33||41|
AND PROJECT MANAGEMENT
Table 12 - (CONT.)
|SPSS Science Products||39||18|
|Adobe Sabon Font||89||390|
|WORD AND DOCUMENT PROCESSING||1,206||1,411|
|Microsoft Select Advantage||89||658|
|TOTAL for Software||13,730||15,100|
|Sun Hardware maintenance|
(Total Number of machines)
|Sun Software under maintenance|
(Licence, media, and documentation)
|Silicon Graphics maintenance||80||97|
Table 13 shows the number of contract and non-contract repairs undertaken for different types of equipment. Tables 14(a) to 14(c) show both the number of repairs and the number of staff hours spent on contract and non-contract repairs distributed amongst institutions.
Figures are for 2000-01; 1999-2000 figures are in italics.
Table 13 - Equipment Repaired By Type
|Other PC clones||128||92||122||114|
Table 14(A) - Equipment Repaired By Institution
|No Hours||No Hours||No Hours||No Hours|
|Arts and Humanities||5||7||13||15||45||62||95||127|
|Humanities and Social Sciences||4||18||15||37||98||133||191||204|
|Other GB Institutions||40||39||112||126||27||18||84||28|
|No institution specified||17||9||32||30|
Table 15 shows the monthly totals of calls dealt with by the Help Desk. Tables 16(a) to 16(c) show the number of Help Desk calls and users by institution. In Table 15 only, the figures exclude those calls made direct to the NT Support service (172 in 2000-01; 83 in 1999-2000).
Figures are for 2000-01; 1999-2000 figures are in italics.
Table 15- Help Desk Calls By Month
|TOTAL FOR YEAR||9,886||(100.0%)||9,943||(100.0%)|
Table 16(A) - Help Desk Calls By Institution
|Arts and Humanities||1,076||(10.7%)||1,243||(12.4%)||432||(9.6%)||475||(10.2%)|
|Humanities and Social Sciences||1,977||(19.7%)||1,907||(19.0%)||749||(16.6%)||801||(17.1%)|
|Other GB Institutions||477||(4.7%)||433||(4.3%)||199||(4.4%)||189||(4.0%)|
|Other external (world-wide)||489||(4.9%)||263||(2.6%)||24||(0.5%)||32||(0.7%)|
Table 17 lists the training courses provided by the Computing Service; in each case the figures show the number of times a course was offered and the total number of persons attending. Courses involving more than one session count attendances at each separately. Tables 18(a) to 18(c) show how attendances at courses were distributed amongst institutions.
Figures are for 2000-01; 1999-2000 figures are in italics.
Table 17 - Training Courses
|IBM PC: basic intro to Windows 95/98/NT4||4||4||61||86|
|IBM PC: further on Windows 95/98/NT4||3||3||55||83|
|IBM PC: Windows 95/NT4/98 top 10 tools||2||25|
|Macintosh: top 10 tips||1||10|
|Unix: further use||2||2||41||40|
|Unix: introduction to Perl||1||1||37||31|
|Unix: configuring X||1||9|
|Linux system administration||2||3||530||460|
|TABLE 17 - (cont.)|
|World Wide Web: introduction||4||7||70||144|
|World Wide Web: further exploration||5||4||53||76|
|Setting up a WWW server (Windows NT4)||1||3||10||37|
|Managing a WWW server on a Unix system||3||3||94||72|
|Designing and managing a Web site||3||3||74||82|
|HTML: advanced (overview)||3||3||141||112|
|HTML: advanced (practical workshop)||3||51|
|Dreamweaver: for advanced HTML||2||50|
|Netskills images and multimedia||1||24|
|Netskills content management||1||26|
|Graphics manipulation for the World Wide Web||3||3||48||43|
|Databases and the World Wide Web||1||11|
|Email using Pine: introduction||1||3||14||36|
|Email using Mulberry: introduction||3||1||38||22|
|Pine tricks of the trade||1||1||5||13|
|Microsoft Excel: introduction||8||9||139||171|
|Microsoft Excel: introduction, self-paced||5||4||225||180|
|Microsoft Excel: further use||5||6||95||136|
|Microsoft Excel: advanced||3||1||48||16|
|Access: further use||4||3||138||84|
|Filemaker Pro: introductory practical||2||1||35||14|
|SPREADSHEETS & DATABASES||35||34||1,091||1,064|
|Microsoft Word for Windows: basic||4||4||48||57|
|Microsoft Word for Windows: intermediate||3||3||53||78|
|Microsoft Word for Windows: advanced||3||3||55||68|
|MS Word for Windows: introduction, self-paced||5||4||171||135|
|Microsoft Word on Macintosh: introduction||1||2||8||24|
|Microsoft Word on Macintosh: advanced||1||1||5||10|
|Powerpoint: quick start, self-paced||3||3||63||86|
|Powerpoint: advanced workshop||2||3||30||45|
|Quark Xpress: introduction||3||2||62||35|
|Photoshop for Photographers: demonstration||1||1||16||12|
|WORD & DOCUMENT PROCESSING||42||41||931||826|
|Simple questionnaire design and data entry||1||1||14||8|
|SPSS: further use||1||1||16||7|
|S-PLUS: jump start workshop||1||54|
|Simple C Programming||2||2||237||210|
|Visual Basic for applications||2||104|
Table 18(A) - Course Attendances
|Arts and Humanities||285||(5.5%)||271||(5.5%)|
|Humanities and Social Sciences||649||(12.6%)||682||(13.8%)|
|Other GB Institutions||326||(6.3%)||208||(4.2%)|
|No institution specified||3||(0.1%)||3||(0.1%)|
|Previous page||Table of Contents||Next page|
Copyright © 2002 The Chancellor, Masters and Scholars of the
University of Cambridge.
Comments should be sent to email@example.com