IT Management

Ever since Service-Oriented Architecture (SOA) came on the scene, it’s been the subject of much discussion and anticipation. Organizations that understand it well know it isn’t just another new-fangled technology, but an evolutionary discipline of application architecture and design. But whether SOA was pursued with a sense of responsibility or just as a response to the hype, everyone hoped it would become the next “killer app.”

Obviously, that hasn’t happened—partly because more tactical and pressing issues crept into the agenda of the CIO community and also because the road has been treacherous for large-scale SOA projects. Although most effective when applied with SOA, virtualization was one of those other bigger priorities, with a much clearer ROI proposition of its own. Often, it was pursued in isolation.

Large-scale SOA comes with the rather daunting challenge of coordinating activities across such tracks as:

  • Virtualization (by itself, and/or in the form of cloud computing)
  • Business Process Modeling (BPM)
  • Complex/Business Events Processing (C/BEP)
  • Federated IT/services governance
  • Master Data Management (MDM)
  • Identity management
  • Business Activity Monitoring (BAM).

Given that, and taking into account the tepid outlook for IT spending, one wonders if SOA will fade away, push along, or be an enabler for much-needed IT efficiencies during the tough times ahead. An April 2009 forecast from analyst IDC shows a “relatively robust market for service offerings related to SOA.” IDC forecasts that SOA-based services will grow by 26.5 percent to $16 billion in 2009 and that the growth will be sustained, rising to a projected global total of $33.7 billion in 2013 (see Figure 1).

Against that upbeat background, let’s consider SOA’s future in terms of business drivers and technical advancements.

The American Recovery and Reinvestment Act (ARRA), signed into effect earlier this year, makes available $787 billion for economic recovery and supports expansion of programs in a broad array of areas. In the technology arena, it includes funding for increased use of Electronic Health Records (EHRs) as well as expanded public access to broadband, wireless Internet and smart grid technologies, to name a few initiatives. IDC estimates that the ARRA will infuse roughly $100 billion into IT to support these programs over the next five years. Many of these initiatives will be comprehensive IT programs that could benefit from large-scale SOA projects.

Also driving activity related to SOA are several legislative and regulatory compliance initiatives such as:

  • International Financial Reporting Standards (IFRS)
  • Solvency II in Finance/Banking |
  • The New International Classification of Diseases (ICD-10) and Certification Commission for Healthcare Information Technology (CCHIT) in Healthcare.

IDC’s report also points out commercial efforts that could propel SOA adoption in a big way. One such effort is the SAP/Microsoft-led Banking Industry Architecture Network (BIAN) that’s specifically aimed at sharing domain and technical expertise on SOA in the banking industry. In addition, traditional business drivers, such as cost savings, improved business agility, real-time access to information, enterprise mobility, etc., are expected to continue to contribute to greater adoption of SOA.

In the technical arena, there’s never been a duller moment in activities related to fine-tuning and optimizing the standards and related design patterns. There’s so much debate and discussion around these topics that, to the naysayer, it looks as if SOA is falling apart. But a more practical assessment of this continuous improvement would lead one to believe that SOA is actually evolving to encompass a broader set of design principles that can be service-oriented, resource-oriented (Resource-Oriented Architecture [ROA]), and Web-Oriented (Web-Oriented Architecture [WOA]).

In this new world of SOA, lightweight and heavyweight approaches would be deployed in a complementary fashion to meet the distinct usage requirements (access, performance, and security) of the business. In the lightweight approach, resources are manipulated directly by HTTP’s simple verbs using a technique known as Representational State Transfer (REST). In the heavyweight approach, services are manipulated by a messaging layer over HTTP in the form of Simple Object Access Protocol (SOAP).

The technical architect must exercise a greater sense of responsibility by making the right choices to ensure the architecture is simple yet robust enough. Far greater consideration will be given to the emerging lightweight options (‘REST FIRST!’) in developing next-generation SOA applications. They score high in ease of access and expediency of development and lend themselves to simpler integration with the social networking technologies of Web 2.0, which often are central to the business expansion goals of new SOA endeavors.

A service or resource is no good unless it can be easily consumed. The whole discipline may well be all bundled up together as Consumer/Client-Oriented Architecture (COA). Meanwhile, some fresh ideas are emerging in the area of SOA governance for the design and run-time environments. The role of Universal Description, Discovery and Integration (UDDI) as the registry is being seriously questioned, and innovative, newer and lighter registries are emerging that are based on REST. Run-time governance also is facilitated by Really Simple Syndication (RSS)-like protocols such as Atom and the Atom Publishing Protocol (APP).

On the heavyweight side, standards such as WS-* are being simplified and streamlined to promote policy-driven SOA implementations for serious and secure enterprise applications. Modeling techniques and tools are being further refined to facilitate top-down SOA implementations that are vital to achieving improved business agility. Newer appliances and private enterprise clouds are addressing performance and cost barriers for these heavyweight SOA projects.

All of these business and technical developments bode well for SOA and should further encourage CIOs/CTOs to give it serious consideration as an application architecture and design discipline, There’s a strong case to focus on SOA and be ready to embrace its promise and potential as a pivotal technology in the years to come. ME