For network administrators, the goalposts keep moving. It is no longer enough that they monitor and adjust the availability and performance of network devices such as routers, hubs and switches. Now, they are increasingly being called upon to integrate their network management activities into an overall, service-centric framework. The idea is that with a unified service management approach, IT support staff can correlate performance data from network devices, platforms, applications and storage systems.

	Paul Arthur, BMC: “Organisations have become towers of technology in their monitoring.”

By doing so, they are able to see how the performance of each individual component impacts the delivery of the business services they support. This is the new world of service management. In terms of network management, it gives an organisation the capability to manage and "proactively" identify problems on its network, such as a bottleneck, to prevent the degradation of its online services. The approach will also help organisations to meet their service level agreements (SLAs) with customers more consistently, says Gerhard Haberstroh, European market impact manager at systems giant Hewlett-Packard (HP).

Certainly, software vendors sense a huge opportunity with service-centric network management. Since 2001, the largest vendors in the network and systems management market, including IBM's Tivoli division, Computer Associates and HP, have made significant moves into service management.

For example, HP acquired network performance management software supplier Trinagy, in August 2001. In addition, it unveiled a set of 45 product enhancements, and new products, as part of its HP OpenView Integrated Service Management strategy in November 2001. OpenView is HP's suite of network and systems management software.

Network management specialists are also keen to get in on the act. During 2001, Micromuse and Riversoft signed a number of partnerships with other technology vendors and expanded their product lines to add quality of service and network performance monitoring capabilities. And, application and systems management specialist BMC joined the fray by acquiring French network performance specialist Perform, in February 2001.

But aside from vendor marketing, just how big an opportunity does service management present? And how difficult is it to implement?

Total isolation
In the past, most organisations have managed their IT systems as if they were a collection of isolated segments. This has often led to a distinct lack of communication between departments or administrators working on specialist areas such as network devices, systems, applications or an organisation's help desk, says Paul Arthur, UK marketing director at systems and application management specialist BMC Software.

This problem becomes more acute when data about the performance of network devices cannot be tied to a specific business process such as a customer accessing their bank account balance online. "Organisations have become towers of technology in their monitoring," says Arthur. "The network operations people look at the routers, switches and hubs to gauge the performance of the network, but they cannot relate that back to the business process.

All they can say is that this router is busy… they do not know which end-users are being affected."

To address this problem, the large network and systems management vendors are claiming that their product sets encompass the full range of tools needed to provide end-to-end service management – and do so in an integrated manner.

For example, the Tivoli Business Systems Manager (TBSM) is server-based software that can provide a view of all of the software and hardware components linked to a specific application, claims Angus Jamieson, Tivoli's ambassador for northern Europe. "The graphical view of how boxes on a network are linked is easy to provide for hardware and network devices, but it is very difficult to do for logical applications," he adds.

This task can be made easier when organisations can set different performance priorities for different levels of an application, says Jamieson. TBSM allows an organisation to set levels one, two and three for an application such as online banking. "An organisation may not care if a system is running slowly for level one, but at level three it could generate a red alert because a gold card customer may require an application response time of 1.5 seconds per transaction. This could mean that if an organisation's bandwidth is struggling, the gold card customer could be provisioned over to a bigger server," says Jamieson.

Patchy uptake
Despite these dangers, analysts remain sceptical. Lance Travis, an analyst at IT market research company AMR Research, says he has seen little change in the number of organisations implementing service management products since the end of 2000.

Martin Rosenberg, a program director of service management strategies at analyst company Meta Group, believes that many early efforts are – at best – patchy in their approach. For example, "there may be a couple of routers on a network that are not covered by the organisation's network monitoring capabilities." These standalone routers, he explains, must be checked manually on a regular basis by administrators, as a traffic bottleneck on one of them could prevent critical data being made available across a network, but not show up on the IT department's systems and network management console.

Another problem, Rosenberg adds, is that, "vendors, in general, have been successful in providing an end-to-end view of an application and response time measurements, but what is difficult to achieve is to do something about a problem – and in real-time. In the best case scenario, an organisation may be monitoring everything, but they need some commands at the technology level to automatically or manually reconfigure a router, for example."

	Angus Jamieson, Tivoli: “Organisations have to be able to prioritise network response.”

Only when vendors add more sophisticated, automated mechanisms for resolving network problems will these service management products become a compelling proposition, says Rosenberg.

Naturally, suppliers are keen to play down the technical complexities of service level management, preferring to highlight the business benefits. "Service management is not technical. It is about improving your business, customer relations and driving down costs," insists George Bathurst, UK marketing manager at Hewlett-Packard. A major benefit, suppliers say, is that service management tools can help IT executives and business executives agree about the level of service provided to end-users, be they customers, trading partners, or internal employees.

For example, the business side of the operation may receive numerous complaints about a particular online service. With tools such as Tivoli's TBSM, an IT manager can provide a detailed report on network performance that shows that, while that service can boast 99.9% availability, it performs slowly at certain times of day when traffic volumes are heavy. With this information, business managers are provided with valuable data about customer behaviour, while IT managers are able to identify where network ‘tweaks' are likely to enhance the performance of that service.

This highlights the fact that while suppliers may be disingenuous to downplay the huge technological complexity involved in delivering service management, the business benefits may soon start to justify the effort involved.