I feel that all UCT academic staff should be made aware of the proposed Core ICT Services Service Level Agreement. Under this agreement, unplanned downtime for ICT services is restricted to 2%. No doubt those who plan to rubber stamp this document believe that this number is insignificant.
The reality is that academic staff (and postgraduate students) rely on ICT services to do their jobs. With more and more resources becoming available online exclusively, it is nearly impossible to do research without internet access. Two percent downtime means that researchers will be unable to do their jobs for 48 minutes per forty hour work week. Considering that staff are still being paid during this time, it is only fair to consider this time as paid leave. All in, this ICT downtime accounts for about 6 days of paid leave per year.
I would encourage UCT management to calculate the amount of money this downtime costs the University. Once this value has been calculated, perhaps it will provide justification for the University to insist that ICTS provide a truly world class level of ICT services. Slow internet is forgivable given the realities of the South African telecoms industry, poor service is not.
Thank you for engaging with the Core Services SLA and for offering your comment. At this stage the proposal is just that - a proposal - and ICTS welcomes all input.
ICTS does not believe that the 2% unplanned downtime is insignificant. We realise the implications of any downtime to the business of the university.
However, the writer's assumptions that the percentage relates to a working week is incorrect. The services are measured 24x7. In other words, although UCT officially works a 37.5 hour working week, we are measuring the services for the full 160 hours a week. For every work hour, there are 4 non-work hours that we are still measuring against.
The proposed 98% target is a "best guess" target, based on our experiences with the Faculty of Engineering & Built Environment SLA. Our new infrastructure is still being bedded down and we don't know how well it will perform. However, our expectations are that it will in most cases exceed the 98% uptime.
Availability is measured per service. If we take the service "Browsing the Internet" as an example, one would need the following to be in place and working correctly:
If any one of these supporting components experiences a problem, then the whole service is deemed to be unavailable. The individual components must thus perform at a much higher availability percentage in order for us to be able to measure the full service at 98% availability.
Our experience is that much of the downtime will occur outside of work hours. ICTS operates during standard UCT core hours, i.e. the 37.5-hour working week. If a service fails outside of these hours, then the likelihood is that it will cause extended downtime as we won't be here to fix it.
ICTS drafted a proposal to extend our support hours and took this to the University ICT Committee. The committee decided on a "wait-and-see" approach before allocating considerable recurrent budget to extend ICTS support hours.
We will monitor performance against the SLA, and where higher service levels are required, we will analyse all data and decide what the best approach would be. In some cases, extending the support hours may not be the remedy - it may be that new architecture or hardware will resolve the problem instead. The old adage holds true: "If you spend more, you get more".
The Core ICT Services SLA is a living document. It will be informed by measuring and analysing data and reporting against the targets. UCT may also decide that it should favour certain services over others, resulting in lower availability targets on some services and higher availability targets on others.
For more information about the Core ICT Services SLA, click here.
Information & Communication Technology Services
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.
Please view the republishing articles page for more information.