Agile IT systems are systems that are malleable enough to address business uncertainties. Such systems can effectively respond to internal and external stimuli in a very short period of time. Flexible IT systems imply that the IT architecture underlying them is itself flexible and lends itself to incorporating changes in a dynamic fashion. Architectural approaches such as Service Oriented Architecture (SOA) are transforming the way IT systems are designed by bringing in a high degree of reuse and loose coupling of applications.

IT architectures include the technology, strategies, plans, and principles that guide an organization's new technology investments as well as manage the existing technology investments.

The objective behind having a good IT architecture is so an organization can operate with a high degree of flexibility while keeping the cost for the technology investments within justifiable limits.

IT architectures are derived from the business architecture and once the architecture is defined, it supports the business. Organizational flexibility is a key requirement supported by IT architectures. Organizational flexibility and agility is emerging as a key constituent of sustainable competitive advantage. The business processes supporting organizational activities have to be flexible enough to enable organizational flexibility. Business processes are dependent on the underlying IT systems for execution. These IT systems and architectures should be flexible enough to deal with uncertainties in an unstable environment. They have to respond rapidly to internal and external stimuli. Agility refers to an organization's ability to thrive in a continuously changing business environment. A flexible enterprise is qualified by the cost and the speed of response to changes in business models, business opportunities, and market conditions.

The Business Value of SOA-based IT Architectures
The new enterprise architectural approach termed Service Oriented Architecture (SOA) envisages delivering IT functionality as services over a distributed network. The idea behind SOA is to treat business and IT functionality being delivered as a set of services, wherein the services are self-contained and independent of the context or state of other services. SOA-based systems can be seen as a collection of services having well-defined interfaces. In technical terms, SOA is an approach to loosely coupled, protocol independent, standards-based distributed computing where the IT resources are available on a network as services. The key idea differentiating SOA from earlier models of distributed computing lies in the notion of a service as the least common denominator. Here, a service refers to a piece of functionality, which is defined by a strict contract via a well-defined interface that is independent of any underlying implementation platform of the service. This kind of interface is loosely coupled in that a service is a standalone entity with no tight coupling to the underlying environment or to the other services. These characteristics are key to delivering the required flexibility for enterprises by adoption of SOA.

Service Oriented Architecture has three key constituents: service provider, service requestor, and service registry and three fundamental operations: publish, find, and bind. A service provider makes the service available and publishes the contract that describes its interface. It then registers the service with a service broker. A service requestor queries the service broker and finds a compatible service. The service broker gives the service requestor directions on where to find the service and its service contract. The service requestor uses the contract to bind the client to the service. The connection between these entities is loosely coupled offering the maximum decoupling between any two entities. For the loose coupling it is mandatory that the services be described in ample detail to be self-describing and self-contained. These factors lead to the natural question of how to measure the business value of SOA based IT architectures, which has a combination of unique characteristics, prominent among which include re-use, flexibility, and open standards.

Cost and Benefit Dimensions
Our primary research revealed six cost dimensions and five benefit dimensions for the posited relationship between the costs, benefits, and business value. (See Figure 1.) The cost dimensions are related to the efforts required for an enterprise-wide IT effort such as SOA, while the benefits are potential rewards of the SOA deployment.

The cost dimension is comprised of variables such as Organizational, Business Process, Technology, Servicification, Integration and People. The benefit dimension is comprised of operational measures such as Re-use, Flexibility, Integration, Open Standards and People. Table 1 gives the definitions of these dimensions.

On the basis of our research and interviews with industry experts, we have identified the following propositions regarding the cost dimension:

P1 Organisational, business process, technology, servicification, integration, and people costs are the most important cost variables influencing SOA architecture decisions

P_1A Among organisational costs, understanding business context is the most important cost variable influencing SOA architecture decisions

P_1B Under Business Process analysis costs, identifying business processes that can be shared is the most important cost variable influencing SOA architecture decisions

P_1C Among technology costs, application portfolio analysis is the most important cost factor influencing SOA architecture decisions

P_1D Among servicification costs, determining the optimal level-of-service granularity is the most important cost variable influencing SOA architecture decisions

P_1E Among integration costs, interface development is the most important cost variable influencing SOA architecture decisions

P_1F Among people costs, ensuring business continuity is the most important cost variable influencing SOA architecture decisions

The following were the research propositions regarding the benefit dimension:

P2 Re-use, flexibility, people, integration, open standard benefits are the most important benefit variables influencing SOA architecture decisions

P_2A Among re-use benefits, business process re-use is the most important benefit variable influencing SOA architecture decisions

P_2B Among flexibility benefits, time to market is the most important benefit variable influencing SOA architecture decisions

P_2C Among people benefits, development of a standardized employee skill set is the most important benefit variable influencing SOA architecture decisions

P_2D Among integration benefits, inter-organizational collaboration is the most important benefit variable influencing SOA architecture decisions

P_2E Among benefits from open standards, platform and technology independence is the most important benefit variable influencing SOA architecture decisions

Conclusion
In this research, we have studied the relationship between IT architectures and business value. By breaking down the cost constructs into variables such as organizational, business process, technology, servicification, integration, people and benefit constructs into variables such as re-use, flexibility, integration, open standards, and people, we attempted to find the most important cost and benefit variables that have to be considered while measuring the business value emerging out of SOA-based IT architectures. Our study reveals that the most important cost variable in SOA-based architecture implementations is the identification of new business services. This is followed by variables such as understanding business models and strategies, organizational buy-in, service realization through interface development, identifying existing services, and identifying the optimal level of service granularity.

The most important benefit is derived from low vendor lock-in (allowing enterprises to replace customised applications and products). This is followed by variables such as platform and technology independence (leading to greater collaboration and integration, with less spending on middleware infrastructure), time to market (which enables enterprises to come out with products/services in a shorter timeframe), adaptability to emerging business scenarios (which facilitates the enterprise to realize faster growth by speedier and responsiveness to market conditions), benefits due to a reduction in application redundancy (emerging from rationalization of applications), and re-use of business processes, applications, and infrastructure (through the potential reuse of services and infrastructure), standardized employee skill sets (as the SOA interfaces is abstracted from the implementation), benefits from loose coupling (as SOA enables systems to be assembled and disassembled easily as they have less dependency on other systems). Finally there are benefits from reduced cost and time of introducing new business processes, reduced cost and time of modifying existing business processes, reduced cost and time of introducing new applications, reduced cost and time of modifying existing applications, reduced cost and time of introducing new IT infrastructure, and reduced cost and time of modifying existing IT infrastructure.

Our research implies that the IT investment decision makers need to have a comprehensive understanding of the cost and benefit variables adopted in the study to arrive at the business value measure to justify SOA-based IT architectures. Understanding the role of SOA-based architectural approaches in achieving flexibility across the three dimensions can guide appropriate investments in flexible IT systems and strategic IT planning. Studying the influence of IT architectures on organizational flexibility poses a number of challenges. In understanding the impact of IT architectures on organizational flexibility, we have studied business processes, IT applications, and IT infrastructure as the three dimensions to capture organizational flexibility. But organizational flexibility could be influenced by a number of other factors.