Technology Industry
Industry: Email Alert RSS FeedToward an on demand service-oriented architecture
IBM Systems Journal, March, 2005 by C.H. Crawford, G.P. Bate, L. Cherbakov, K. Holley, C. Tsocanos
Over the last 40 years, information technology (IT) architectures and development approaches have dealt with increasing levels of IT complexity and integration challenges. Constrained budgets continue to mandate that legacy systems be reused rather than replaced. Growth by merger and acquisition requires that entire IT organizations be integrated and absorbed. Additionally, easy access to the Web has created the possibility for new business models, which must be evaluated for their potential.
- Most Popular Articles in Technology
- An overview of continuous data protection
- Why all those current ratings?
- Many countries now have a mobile penetration rate above 100%, report says
- The Tata Group's big telecom gamble: VSNL's recent acquisition of Tyco ...
- MEASURING BANK BRANCH EFFICIENCY USING DATA ENVELOPMENT ANALYSIS: MANAGERIAL ...
- More »
At the same time, the traditional needs of IT organizations persist--primarily focused on quick response to new requirements, typically consisting in turn of corporate management pushing for better IT utilization, skills simplification, greater return on investment (ROI), continued integration of historically separate systems, and faster implementation of new ones. The endless varieties of hardware, operating systems, middleware, languages, and data storage create an environment where accommodation of heterogeneity is fundamental to system development. The cumulative effects of decades of growth and evolution, encompassing multiple computing architectures, programming languages, and connectivity products, have produced the complexity that now challenges all IT organizations.
As the level of complexity continues to increase, traditional approaches are reaching their limits.
Powerful new technologies such as Web services, (1) autonomic computing, (2) utility computing, (3) and grid services (4) provide partial answers, but the problem in many cases is the lack of a consistent architecture within which applications can be rapidly developed, integrated, and reused. SOA (service-oriented architecture) holds the promise of being the consistent architecture required for future development. In isolation, however, neither the application of these new technologies nor the use of the SOA approach, itself, provides a complete solution.
ON DEMAND SOA
Much has been written about why SOA, Web services, autonomic computing, utility computing, and grid technologies and standards can be beneficial, but a holistic view, unconstrained by technology, is currently lacking. On demand SOA provides such a view, namely a distributed computing model infused with the building blocks of these new technologies. Taken separately, each provides significant benefits, but the integration of these technologies promises far greater impact and provides the foundation for on demand SOA.
To demonstrate on demand SOA, we use a real-life example of a business-to-business (B2B) process in which services are used to automate the processing of an electronic purchase order request (POR). Our buyer, Acme, Inc., is a large manufacturing company. Our supplier, Pens R Us, is a large stationery company. A contract already exists between the two parties. Acme, Inc. wants to use an electronic POR to buy 500 reams of paper from Pens R Us. The first step is to determine which services are required to fulfill the process. At one extreme POR may be viewed as a single service; at the other extreme the service granularity could be so fine that the POR might be constructed from multiple services. The choice is made by balancing quality of service (QoS) characteristics, ubiquitous service reuse, and reduction of complexity for service composition.
The supplier may choose to view the process as the following steps:
1. Supplier authenticates the purchaser.
2. Supplier looks up the buyer contract based on purchaser ID.
3. Purchaser browses the product catalog with negotiated prices from the contract.
4. Purchaser adds items to the shopping cart.
5. Purchaser checks out, providing payment description and delivery information.
6. Order information is sent to the fulfillment department.
7. Confirmation of order with expected delivery date is sent to the purchaser.
From such a process description, we can further list the software or application services that are required:
1. Login
2. Contract lookup
3. Catalog browsing with shopping cart and checkout
4. POR data creation (from login ID, contract ID, shopping cart data, and other information supplied by checkout)
5. Information delivery to fulfillment process
6. Message to purchaser to confirm order
The enablement of SOA with open standards, for example, Simple Object Access Protocol (SOAP), Extensible Markup Language (XML), Web Services Description Language (WSDL), and Open Grid Services Architecture (OGSA), among others, offers the ability to fulfill the promises and value propositions of SOA implementations. These open standards allow a service to be decoupled from the IT infrastructure. As long as vendor support for the standards exists across resources, the service composer need not be concerned with where the service will run--only with how to assemble flows between services. Additionally, dynamic service lookup means that service consumers need not be concerned with where underlying software resources exist on distributed, heterogeneous systems.