BPM-1 is finished. The second generation of Business Process Management software is coming. Like its predecessor, it focuses on building and integrating agile software driven by business processes. Unlike its predecessor, it could succeed.

	express delivery
	• print  • comment
	BPM-2 is designed to be a digital framework describing your business and systems. Think of it as a nerve center that sends signals to all of the organs of your system. Jeff Schneider is CEO of Momentum Software, which provides consulting services for IT architecture development, business process optimization, application development and systems integration. Read more of Jeff's views on the 'service oriented enterprise' in his weblog.
	Glossary terms: BPM, business process, packaged software, BPEL4WS, web services, lookup tool Sponsored links: your text link here

BPM-2 retains the best of what BPM-1 offered:

• The process drives the software, not the other way around (which is the major complaint about the large packaged software vendors).
• It operates at three layers:
  • Process maps support strategy and business flow;
  • workflow drives human flow;
  • integration orchestration drives system-to-system flow.
  
  
• The process is described digitally so that it may be executed.
• Legacy applications exist and must be woven into processes through integration facilities.
• Business processes have boundaries, goals and metrics; and these items should be visible to the management team.

Many of the first-generation BPM tools adopted these concepts. They failed because — driven by customer demand — they went far beyond this list. They provided facilities for transactions, security, presentation, transformation and many others. In essence, they became large monolithic layers that sat on top of application servers, tightly glued together in order to facilitate rapid creation of applications according to a business process. The result was really a next-gen rapid application development (RAD) tool, which was tied to a specific platform, usually Java.

Just as the ERP vendors had been accused of providing 'Big Bang Functionality', the BPM-1 players provided 'Big Bang Architecture' — everything you need in one tidy box.

So what is wrong with Big Bang Architecture? Perhaps nothing, if all of your applications choose to use it. In most organizations, this just isn't realistic. It is for this reason that we have recently seen more emphasis on 'Point Architectures' that are more granular in nature and may be woven together at runtime. Service oriented architectures (SOA) and web services promote the concept of having stronger separation between the architectural components. Thus, the application is described not by using platform specific terminology (JMS, EJB, .Net remoting, etc), but by referring to interfaces to the services that provide the functionality — regardless of the language, platform or vendor.

BPM-2 decouples

BPM-2 embraces many of the concepts of BPM-1, but attempts to remove the language and pre-specified platform dependencies. These are the major changes:

• Because languages and platforms change, the BPM tool must be completely agnostic to these concerns.
• BPM-1 systems were based on programming models that were designed for 3-tier solutions; BPM-2 erases the 'logical tier' concept and replaces it with black-boxed services.
• The digital process must be described in a standard, platform neutral manner. It should be exportable from your solution.
• BPM-1 largely attempts to bypass the existing enterprise architecture; it comes bundled with its own. BPM-2 honors the existing architecture and encourages a best-of-breed architectural solution for a given process.
• Corporations are adopting a new ubiquitous fashion for integrating software: Web Services. This means that:
  • BPM-2 must integrate to other systems and processes using services at the edge of the application.
  • BPM-2 functionality must be woven at the core via web services.
  • BPM-2 must facilitate inter-business processes as a first-order concern.

BPM-2 is designed to be a digital framework describing your business and systems. Think of it as a nerve center that sends signals to all of the organs of your system (the services). An organization will have distinct services for presentation, authorization, access control, notification, integration, directory lookup, persistence, data transformation, and likely scores more. Rather than buying all of these from a single vendor, pre-baked into a BPM platform, you will find them from specialists and weave them together.

Moving ahead

There are three reasons why BPM-1 was not well accepted:

1. It was a solution designed to span the enterprise, yet its architecture failed to recognize the existing enterprise infrastructure.
2. It grew up in one of the worst periods to sell commercial software.
3. The all-encompassing nature of BPM-1 made it a large application, requiring a hefty price tag and long release cycles.

BPM-2 will face its own challenges:

1. It is largely based on the acceptance of a process execution language. I am a fan of BPEL4WS, but I believe that it is limited and will either require additional enhancements or it will give way to a standard that does.
2. BPM-2 is based on web services. And although the software community has come to agreement on the WS basic profile (XML Schema, UDDI, WSDL & SOAP), much more is needed.
3. Most software specialists, including business analysts, are not aware of how to drive software creation from a process perspective. They believe that software is driven from the Use-Case, not the Business-Case. This will require a change in methodology.

Lastly, don't count out the BPM-1 vendors. Many of them understand this new wave and are still cash rich. It may be challenging for them to retool their products but I believe that most of them realize the writing is on the wall if they don't make changes. They also realize that the largest threat will likely come from IBM and Microsoft, who are pushing web service standards and backing process languages like BPEL.

BPM-2 will prevail because it is the culmination of decades of work involving rapid design, model driven development, component modularity, distributed computing, analytics and process improvement. It is the next generation, with a rich inheritance. It will create a disruption in the packaged applications space and create openings for new entrants. But in the end, corporate customers that adopt such solutions will likely be the true benefactors.

More on this topic