The focus of service level management (SLM) is starting to
change from using metrics that are convenient to either internal or
external IT providers to those that are important to the business -
although such a shift is still in its early days.
Although most enterprises already have some form of IT service
level agreements (SLAs) in place, most today tend to concentrate on
measuring the performance of individual technical components. This
means establishing targets based on factors such as availability
and how quickly the IT department or an outsourcing provider has
pledged to bring a given system or network up should it go
down.
But there is a growing impetus behind changing the balance by
attaching service levels to entire processes and activities that
are more meaningful to the business. Such activities include
performance response times, particularly during peak workloads,
number of transactions processed per hour and data security.
Tom Shave, a principal consultant at Morse Consultants,
explains, "It is about agreeing on appropriate and reasonable
targets for providing services to meet the organisation's
requirements, documenting those targets and using software to
measure and manage conformance in order to improve the service on
an ongoing basis."
Activity of this type is key to adopting a wider service
management approach to IT, which has been much talked about in the
industry over at least the last three years. But to date, the
number of organisations embracing such a philosophy has been
limited, although it is gaining traction.
For those enterprises that have gone down this route, however,
key drivers include moving to shared services, which often requires
that IT departments or outsourcing providers become more visibly
accountable as part of the transformation process, and regulatory
compliance issues. As Mathieu Poujol, director of technologies at
Pierre Audoin Consultants, says, "Compliance people do not like
there to be any risk around the IT infrastructure, but to be able
to channel risk, they need reliable metrics and clear agreement on
IT service levels."
Moreover, Tony Lock, programme director at analyst Freeform
Dynamics, indicates that there is "a distinct correlation" between
effective service management and improved ratings from customers.
"Once IT starts considering business operations in terms of how
they deliver services to users, they become much more in tune with
what users want and there are far fewer problems in terms of the
business perception of IT," he explains.
But, despite its importance, the SLM element of service
management has so far been lower down the priority list than other
areas such as incident and
change management. This situation has not been helped by the
fact that many IT departments find it difficult to disentangle and
identify the components of any given service.
To do this requires undertaking an audit of the organisation's
hardware and software assets and understanding underlying processes
in order to feed such information into a
configuration management database (CMDB). CMDBs provide a
single view of the authorised configurations of each component of
the IT infrastructure as well as the relationships between them,
which makes it possible to understand where bottlenecks are taking
place. Such information is, in turn, fed into the SLM system in
order to monitor what is happening with the individual components
that make up a broader IT service and alert staff to take action
quickly should a problem occur.
"The tool is aware of the service that the customer has access
too and how critical it is, but it also understands the underlying
components that make up the service. So the tool should
automatically prioritise assignments to the relevant support teams
and ensure that they are dealt with as soon as possible," Shave
explains.
Although
systems management tools have traditionally focused simply on
managing infrastructure, most have incorporated at least some form
of SLM functionality over the last few years. But such offerings
are still likely to require some form of adjustment to measure each
organisation's particular requirements and/or some level of
integration with existing offerings such as the helpdesk. It may
also become necessary to make additional purchases in order to fill
current gaps in functionality.
"The tools may not give you everything that you need depending
on what you are using and what you are measuring, which means that
you may need to acquire something else. You also need to report on
service levels in a way that the user will understand and while
some systems have good reporting tools, others do not," says
Lock.
Other important considerations include being able to provision
and de-provision services as required, which again may require the
introduction of datacentre automation tools to both speed up the
process and reduce costly manual intervention.
Given the complexity of such initiatives, however, the SLM
market is, unsurprisingly, an immature one today. This means that,
if projects are taking place at all, they tend to focus on
individual critical services rather than being more universal in
nature. As Poujol indicates, "This is not a simple project and if
you want to introduce an enterprise-wide SLM strategy, it will
realistically take several years."
As a result, Linda Symons, manager of IBM's Tivoli user group
believes that while SLM is a logical progression from present day
systems management, it will take some while yet to take off. "The
majority of companies have been struggling to manage the
constituent parts and having almost got that working are now
looking for the information they collect to show the whole picture.
I still do not think people are quite ready, but they are
interested and dabbling," she says.
Moreover, SLM has slipped on the radar since the credit crunch
and tends to be seen as the icing on the cake, says Symons. One of
the issues in this context, adds Symons, is that an SLM tool is
only as good as the data that goes into it, but most companies
struggling to get this right.
Another challenge is that such a move requires a reworking of
both business and IT processes, which can be a daunting and
political task. Although best practice frameworks such as
ITIL (Information Technology Infrastructure Library) for
service management and CoBIT (Control Objectives for Information
and related Technology) for IT governance can help, it is important
at the very least to establish a service catalogue to act as a
foundation.
This catalogue should include information based on discussion
with users such as how critical each given service is to the
business and why, how good service delivery is today as well as a
baseline for the future.
A problem for many organisations, however, is defining what a
service means in the first place, says Shave, as it means different
things to different organisations.
"Is it a technical service such as a network or an end-user
service, such as the provision of e-mail? In practice, it will be a
combination of the two as the technical service supports the
business one," he explains.
But what that business service comprises in terms of hardware,
software and networks will also need to be documented and processes
developed to ensure that all new services are created in a
consistent manner. It is also important to explore how such
services will be provided and to devise performance metrics that
are meaningful to customers themselves. The capabilities of the
support team likewise has to be realistically evaluated and thought
given to how service performance will be reported on.
A further important factor, however, is understanding where to
start and what services it is important to start off with. "You are
talking about managing the IT infrastructure as a whole and few
people have done that," says Lock.
"Those kind of questions can become deeply political and will
require IT service deliverers to talk to their customers in,
perhaps, a completely new way. It is more philosophical than
anything else and it is about two-way communication," he adds.
Such discussions can be fraught with difficulties, not least
because, if users are asked whether a service is good enough, they
are likely to say that it could be better. "But if you push them
and ask them what is important for the business as a whole, they
can start taking decisions on what gets priority. Not all services
are created equal, but it is up to the business to decide on that
inequality," says Lock.