The 2007/08 JISC Monitoring Unit Annual Survey of Network Infrastructure Services was the sixth in an ongoing series designed to investigate customer satisfaction with, and inform future development of, the JANET network and network services provided by JANET(UK). The survey ran from 3^rd December 2007 to 25^th January 2008, and all institutions with JISC-funded connections to JANET were invited to participate.

The survey questionnaire comprised 66 questions in 12 sections and followed the same format used since 2003/04, albeit with a number of minor changes. Of the 667 institutions invited, complete responses were received from 281; representing a 42% response rate, which is a modest decrease from 52% in 2006/07. Nevertheless, the size of the sample is sufficient to ensure the validity of the statistical analysis.

The results from this survey were appropriately coded and analysed alongside data from previous years to examine year-on-year changes and differences between sectors (FE vs. HE only due to small sample size of RC and SC). Linear regression was used to investigate the presence of any long-term trends in the data.

The JANET network and network services remained highly regarded this year, with the network universally considered 'leading-edge'. The quality of liaison with JANET(UK) staff decreased slightly this year but was nevertheless well regarded and there is no evidence of any long-term decline. Regarding ways in which the quality of the network or JANET(UK) services could be improved, the main theme was to provide greater bandwidth to Further Education and Specialist Colleges in England, followed by a desire for more pro-active engagement with the FE community.

Increased use of the network (24%) and remote access (22%) remain the most common planned developments affecting institutions' use of JANET, with the implementation of shared services (19%) being an important new plan this year. Collaboration, rich media and Web 2.0 are also moving up the agenda. Sector differences are unremarkable other than FE now contemplating collaborative projects as much as HE.

The enhanced capacity of SuperJANET5 has been delivered to users desktops without problem by the majority (58%) of institutions, and where problems are reported, these mostly relate to 'final mile' bandwidth restrictions that SJ5 did not necessarily address.

The perceived quality of JANET connections increased further this year, with gradual long-term increases (from a high start) now well established. Reported saturation of JANET connections remains persistently low (29%), which along with consistent estimated times to saturation of circa 18 months, indicates JANET upgrades are keeping pace with demand. Nevertheless, at the time of questioning, there remained an endemic saturation problem within English FE (38%), due to external funding decisions.

Institutions reporting purchasing extra bandwidth on their JISC-funded JANET connections remain rare (17%) but are increasing, particularly in HE. The acquisition of additional network connectivity is more common (41%), particularly in HE (62%), and is becoming increasingly so. This suggests, along with other comments in the survey, that resilience is of increasing importance to institutions.

Commercial providers supply most additional connectivity (79%), typically via DSL links (52%), and this has not changed in recent years. The reasons for choosing commercial providers are cheap resilience in FE, and non-academic/AUP unfriendly traffic in HE. Resilience in HE is, therefore, more likely provided by additional JANET or regional network connections, which are much more common than in FE (34% vs. 8%).

Satisfaction with the availability of JANET connections increased slightly this year, with a small long-term increase (from a high start) now established. The recent increase in the availability Service Level to 99.7% was well received; 93% agreed it was sufficient for their institution's needs, reinforcing the perceived quality of the network.

The timing of the scheduled maintenance period remains convenient for the majority (89%) of institutions, with 96% indicating they receive sufficient notice of such maintenance. Only 3% indicated they had received an unreasonable amount of scheduled maintenance.

Notably, HE institutions were generally more demanding of network availability than FE, demonstrating slightly lower satisfaction with both the Service Level and the timing of scheduled maintenance.

JANET(UK)'s fault handling and resolution process remained highly regarded this year, with evidence of longer-term increases in satisfaction. Again, however, HE institutions tended to be somewhat more critical than FE. The current system for providing network outage information was considered acceptable but more targeted and concise email alerts were requested, and SMS messages suggested as an addition/alternative, as they were last year.

Reported Netsight use increased to 78% of institutions this year, and availability of the system is still highly rated. Demand for more detailed traffic data remains, however. Most institutions (79%) employ some form of bandwidth management, with caching proxies remaining the most common method (61%).

The perceived quality of JANET CSIRT's services decreased slightly this year, although the long-term trend is for it to remain high (NB: JANET(UK) have previously reported staffing problems in this division during 2007). Most institutions (54%) do not have contact with JANET CSIRT in an 'average month', although HE have more frequent contact than FE. As has been the case in previous years, respondents consider it 'important' to 'very important' that JANET CSIRT, rather than an external body, provides security services to the community.

The proportion of institutions agreeing that JANET(UK) should pre-emptively filter network traffic has decreased to 61% this year, although it still represents a long-term increase. Filtering between external networks and the backbone remains more acceptable than filtering between institutions and the backbone, and FE institutions (65%) remain more in favour than HE (50%).

Email scanning is employed by almost all institutions (97%), with on-site solutions the most popular (78%), regardless of sector; DNS blocklists are used by 66% of institutions, although more so by HE (81%). Provision of email security services by JANET(UK) remains 'important' to the community, although usage of the JANET DNS Blocklists (29%) and JANET Mailer Shield (3%) remains low in comparison to similar non-JANET services.

The JANET Roaming Service was regarded as 'important' by HE institutions but FE were 'neutral' towards it, as was the case last year. Use of the service was in line with JANET(UK) figures, and has increased from last year. Most respondents were satisfied with the service, although a few indicated dissatisfaction due to its limited coverage and technical problems.

Reported use of the JANET Videoconferencing Service has increased from 22% of institutions to 34% this year; although it remains HE dominated (67% vs. 20%). Satisfaction with the service remains high.

JISCmail usage was reported by 56% of respondents, with a HE bias (75%). Satisfaction has apparently increased slightly this year and FE remain more satisfied with the service than HE.

Reported use of the Primary (27%) and Secondary (49%) Nameserver services was higher and lower than JANET(UK) figures, respectively, whilst satisfaction with them was high.

The JANET Managed Router Service was confused with a similar service previously operated by the JISC RSCs. Usage was reported by 43% of institutions, and satisfaction was high.

Reported use of the JANET Network Time service remained static (59%) this year, although there is a long-term trend for it to increase. Use remains HE biased (87% vs. 49%), and there were still a number of FE respondents indicating they were unaware of the service.

The JANET Web Services remain 'satisfactory' to 'good'; use has increased in line with JANET(UK) figures and is biased towards Specialist Colleges, as intended.

The JANET Usenet News services remain little used (20% of institutions, mostly HE) and are generally considered 'satisfactory' to 'good' despite a long-term trend for satisfaction to decrease.

Reported use of Voice over IP (VoIP) has increased slightly this year, with 40% of the community using it in some form with a further 18% planning to implement it. Usage remains HE dominated and is mostly over the 'wider internet' or 'breakout to the public telephone system', i.e. to destinations outside JANET. Comments indicate a demand for advice, such as case studies and best practice, from JANET(UK) and the provision of a central SIP/directory infrastructure.

Eighteen suggestions were received about services JANET(UK) should provide. Extending web mail/filtering services to the entire community and providing a central backup/disaster recovery service were the most common suggestions. Notably, five requests were made for services JANET(UK) already offers.

Satisfaction with the JANET Service Desk remained high this year, although FE institutions were again significantly more satisfied than HE.

JANET(UK)'s web site was considered a valuable resource by almost the entire community (96%), and its content and design were rated 'good', as in previous years. Notably, 'ease of finding information' was rated significantly lower than other aspects, and FE institutions were more satisfied with all aspects of the site than HE.

All the Technology Advisory Services were included in the survey for the first time this year. About 10% of institutions used each of the services, with satisfaction in the range 'neutral' to 'fairly satisfied'.

Documentation published by JANET(UK) continued to be highly rated, with long-term trends for increasing quality. Notably, FE institutions no longer rate documentation more highly than HE.

Most respondents (77%) were satisfied with JANET(UK)'s programme of workshops and conferences, with HE institutions being more satisfied than FE. Comments indicated that this may be due, in part, to problems in staff from small institutions attending often distant events.

Satisfaction with JANET(UK)'s training courses decreased slightly this year, although the high long-term trend remains static. Notably, differences in satisfaction between FE and HE were absent this year. Only a few comments mentioned staffing or location problems, suggesting these are less of a concern when attending training courses, as opposed to other events. The provision of training courses by JANET(UK) was considered more important this year, and there is a long-term trend for this importance to increase.

For the fifth consecutive year, survey respondents have indicated the JANET network is of very high quality (Figure 1.1), with long-term trends for increasing quality apparent for most aspects of the network and its operation. Development of the network is keeping pace with demand, and SuperJANET5 upgrades have been delivered successfully to end-users. Nevertheless, demand looks set to rise as more institutions implement remote access, establish collaborations, and the use of rich media/Web 2.0 grows. Furthermore, at the time of questioning, a core of English FE institutions (38%) remained constrained in their use of JANET due to bandwidth restrictions.

Figure 1.1: JANET Network perceived quality/satisfaction [1]

The network services provided by JANET(UK) are also well regarded by respondents but not to the same extent as the network itself (Figure 1.2), and the limited longer-term data available suggests satisfaction is generally static. There is no evidence of any endemic problems with the services, other than a recurring issue over their promotion to the community. This seems to be a greater problem in the FE sector, where the use and even knowledge of services such as the JANET Network Time service and DNS Blocklists remains particularly low, suggesting JANET(UK) needs to continue its promotional efforts in this area.

Figure 1.2: JANET(UK) network services perceived quality/satisfaction

JANET(UK)'s customer services similarly remain well regarded this year (Figure 1.3), and there is little evidence of any longer-term change in satisfaction with them. The Technology Advisory Services, which were included in the survey for the first time this year, are perhaps less well regarded, although each was judged at least 'satisfactory' and their low reported use (small sample size) makes further inferences inadvisable.

Figure 1.3: JANET(UK) customer services perceived quality/satisfaction

In conclusion, JANET(UK) are continuing to provide a high quality network and associated services that meet the demands, including those going forwards, of the community they serve. Notably, the JANET network is considered to be of greater quality than the other JANET(UK) services, although the reasons for this are not apparent from this survey. Similarly, FE institutions tend to be more satisfied with JANET(UK)'s services than HE ones, despite making the greater proportion of negative comments; again, the reasons for this are not clear.

The annual Survey of the use of JANET and JISC-supported Network Infrastructure Services now has five years of detailed data showing, at worst, no decrease in the high standards of JANET(UK)'s services. The Monitoring Unit suggests this provides a more than adequate evidence base for the success of JANET(UK). Future survey work may better concentrate on identifying problems, assessing developing areas, and investigating more intractable issues (such as those mentioned above), rather than the current exhaustive assessment of satisfaction. The maintenance of long-term data is, of course, important but could be achieved with a 'lighter-touch' approach.