The service economy.

From the customer's perspective, it matters little that their online bank's mainframe system is experiencing availability problems. Similarly, if a mobile salesperson for a pharmaceuticals company cannot access his list of sales prospects as he is approaching an important meeting, the news that the network is down is of little comfort. For either user, the message is that they cannot access the service they need. For the customer, this might result in a switch to a competing provider. For the sales executive, it could result in the failure to win a valuable new account, or keep an existing one.

The realisation that poor IT performance can have a direct impact on a company's prosperity is not news for most organisations. However, it is a concept that is only now starting to be embraced by the team that manages the technology underpinning that service. Until recently, most IT departments have focused on the performance of individual components that support a service - the hardware, the network, and discrete applications - rather than the performance of that service as a whole. But this is beginning to change.

To manage their systems, organisations have historically implemented systems management tools from established vendors such as Hewlett-Packard, Computer Associates and BMC that enable them to keep an eye on elements of their infrastructure and respond if a particular component is not performing to scratch.

Now though, the emphasis is starting to migrate towards an end-to-end, 'service management' view of how these components perform together, and how they impact the service as a whole. "Customers are realising individual technologies do not sit in isolation, but are part of a wider service," explains George Bathurst, software marketing manager at systems and software giant Hewlett-Packard. "The customers or users accessing a service do not care whether the server is down, they just want to access their emails or perform whatever task they need to get on with their work."

This situation has been complicated by the fact that organisations now rely more heavily on certain software applications, and also expose a greater portion of their internal systems to the outside world as customers and partners increasingly use the Internet for business transactions. In addition, IT managers find themselves having to manage an ever-expanding number of components within the infrastructure, including application servers, and web and print servers, that are not normally dealt with by traditional systems management tools. The potential speed with which web-based applications can handle transactions has also raised user expectations. "Before, it never used to be an issue if you didn't make payroll on a certain day. Now, if there's a problem with a service it goes straight to the board," says Rick Sturm, president of US-based technology analyst company Enterprise Management Associates.

Too much information One of the problems is that technologists are used to supplying mechanical data such as network uptime, number of transactions per second or processor utilisation, says Sturm. A more service-centric view will require a change in focus away from the technology. "Often IT departments provide too much information - they might as well be talking Swahili to business people," he says. This is partly because IT departments have become used to defining complex service level agreements (SLAs) with external service providers or their organisation's own internal departments.

"There has been too much emphasis on metrics like 5x9 uptime," points out Paul Arthur, UK marketing manager at systems management software vendor BMC. "Businesses are now looking for metrics that speak the same language as they do - how long it takes to complete a transaction, how many attempts it takes to order a book, that kind of thing."

But how ready are organisations to move to this service-oriented view of systems management? According to a survey of its users in conjunction with analyst company Gartner Group, testing and systems management software company Compuware found that many respondents (37%) are still at a 'reactive' stage, where they detect problems in individual systems and fix them as they arise. Only 7% are at the highest level of service management maturity, where they are aware of the full impact of IT on their business processes. "Early ecommerce efforts highlighted the deficiencies in IT's ability to react to business issues. Businesses are being forced to become proactive," says Mike Lucas, product marketing manager at Compuware.

The major challenge for IT departments in moving towards service management, however, is that they must aggregate data from many different systems within their IT infrastructure. Furthermore, the IT function in many organisations will be made up of distinct teams of network administrators, hardware specialists and application developers, all used to dealing with one particular area. At package delivery company TNT, for example, IT manager Dennis Beard found that in order to report on the performance of the company's package tracking service as a whole, his staff would not only need to track internal systems such as its Oracle database, Unix and NT systems, but would also have to manage the performance and response of the handheld devices used by TNT's delivery team, plus browser-based interaction from its large corporate clients. Systems management tools vendors are now starting to address the service management shortcomings of their products. Hewlett-Packard, for example, offers a product called Service Navigator, which allows users to map business processes in a hierarchy and then align them with the technology underneath. Its core systems management tool OpenView then manages the technology components. Compuware, meanwhile, has adapted its Vantage systems management product (formerly EcoSYSTEM) to alert the IT department when a service is not functioning properly through a series of 'traffic light' indicators, rather than when there is a network problem, or a particular application has a bug in it.

However, organisations will still have to perform a significant amount of integration themselves. "The tools are slowly coming along," says Aidan Lawes, chief executive of the UK arm of the IT Service Management Forum (ITSMF). "But there are so many different technical environments to consider it becomes extremely complex. Most organisations still use point solutions, so will need to aggregate that data to get a complete picture." Lawes suggests organisations adopt formalised service management policies, such as the IT Infrastructure Library (ITIL), which draws out the stages involved in completing particular business processes, and has recently been adopted by Microsoft for the service management of Windows. Kim Gravell, a consultant at systems integrator PricewaterhouseCoopers, suggests businesses first work out what tasks they are trying to achieve - for example, 'Post invoices Friday' -- and then look at what processes have to be fulfilled in order to complete that task. This could be extracting data from customer order systems, printing invoices, and final checking by accounts, for example.

Only then, says Gravell, should organisations begin to look at the performance of the hardware, networks and applications that support these processes. "This will be a bit painful to start with, as most organisations tend to get bogged down with technology," she warns.

Organisations should not ignore the performance of individual components within the infrastructure altogether though, warns Marcel den Hartog, European marketing manager at systems and network management company Computer Associates. At the technology level, organisations should assign 'weight factors' to the individual components of the network, he advises. This will help managers see how the failure of certain aspects of their IT infrastructure impacts the service as a whole. "There's no need to bring down an entire business service just because one of the nodes on the network is functioning more slowly than usual," says den Hartog.

With this service management framework in place, IT departments are able to supply business managers with a more accurate view of how the end user perceives an organisation's service, while still keeping one eye on how individual systems are performing. But the performance data that arises from having a holistic view of a particular service is where the real business benefit lies, according to Paul Arthur of BMC. "It's not just about doing the management but about reporting back to customers," he says. "For example, an organisation could see whether a 10-second decrease in response time is worth the extra $10,000 it had to fork out to increase bandwidth. Furthermore, if it knows how much more the customer is prepared to pay for a certain improvement in performance, it can gauge it accordingly."

At one company, service management had an additional benefit. After pharmaceuticals giant Procter & Gamble implemented a services management strategy, says Aidan Lawes of ITSMF, the company discovered that it was experiencing a lot of repetition of tasks, and as a result, was using 40% more resources than it needed. This meant it could free up resources to work elsewhere. Costs went down, and staff productivity went up. The department that got the credit for it was IT. Staff began to ask the IT department to help with enhancing services, rather than just fixing problems. "The IT department felt its standing had increased in the organisation," says Lawes.

Service management: Key issues

According to US-based analyst company Enterprise Management Associates (EMA), there are a number of issues organisations need to consider when choosing tools to manage their service provision. A product suitable for one environment, says EMA, may be entirely inappropriate for another, so organisations should consider the following issues:

Contracts & agreements Many organisations will have internal or external SLAs that focus heavily on technical metrics, so they need to develop strategies with colleagues or trading partners that ensure they foster an atmosphere of constructive collaboration and trust in which keeping the SLA 'in the drawer' becomes acceptable.

Alignment with business objectives Technical performance is important when managing service levels, particularly where service level agreements (SLAs) are involved. However, on top of this, organisations should ensure that their technical requirements directly support their overall business objectives.

Integrated, end-to-end management Many aspects of the IT infrastructure play a role in service delivery, so when measuring the performance of a particular service, service management tools must be capable of looking across the entire infrastructure supporting it.

Accountability For the individuals in charge of a particular service, it is key to their job security and credibility that the service performs as well as possible. This is an issue that organisations still have to tackle on a broad scale, according to EMA.

Role of the Internet The breadth of the Internet means that neither organisations nor their external service providers can control the entire delivery path of their services. Issues such as 'last mile' network provision mean the experience will often be inconsistent from one user to another.

Hands on: Avaya

When corporate networking specialist Avaya was spun off from its parent company Lucent in May 2000, it had the opportunity to completely rethink the way it handled its systems and network management.

Although Lucent had used Hewlett-Packard's OpenView network and systems management platform for some years, adding new management functions as HP developed them, it found that it had created "islands of monitoring" - different monitoring platforms that could not communicate with each other - dispersed across its IT support teams around the world. This meant it could not get a consistent picture of how its infrastructure was performing as a whole.

As Avaya cut its technology infrastructure ties with its parent company, it realised this would be an ideal opportunity to harmonise its systems management strategy and develop a more service-oriented approach. "It was what our end-users wanted," explains Jean-Philippe Draye, a systems architect at Avaya. "For example, the SAP users in the finance department don't ask us to ensure their CPU capacity runs to 100 people - they just say they want to be able to use SAP all the time. It's a service."

This realisation spurred Avaya to invest in HP's Service Desk application, which sits on top of the OpenView operations and network management systems Avaya inherited from Lucent after it was spun off. Service Desk aggregates 'tickets' that indicate the location of problems from any of the underlying systems. So, where a problem with overcapacity in a server would previously have generated an alert in OpenView Operations, it is now routed straight to the Service Desk. The HP platform manages all the network devices, routers, remote access devices, web servers and content switches within Avaya's vast IT infrastructure from a central console in its Internet operations centre in Orlando, Florida.

Although Avaya inherited certain aspects of the OpenView platform from Lucent, which went some way in making the implementation path easier, it did also evaluate other vendors. Draye says Avaya chose the Service Desk application because it was based on the formalised service management policies laid out by the IT Infrastructure Library (ITIL), and would integrate "almost out of the box" with Open-View. "We wanted to move from element-based monitoring to service-based monitoring, which this system allowed us to do," says Draye.