Informatics Computing Support
This page outlines the purpose
of support that
is provided to users of Informatics computing systems, how
we provide it and what
If you are looking to get support, try the Support FAQ. If you are here because
you really aren't sure how to get the support you need, trying talk to
one of the local computing support staff
or the local site
There's a separate page on operational and
policy issues, including how to handle escalation of
Computing support in Informatics is intended to help members of
Informatics make effective use of the computing facilities provided.
This includes students, teaching, research and administrative staff.
How it is provided
As a very simplistic separation, support to users, from a user
perspective, falls into two categories:
- frontline, day-to-day support for user issues
- behind the scenes implementation and development
this is provided by the Frontline Support Team, comprised mainly of
Computing Support Officers. Their remit is to do a first-pass on
handling individual queries, sort out machine problems and maintain
the labs of machines at operational levels. There are also daily and
longer-term scheduled activities such as changing backup tapes, doing
upgrades on machines, changing toner cartridges, etc.
- behind the scenes
Support for the rest of Informatics computing is provided largely by
Computing Officers. The work covers a very wide variety of typically
technical tasks, ranging from helping the Frontline Support team with
more difficult issues that users have, to the complex process of
developing infrastructure and implementing services to support
Informatics as a whole.
There are several main areas where support is provided.
- general or
This typically covers routine, day-to-day issues such as logging in,
network connectivity, access to commodity services such as email,
filespace, printing, etc.
The Frontline Support Team expects a significant proportion of such
requests to require prompt attention. However, assessing urgency and
the relative impact of user problems is not a black and white issue.
It will always be necessary to adjust the speed of response in the
context of other current demands.
In general, queries get assessed for their priority within a
few minutes of arrival. Depending on the current workload, there
should be a response to your query within a working day;
usually it will be more quickly than that and provide the solution.
If there isn't an immediately obvious solution, the support team will
aim to give an indication of when a user might expect a solution.
Balancing the speed of response against the amount of time it takes is
not easy - see below.
support in this area relates to the installation of the operating
system and additional software on a computer. (For information on how
equipment is allocated please see the computing
How such management is implemented heavily
reflects two key and interrelated issues, namely
security and scalability.
- in order to have centrally
manageable systems that computing staff can trust are secure, it is
important to know the integrity of anything installed with system
A key principle of DICE is the automation of machine installation and
subsequent management (eg, updating and status monitoring), using LCFG . Unless installation and
management are automated (and, by implication, scalable), it is not
possible to provide large numbers of machines whose behaviour,
performance and status is predictable.
With these two factors in mind, the support of machines currently
falls into two categories.
- standard installation
and configuration with respect to a particular category, (eg staff,
research student, etc). The machine will be fully managed centrally.
With respect to security, this model currently requires that users do
root privileges in order to have confidence in
the integrity of the system.
- the user has total
responsibility for installation and management of the machine.
Having provided the DICE commodity model as planned, it is a high
priority to provide further options to meet the varied needs of users.
These options support different aspects of security/trust and the
degree of scalable management. It is hoped to provide at least two
additional options which would lie between
- the user can opt for installation of additional, user-contributed
packages (not necessarily contributed by the same user). These
packages are not supported centrally. However, they are
vetted and restricted in operation to avoid security vulnerabilities
(eg, post-install scripts will be prohibited).
- the machine is managed according to a central profile but the
user has root access to allow installation of privileged packages. In
the event of problems the machine can be reinstalled to the state of
its initial profile, rather than its current state.
- support for installed software
Much software available on centrally managed systems in Informatics is
available as part of a standard distribution (eg, RedHat 9). Since
there are typically hundreds of installed packages, it is not
practical for them to be individually supported locally. Typically
support for them is available through manual pages or associated
material on the web.
However, some applications or packages, such editors, mail programs,
browsers, are key to provision of/access to local services and their
use in the Informatics context is given greater support. Typically
there will be configuration and use information specifically
documented on local computing web pages. Ultimately though, the vast
majority of these packages are provided `as is' and while the aim is
to provide as much support as practical it may not always be possible.
- connection/access to Informatics services
There are a number of core services provided to users which they make
use of from both managed and non-managed machines, within Informatics
or elsewhere. Such services include email, personal web space, local
network access (wireless or wired) (eg DHCP, DNS), remote access via
secure gateways, printing, etc.
For these services, information is provided about how to access them
but diagnosis of problems with access from non-managed machines are
supported only up to the point of connection.
Balancing the speed and frequency of responses to support requests
against the amount of work that generates is not easy.
On the one hand a fast response can reassure the user that the problem
is being attended to. However, if that requires handling a request
several times (eg, two or three replies rather than one) that
generates extra work. It can often be the case that information is
required from other computing staff who are focussing on longer term
activities and understandably don't respond to queries instantly. In
our view it is, on balance, better to respond when information becomes
available. However, as stated, it is intended to give some indication
of progress within the working day and usually it will be quicker than