Enterprises have started to realize the value of Service Oriented Architecture (SOA) and have adopted it as a mainstream enterprise IT architecture. However it's important for enterprises to realize that SOA presents a significantly different business and technology adoption model from a management perspective.

Business Service Management (BSM) is an approach to managing and making visible the impact of changes in business and IT. The IT infrastructure library is a framework for managing IT services. In this article we will explore the challenges in managing services in SOA and propose a way for BSM and ITIL to be used to build a platform for managing services in an SOA ecosystem.

Service Management in Enterprises
Broadly put services can be categorized into two types, services offered by the business and services offered to the business. A service offered by the business refers to the way capabilities are organized so they add value to the customers of the business. Services offered to the business are those that enable the enterprise to carry on its daily operations. It's important to note that services offered to a business diminish in significance if they're unmanaged and can't guarantee optimal performance, in turn greatly affecting the services offered by the business.

From the IT perspective , enterprises offer their business services through processes, applications, systems, and networks. All these are lined up to meet specific business needs. Traditionally service management has referred to the services offered by the IT support community that managed the software assets as well as the infrastructure. The most widely accepted framework for this IT service management (ITSM) has been the popular IT Infrastructure Library (ITIL) approach. Some ITSM implementations have begun to look at managing the processes and information part of an enterprise besides the IT support services. Complementarily, Business Service Management (BSM) is a process that provides visibility into the activities performed by the business through measurement and monitoring.

BSMbased approaches for measuring services offered by the business are vital for handling dynamically changing businesses, but they're still local in nature and aren't comprehensive enough to handle the management of all parts of the enterprise as required in the face of business changes. The divide between business and IT is the root cause of this so a bridge should be built between business and IT that finds comprehensive ways of addressing the management of the services offered to the business alongside the management of the services offered by the business.

The IT Landscape in the Service Ecosystem
SOA stands on the twin pillars of reusability and interoperability. The "reusability" factor derives from provisioning reusable business services that can be consumed by a number of players both within and at the edge of the enterprise. SOA promotes at its roots a loosely coupled architecture as opposed to the point-to-point-based tightly coupled architectures in traditional enterprises.

SOA can be manifest at multiple levels of enterprise IT, namely resource, application/data, middleware, and the processes that result in a layered architecture as shown in Figure 1.

Services created at each of these layers cater to top-level business functions. However, an important consideration even in such a service ecosystem is that the traditional applications that are either custom-built or commercial off-the-shelf (COTS) would still continue to co-exist with new composite applications that are process/service-centric.

Another aspect of this paradigm shift is that business processes no longer exist in a single application but might be created by assembling services provided by multiple units in the enterprise and stakeholders outside the enterprise. As a result, there are multiple owners in a service ecosystem each with service owner owing a part of the business process. In addition there are other shared infrastructure elements in a SOA, for instance, the enterprise service bus infrastructure owned by corporate IT that provides value-added services such as translations, content-based routing, reliable messaging, message tracking, etc. Hence the shared features of a process like security, reliability, change management; incident management, etc. depend on the service providers of these shared infrastructure/services.

SOA Management in the Enterprise
Traditionally the application-centric or infrastructure-centric view of IT management has been adopted in enterprises. In the existing IT management approach an application usually has a known set of users and the application generally implements a single business function where the application in question is deployed on a single piece of infrastructure.

In such a scenario IT management is simple, involving the monitoring of the related infrastructure and applications, which can be taken care by very high-level SLAs (application or infrastructure level). In the event of a change or incident at both the infrastructure and application level only a certain set of users is affected and the change/incident is limited to the application or within infrastructure boundary. Hence policies to mange such incidents and related changes aren't complex.

However, in a service-based ecosystem processes span services deployed on multiple machines, where diverse and large numbers of users consume these services. In event of an incident or change the effect is no longer limited to a service or piece of infrastructure, for instance, if a service interface changes (say a new release) then all the processes that consume the services need to incorporate the change. Hence the existing approach of application- or infrastructure-centric management of IT isn't adequate for managing services.

Service management needs to take into account the impact of layering in the service ecosystem. For instance, resource tier services can be managed by the enterprise infrastructure team, services created at the business service layer can be managed by business owners whose SLAs are in turn driven by individual application level service SLAs that are managed by the service owners. This clearly identifies the need for properly demarcating service management for each layer and across the different stakeholders. This requires a top-down view of SOA management, linking business processes and the underlying IT resources.

BSM and ITIL for SOA Management
The value proposition of using BSM and ITIL for a service ecosystem is that combined they provide a comprehensive approach to SOA management.

In this IT-enabled business world these two together provide a recipe for Business Technology Optimization (BTO). Inherently both BSM and ITIL are driven from the business perspective and take a business-centric view of IT capabilities in an enterprise. As SOA is adopted to align business and IT, both BSM and ITIL are the right fillers to help enterprises achieve true agility. Figure 2 shows a comprehensive SOA management approach with BSM and ITIL.
Together, they provide the ability to:
• Map business objectives the way IT operates, facilitating better visibility
• Create a process-driven ability for management that offers better control
• Create mechanisms to deliver IT operations as services to administer ITSM effectively
• Create a common vocabulary for the management activities for better communication