• United States



by No Analyst or Consultant

A Services-Based Approach to IT

Jun 27, 20036 mins
CSO and CISOData and Information Security

“Many of the inquiries META Group currently receives from clients relate in some way to the concept of reusable IT ‘services,'” says META Group analyst Daniel Sholler. “Clients have invested large sums in technology, and they are searching for ways to maximize the return they get from those investments. To achieve this, they want to delegate tasks to shared services to maximize reuse and efficiency. These services include not only operational groups and processes (such as the IT help desk for its problem resolution process), but also shared infrastructure hardware and software components. In addition to being shared and delegated, an essential attribute of these services is that there is a defined interface for how to invoke them.”

IT operations is the domain where a services-based approach is best established (such as providing central IT services, not unlike an outsourcer). Help desk and desktop management groups are examples of shared operational services. These services are typically provided by IT personnel with the assistance of technology tools, and are provided to IT users throughout the organization. Leading IT operations groups have a service catalog that specifies which operational services are available to business users and how they can be obtained and paid for. Established process/service catalogs, such as the IT Infrastructure Library, are valuable in advancing a service-based IT operations structure.

IT technical or infrastructure services consist of hardware and software infrastructure that is built, designed, and operated separately from individual applications and shared across multiple applications. Examples include Web server farms, storage-area networks, shared databases, and enterprise application integration (EAI) middleware. The “consumers” of infrastructure services are groups within the IT organization, such as an application development project team, that must use these shared infrastructure components to enable their applications. Although infrastructure services are well established for specific purposes such as shared WAN, LAN, and remote access as well as shared DMZ and Web server farms, a broader-scope application of this approach has been adopted systematically only by a few leading IT organizations.

“An inventory or catalog of available technical services enables the design of repeatable infrastructure patterns (for example, n-tier transact solutions, repeatable blueprints) that can be decomposed into new components to buy, such as a new DBMS server, and existing shared solutions to be leveraged, such as a Web server farm or WAN,” says META Group analyst Bruce Robertson. “Shared technical services must be mapped into end-to-end infrastructure pattern blueprints to ensure optimal reuse during application delivery projects.”

Business application services are capabilities that can be shared among multiple implementations to achieve similar purposes. These services can be large-grained, such as entire software packages (ERP, CRM, for example), or key modules, where the customer is the business unit. However, a more composite application approach will increasingly define finer-grained software components (such as parts of applications) that can be integrated into different solutions to achieve a particular task (for example, a credit card authorization routine or pricing engine).

Although business application services have the least mature services-based approach, the most value can be gained from their increased reuse. The introduction of XML-based Web services technologies will enable significantly increased integration of software components and drive cultural change toward reuse and service-oriented architectures (or designs) in both business application components and hardware and software infrastructure.

“Many clients are struggling to define how granular business application services should be,” says META Group analyst Janelle Hill. “If the service is defined too broadly, sharing can be inhibited. On the other hand, if the service is too narrow, its usefulness and impact will also be limited. I advise clients to ‘think big’ – to look for the greatest amount of agreement they can get among different groups on the definition of a service that can be shared.”

As delegated tasks, all three types of services – operations, infrastructure, and applications – should be defined by what they provide to the user via a service interface, not by the technical details of how the task is achieved. “The service definition should include what the service does and does not provide, who the users are, and what service levels will be provided to different classes of users,” says Robertson. “Although the service provider will need to define the technical details as part of the implementation, this is less important than defining the interface. The interface is the key factor in determining agility: how easily the consumer and provider can change without deranging their dependencies.”

In many cases, the mature shared services that are deployed will be higher-order aggregate services that combine technical, operational, and/or business application services created by different IT groups. For example, EAI technical services (server and software) should be explicitly mapped to operational services required for EAI (format and routing policy updates, monitoring and reporting, server administration, capacity management, administration/support, and plan/build services).

The shared-services concept is hardly controversial. Many IT operational processes, application functions, and infrastructure components can be shared by different “customers” or used for different purposes to reduce costs and enhance overall performance. The challenge is not conceptual, but practical: how to adapt the IT organization to enforce and manage reuse of services across the three domains of operations, infrastructure, and applications.

“Sharing and delegation of services force the IT organization to operate differently,” says Robertson. “The organization needs to face head-on the proposition that central IT must actually exist in a strong and meaningful way. The various groups within IT must manage according to the same goals: creating reusable services that increase value, establishing reasonable mechanisms for users to pay for these services, and ensuring that these services are used and managed appropriately. The mandate is to move away from complexity toward simplicity and from a distributed approach toward a consolidated one – enforcing this requires strong leadership both inside and outside the IT organization.”

As leading IT organizations leverage Web services and instill best practices throughout their enterprises to maximize service sharing and reuse, the pressure on others to emulate this approach will increase. Consequently, META Group believes the trend toward a services-based approach to IT, still only in its early stages, will be one of the most important developments in the industry over the next decade.

USER ACTION: IT organizations must inventory and define their catalogs of operational, technical, and business application services. Each service definition should stipulate the appropriate users and use cases for the service, the service interface, and the service level provided as well as the value structure, chargeback and funding approach. These service catalogs should be documented and “marketed” to the intended users both inside and outside the IT organization.

CIOs should charge their enterprise architecture teams as well as the heads of their infrastructure, operations, and application development groups, and their top executive lieutenants (e.g., IT transformation leader, IT controller) to inventory the various types of reusable IT services in the organization, optimize how different services are aggregated and consolidated, increase the level of reuse within the organization, and measure the efficiencies obtained.

META Group analysts Daniel Sholler, Bruce Robertson, Janelle Hill, Karen Rubenstrunk, Craig Roth, and Hollis Bischoff contributed to this article.