Realising the promise of SOA and BPM

Realising the promise of SOA and BPM Madan Sheina October 2008 1 REALISING THE PROMISE OF SOA AND BPM Realising the promise of SOA and BPM Ensuring success with process-driven data integration and data services Much of the technical effort around implementing service-oriented architectures (SOA) and business process management (BPM) systems has centred on application integration and workflow. But many enterprise architects are now grappling with performance issues caused by an inability to access business data and to move it around in the same agile way that they handle their business application logic. Process-driven data integration can help to enrich and quicken up SOA business services and BPM processes through sophisticated sequencing of data services into meaningful reusable combinations that incorporate human task intervention to ensure that they are being fed accurate, consistent and timely information. Overview The Ovum view Service-oriented BPM is a transformative force in IT that offers enormous potential for enabling organisations to achieve greater agility across their applications and business processes. But its success is only as good as its processes (web services), data foundation (data services) and overall SOA design. Most business processes supported by SOA are inherently data-driven. Yet many organisations don’t intuitively think about their SOA or BPM problems as data problems. Far too many SOA initiatives are undertaken without the foundation of a data services layer, resulting in a further proliferation of the siloed systems they were meant to integrate and consequently hampering business agility. A SOA web service that is orchestrated by BPM often assumes the availability of accurate, consistent and timely data. But that can’t be taken as a given in all cases. Companies therefore need to ‘wake up and smell the data’ if they want to get full payback from their SOA and BPM initiatives, which means transitioning from a siloed application-centric view to an enterprise data-centric view. It requires an up-front emphasis on data integration to ensure composite SOA applications and processes are fed by authoritative and trusted data sources. © Ovum 2008. Unauthorised reproduction prohibited 2 REALISING THE PROMISE OF SOA AND BPM Realtime data integration technologies certainly help to feed SOA services and business processes with the right data at the right time in order to make them work as intended. But rising data volumes coupled with a realtime imperative and increased complexity of an organisation’s applications and business processes does not make this a trivial task. There are simple business processes and complex ones. Likewise the underlying data integration processes might be relatively simple, requiring a single data feed, while others can be more complex requiring access to an integration of multiple data sources. Moreover there are business processes that are suited for complete automation while there are others that require varying degrees of human workflows. So while SOA needs to become more data aware, data integration as a process can benefit from a greater SOA and BPM awareness. In the same way that BPM orchestrates web services at the application layer, it can – and should – also perform the same function at the data services layer. But that can’t be achieved by extending an existing BPM tool. While traditional BPM tools can access data that is exposed as data services, they are not designed to orchestrate its integration at the data layer. Nor are they well-equipped to handle the often complex, multi-faceted data treatments that a dedicated data integration platform provides. Moreover, process and data integration are often handled by separate teams using different tools with little or no interaction between them. BPM projects tend to be narrowly focused along line-of-business functions. That makes it difficult to construct data services that can be used across enterprise processes. What’s needed is a process-driven approach that grows out of the data layer. Process-driven data integration builds on a foundation of a unified and scalable standards-based, data integration platform that can sequence sophisticated data services into meaningful orchestrations removing the burden and complexity of data manipulation and integration tasks from the SOA and BPM layers. That is where the benefits of a process-driven data integration approach can pay great dividends. And because process-driven data integration follows the sound principles of SOA and BPM design, it manages dependencies as the data integration process is further abstracted away from the data. The result is a reusable and easily consumable set of data integration services (i.e. access, cleanse, integrate, and deliver data at any speed) to ensure the success of SOA and BPM initiatives. Squaring up to a tough macro-economy Business enterprises today compete in an increasingly tough and uncertain global economy. Constantly changing macro-economic market conditions such as globalisation, mergers & acquisitions, rising customer expectations, stringent corporate governance & regulatory compliance and © Ovum 2008. Unauthorised reproduction prohibited 3 REALISING THE PROMISE OF SOA AND BPM faster product/service development & delivery are pressuring companies of all sizes to ‘do more with less’. Overcoming these challenges requires companies to be nimble and agile to safely navigate a steady course through rough waters and meet strategic and operational performance goals and objectives. In response companies are leveraging IT in order to maximise the value of their business data and processes to control costs, mitigate risk and be more adaptive, agile and responsive to evolving customer demands and approaching business opportunities and threats. To achieve those goals companies need to be able to access and utilise relevant and accurate information about their business activities and environment in a timely way. This challenging mix of data integrity and data agility (i.e. the timely movement and access to clean, consistent and complete data regardless of its location and format) puts increased demands on corporate IT infrastructures, which are now being asked to bend in ways they were not designed for. In order to address modern-day challenges of business agility, the underlying IT infrastructure of today needs to provide a single, comprehensive view into all business-critical information assets. It also needs to seamlessly handle the complexity of all enterprise data, its varying volume and latencies and its many formats and structures. Data pays dividends for agility Many business processes across industries are inherently information- driven. Data drives most processes running in organisations today. Data is the lifeblood of many enterprise processes to drive its products/services, customer relationships, business partnerships and internal regulatory compliance efforts. An increasing proportion of these enterprise processes relies on data from multiple sources that, in many cases, originate across departmental and enterprise boundaries. Providing timely and trusted data is key to getting visibility into these processes and making them more agile. For example, a manufacturing organisation can better manage its inventory by accurately shrinking order lead times by transparently accessing and integrating business-to-business (B2B) data flowing across its supply chain. Retail organisations that have fingertip access to up-to-the-minute sales data can better plan replenishment schedules and eliminate potential stock-outs. And with a single view of the customer, financial services organisations can effectively cross-sell and up-sell additional services. Business processes that need to be effective in driving competitive advantage thus require the integration of a wide range of data sources, applications, protocols, people and partners. By design these processes must be able to access a variety of applications and different types of data at different times to drive business agility. © Ovum 2008. Unauthorised reproduction prohibited 4 REALISING THE PROMISE OF SOA AND BPM The ability to effectively execute requires the systems and people involved in the business process to have a complete, consistent and accurate view of information about business activities. Hence enterprises must ensure that trusted information on products, customers and partners is available for use by these critical business processes in a timely (i.e. right time) manner. Data challenges Unfortunately that’s a luxury rather than the norm today – unless of course when the process itself runs within the confines of an application, such as an enterprise resource planning (ERP) or customer relationship management (CRM) system, where all/most of the data required already resides. However, the ambition to manage many processes across multiple application domains involving customised process steps and workflows that span multiple data domains highlights a critical need for data integration – a task that is exacerbated by the rising complexity and diversity of today’s corporate IT systems. The heterogeneity of data sources distributed across the enterprise and beyond is a big problem.