Since my recent post, a bit more has dribbled out into the blogosphere about the negotiations over BPMN 2.0, most of it completely off track. But now SAP’s David Frankel, definitely an insider, is shining a welcome light in those dark spaces with his BPMN 2.0 Update.
The biggest difference between the two submissions is in how they define the BPMN 2.0 metamodel. The BPMN-S submission positions the OMG’s Business Process Definition Metamodel (BPDM) as the metamodel for BPMN 2.0. The BEA-IBM-Oracle-SAP submission defines a dedicated BPMN 2.0 metamodel, and proposes a mapping between the dedicated metamodel and BPDM.
Here the BEA-IBM-Oracle-SAP submission is what my post called Approach 1 or the BPMS approach, and BPMN-S is what my post called Approach 2, or BPDM. David gets right to the key point:
The BPMN-S submission uses BPDM as the metamodel, and uses BPDM’s mapping to BPMN notation.
The BEA-IBM-Oracle-SAP submission takes the position that BPDM is not a metamodel of BPMN; rather, it says, BPDM is a metamodel of a new, abstract language for process that, as envisioned by the BPDM RFP, was intended to be mapped to multiple concrete languages. BEA-IBM-Oracle-SAP approach is that BPMN, as one of those concrete languages, requires its own metamodel, whose constructs are clearly recognizable as BPMN elements.
But I think he is being polite here. Since custom behaviors can be expressed in BPDM’s new abstract language, BPMN-S takes the position that the BPMN should be user-definable, referencing my “wish list” post on BPMS Watch for non-aborting attached events. BPDM can express this behavior, and that’s a good thing, says BPMN-S. But I don’t think that’s a good thing, if you can’t understand the semantics from looking at the diagram, and I believe that is also the philosophy of the IBM-SAP team as well. BPMN-S seems to throw down the glove with statements like this:
BPMN2 provides capabilities… needed for effective functioning of organizations and productive interaction with industry partners. These capabilities cannot be supported by typical language modeling techniques that simply capture pictures in XML with a computation-dependent semantics.
Ouch. But David goes on to say that despite the trash talk, merger negotiations between the two groups have been ongoing since March, in which the goal is a unified submission.
May 6th, 2008
If you want to jump-start your BPMN efforts, I’ll be offering a half-day pre-conference workshop on Process Modeling with BPMN at the upcoming Gartner Application Architecture, Development & Integration Summit in Orlando. This Gartner event is the leading independent SOA and application infrastructure conference, and the agenda’s 6 tracks and 70+ sessions cover future trends and latest best practices in application development, application integration, SOA, Web Services and Web2.0, as well as SaaS/Cloud Computing. My workshop is on the afternoon of June 8; the regular conference runs June 9-11.
The BPMN workshop provides an in-depth tutorial on what has become the key BPM standard, used for both analytical modeling and model-driven implementation in BPM Suites ranging from Oracle, SAP, BEA, and SoftwareAG to Lombardi, TIBCO, Savvion, Appian, and Adobe. You’ll learn not only the semantics behind the notation, but patterns and best practices for modeling events and exceptions, flow control, and organizing complex end-to-end models.
There is an additional $495 fee for participating in the workshop, but BPMS Watch has secured a saving of $400 off the standard registration fee for the conference. For complete conference details and to view the agenda visit www.gartner.com/us/aadi-spring. To register and claim your discount call 866-405-2511 and mention priority code: ADISA.
Beyond the workshop, the event is a good opportunity to understand the future of SOA and the web, new application and architecture models, and discuss your specific issues through Gartner 1-1s, roundtables, facilitated working sessions, and online community. Guest keynotes include Nick Carr, Leading Tech and Business Writer on “The Big Switch,” and Andrew Lippman, MIT’s Media Lab Futurist on “IT Architecture futures.”
I’ll be available to meet with you one-on-one as well to discuss how BPMN can help bring business and IT together in your organization.
Hope to see you in Orlando.
May 6th, 2008
Surprisingly little information has reached public view concerning BPMN 2.0, now under consideration in OMG. Unlike most standards approval processes, the outcome of this one is not preordained. There are two submissions, quite different, and it could go either way.
Oracle’s Vishal Saxena notes that one reason BPMN 1.x has been so successful is that it “keeps simple things simple” by focusing on abstract business-level modeling, allowing developers flexibility in how to implement the technical details, and argues that BPMN 2.0 “should maintain this flexibility.” In response, IDS Scheer’s Sebastian Stein points out that a problem with BPMN 1.x is that it “only has implicitly defined execution semantics,” and BPMN 2.0 needs to make them explicit. He goes on to neatly summarize the competing proposals:
At the current point there are two different approaches discussed within OMG, how to do that:
- BPMN 1.1 already contains an implicit definition of the execution semantics, so just make them explicit by writing them down.
- Use a more abstract meta model with well defined execution semantics like BPDM and map BPMN elements to it.
Approach 1 is a simple solution, because in the end it does not change much for the BPMN user. The implicit execution semantics are just made explicit, but they are not modified. Approach 2 is more complicated, because a second language is added to the stack and new concepts are introduced. Having more and new concepts might contradict BPMN’s current strength - simplicity. This will increase the learning curve for BPMN modelling and might hinder a further adoption.
I think this is a fair characterization, but it’s not the whole story. Approach 1 is a joint submission by the large platform/middleware vendors, companies used to having their way in W3C and OASIS. Approach 2 is BPDM, which has been incubating in OMG for a long time, and wedded to OMG fundamentals like MOF and xmi. More significantly, Approach 1 reflects (to me, at least) a BPMS approach to BPM solution development, emphasizing unification of the business-oriented process model with the executable design, while Approach 2 is more aligned to OMG’s programmer-oriented MDA vision, in which models are just a more efficient way to generate code.
It’s a subtle distinction, and not everyone would agree with my characterization. A key element is that in Approach 1, the execution semantics of shapes in the diagram are specified by the standard, while in Approach 2 they can be redefined by the user. That’s what the BPMN Wish List controversy back in March was all about.
As a non-voting guest member of OMG, I am not a party to the contest, but I am not neutral, either. I think a win by Approach 1 would greatly assist BPMS become mainstream and better define the relationship of BPM to SOA. Formal action by OMG on this is not scheduled until August-September.
May 1st, 2008
Regarding TIBCO’s first-ever “analyst summit” at their annual user conference, I’ll leave it to Sandy to record the actual content of the presentations to analysts. I’ll stick to the impressionistic view. Apparently “the analysts” had told TIBCO they wanted to hear executives talk about go-to-market strategy, so we got almost nothing about product and an awful lot about “value propositions.” Are there really analysts who want to spend half a day hearing about value props and selling tactics? Scary. But, having lowered my expectations completely, TIBCO’s “solution showcase” exhibits - open to the hoi polloi after the analyst event ended - actually blew my socks off:
- Spotfire, a business-empowered performance visualization technology, was dazzling. Not BPM-specific, but a perfect fit for BAM. You take a set of instance data, and interactively scatter-plot any metrics against each other, zoom in on a region of interest, filter instances with a slider in real time, and list the instances in the set. All real time, interactive, no code. Totally cool. If they would just put that inside iProcess BAM and get rid of the IDS Scheer analytics, it would make total sense. No they did not announce that.
- iProcess Conductor. Supposedly this has been around for several years, but I missed it completely in the BPMS Reports, and TIBCO never noticed. I call it case management, but it works as well with fully automated straight-through processes, so TIBCO calls it “dynamic BPM.” The idea is a container that holds multiple processes added ad-hoc at runtime either by human decision or event-triggered rules. Each included process has its own expected duration, SLA duration, and actual duration, and processes have dependency relationships with other processes in the case, er… container. The software calculates the Gantt chart for the overall process and can identify critical paths, SLA violations, expected completion dates, all updated as the case progresses. I have had clients looking for this for years, and never knew TIBCO had it. Very cool.
- Business Studio 3.0. This is TIBCO’s Eclipse-based free full-BPMN modeler/designer, now enhanced with an Ajax forms designer and direct deployment to the runtime environment (without going through the old iProcess Modeler tool). The free download will not include the forms piece, and you still need to buy the runtime to do unit testing. It should all be free, in my view, as the competition (Oracle BPA/SOA Suite) lets you download a unit test environment along with the tool, while TIBCO gives you a faster path from model to implementation. You would still need to write a large check to deploy into production. They say they “have reasons” for doing it this way now (but hold open the possibility of making it all freely available in the future).
Bottom line, TIBCO has the piece-parts for a world-class BPM Suite, but BPMS market dynamics appear to run counter to TIBCO’s overall marketing strategy. BPMS favors integrated suites, while TIBCO’s strategy features component “neutrality”, meaning each piece of the stack is sold individually on its own merits, not as an integrated platform that works together as a unit. The BPMS market wants to hear about business empowered implementation, not just IT agility, but TIBCO remains wedded to IT-centered value propositions. TIBCO has features that deliver on core BPM values like cost reduction, cycle time improvement, and quality improvement, but their value prop messaging has moved on to next-generation exotica like predictive intelligence. So from a marketing perspective, BPM appears to be the poor stepchild at TIBCO, but leveraging some unique techology, is succeeding in sales in spite of it.
April 29th, 2008
Alex Toussaint offers a peek at AquaLogic BPM’s improved BPMN support in the upcoming v6.1. The palette will include standard BPMN icons such asXOR, OR, and AND gateways, timer and message intermediate events, and I’m glad to see it. Now I just wish they could figure out a way to drag them out of the palette and into the actual process diagram.
**Note added 4/29 - Jesper suggests this as a better example of BPMN “bling.”
April 28th, 2008
At Impact three weeks ago I just got the drive-by version, but now that I’ve gotten the full analyst deep dive, I have to say that IBM now really does seem to have its act together on BPM. The current v6.1 offering has a lot of the improvement built in already, and the July v6.1.2 has more. We’ll be adding IBM to the BPMS Report series and Ratings in Q3, and they should do fairly well.
One of the strong points is something rarely linked to IBM: user experience. The new offering features a unified business user workspace for viewing tasks, instances, performance metrics, and alerts. It features a large palette of prebuilt Web 2.0 widgets that let business users customize their own workspace and mash up the data. The monitoring components seem especially nice.
Another outstanding feature is the repository for modeling artifacts based on Rational Asset Manager, now federated with WebSphere Registry and Repository, which holds executable artifacts. How many presentations have you seen that talk about business users finding business services in a repository and binding them to their process models? Hundreds, I’m sure. And how many BPM/SOA suites provide the federated repositories, business-oriented metadata and query tools that you would need to do it? Probably none. But IBM now does, and lets business users search for related modeling assets - processes, services, policies, KPIs - and reuse them in their own processes.
One of my long-standing complaints about IBM’s approach has been the jarring discontinuity between modeling and process design, and the lack of round-tripping. They have clearly put a lot of work into this aspect of BPM, and while WebSphere Modeler and WebSphere Integration Developer are still very different tools, they share many artifacts - schemas, forms, KPIs - and IBM provides an elaborate mechanism for round-tripping based on propagating changes between the two environments. The software does not automatically make the changes required, but lists the changes required to keep the model and WID design in sync. It’s a lot better than before, and I suspect will be simpler once BPMN 2.0 comes out.
My other long-time complaint about IBM’s BPM approach as been the whole WebSphere-FileNet dichotomy. While pure content lifecycle processes and document workflow processes are “special cases” in BPM, in reality any BPM process may involve content attachments that should be organized, managed, and securely retained in a true content repository. You shouldn’t have to use a special content-centric process engine to secure process attachments. And now, says IBM, you don’t. While I would still go further, IBM now can tell a single “story” that brings process and content together. Essentially, the steps that involve content operations (e.g. adding/revising/securing documents) are done by a FileNet P8 process, which is invoked via WSDL as a subprocess in the end-to-end WebSphere process. Both the FileNet and WebSphere processes are modeled in WebSphere Business Modeler. Not truly seamless, but the seams don’t show too much. IBM hasn’t told this story very well, possibly for fear of freaking out the FileNet base, but I actually think this gives that base a more comfortable (and plausible) way forward.
Look for the IBM BPMS Report around July.
April 28th, 2008
I’ve just finished the BPMS Watch Ratings, a comparative scoring of the 11 leading BPM Suites written up in my BPMS Report series on BPMInstitute.org. Those reports, which are available for free, include Appian, BEA, Cordys, EMC, FlowCentric, Global 360, Lombardi, Oracle, Singularity, SoftwareAG/webMethods, and TIBCO. I would have liked to get Pega and Savvion - they declined. IBM (WebSphere and FileNet), Adobe, Intalio, and Fujitsu have expressed interest but I needed to get the ratings out; maybe we’ll get them in a supplementary round.
**Note added 4/22: All of the reports are now on the BPMInstitute site.**
The BPMS Watch ratings evaluate the offerings separately for human-centric and integration-centric BPM, and further breaks down human-centric into production workflow and case management. The ratings measure all the suites in 11 categories of capabilities, which are weighted differently based on the process type. The ratings are just for the product itself — nothing for “completeness of vision” or other nonsense like that. We plot the aggregate rating for integration-centric vs human-centric, integration-centric vs production workflow, and production workflow vs case management. The idea is that the market is moving beyond point solutions, and you might want a good compromise between multiple process types. Whatever…
This report gives the weightings and the results, but I still want to put out a more detailed report that gets into more detail behind the scoring and lets users tweak the weightings themselves.
If you know anything about how these Wave/MQ things normally work, the last few weeks are usually an intense negotiation between the vendors - who get to review the draft - and the analysts, with lots of screaming and threats. This one is different. The vendors haven’t seen this yet. So by the time they start screaming and threatening, the cat will be out of the bag.
**Note added April 22: Not too much screaming and threats, but vendors asked me to clarify that the reports on which the ratings are based were written over a 9-month interval from July 2007 to March 2008, and some offerings may have enhanced their capabilities since their report was written. The revised Ratings Report (see sidebar) now lists the date of each product report.**
The report is here on BPMS Watch in the sidebar. You need to register to download it. But if you’re like most people you won’t even read it; you just want to know who won. OK here it is.
April 10th, 2008
It seems my post yesterday, drawn from the press release, keynote slides, and mini-briefing, missed the coded messages in IBM’s BPMS announcement. Here is the decoded version.
The announcement of an “IBM BPM Suite” represents a big deal internally at IBM. It is intended to signify a commitment to a single BPMS based on interworking components from separate divisions - WebSphere, FileNet, Lotus, Rational, GBS, etc. It required signoff from all the various warlords - Rosamilia, Goyal, LeBlanc, Bowden, etc. Apparently a big deal. They know they’re not there yet, but the commitment to get there is new.
Second point is investment protection. The starter sets - one for WebSphere and the other for FileNet - are meant to signal that customer investments in those technologies will be protected going forward as the BPMS takes shape. IBM (unlike FileNet) is well known for this. We’ll see how well it works. I sense they are willing to take their hits in the Magic Quadrants rather than rattle the installed base.
Third point is interworking of components across traditional brand boundaries. It always seems to be harder to do that at IBM than at other places. They obsess over details like unifying the install, test, and support. But again the commitment to integrate across brands in the BPMS is apparently significant.
The June-July v6.1.2 release now seems less like a major step forward and more like the first step on a long journey.
April 9th, 2008
You’re probably saying, wait a minute, didn’t they already have one? Yes, I admit, they were in the 2006 BPMS Report series, in which they agreed (reluctantly, I hear) to let the combination of WebSphere Modeler, Monitor, WID, and Process Server be described as a BPM Suite. But today at Impact here in Las Vegas they actually announced it as an orderable thing. Sort of…
One of the problems for BPMS at Big Blue has always been that the required components cut across IBM brands, which appear to resist integration with each other. So WebSphere had its BPM story, FileNet (Information On Demand) had its own, and I suppose so did Lotus and Rational, as well. Last year IBM marketing decreed WebSphere and FileNet would be presented as a single BPM story, but they never succeeded in getting an analyst to tell it… including me. Because there wasn’t a single story to tell. (That’s what has delayed IBM’s inclusion in the current BPMS Report series, and I met another BPM analyst here who told me his own IBM BPMS report is currently stuck in the mud on the same issue.)
So what is the new offering announced today? It’s called the “IBM BPM Suite” (look ma, no branding!) and it includes both WebSphere and FileNet (with some Rational and Lotus, as well). Does that mean they’ve finally integrated the components? Not really. The new suite “includes a choice of two foundational ‘Starter Sets’ that make it easier for customers to get started with BPM.” One of those starter sets is WebSphere Modeler, Monitor, and Fabric (which includes parts of WID and Process Server). The other is FileNet. Other suite components, including the model repository (Rational Asset Manager), some Lotus collaboration stuff, a new Business Event Processing tool based on another recent acquisition, are labeled “extended value offerings.”
Clearly IBM is interpreting the word “suite” to mean a portfolio rather than an integrated platform. So I doubt the new marketing spin by itself is going to do anything. But the mini-briefing I got on this suggests there might actually be something new here, with Fabric playing a more central role. One of my biggest complaints about the WebSphere BPM story has been the jarring discontinuity between Modeler and WID - different process metamodels, different data models, different programming models, no roundtripping. The new story is that Fabric, formerly known as Webify, is a business-friendlier front end that enables “policy-based service composition” and hides the developer-centric guts of WID and Process Server. I am going to need an actual briefing on this before it makes sense, but here is the marketecture diagram from the keynote:
IBM also throws a small bone to BPMN, saying Modeler will support a new “BPMN visualization format.” Somehow that is not quite saying Modeler will support BPMN. I suspect a bit of BPMN hand-waving like BEA does with AquaLogic BPM today. But IBM is a major player shaping BPMN 2.0, and there is no doubt that standard will play a key role when it emerges in 2009 (or later). The WebSphere piece of the new BPMS is a tweak of v6.1. It should be available later this quarter.
***Correction added April 9: I got a demo of Fabric today and, to correct the post, it does not replace WID and Process Server. It adds a new Service Composition perspective to WID, and it will be linked in to Modeler as well. Fabric allows service endpoints to be selected dynamically at runtime based on policies, and allows business users to tweak policy parameters at runtime (similar to rule maintenance applications in the business rules arena). You still need to go into WID to edit the BPEL, SCA assembly diagrams, and other things that make IBM’s BPM development so foreign to the business side. In the future I think IBM is shooting for direct deployment from Modeler (including the Fabric Service Composer) without dealing with WID… but I don’t think 6.1.2 is promising that.
April 8th, 2008
Antoine of Intalio comments on yesterday’s post re the need to tweak BPMN diagrams to make them more BPEL-friendly. One of my complaints had to do with the nasty “interleaving” error that always pops up when you turn on BPEL validation. He references an Intalio presentation at EclipseCon last year, specifically slide 16. I copied his BPMN into my tool (ITP Commerce) and sure enough, it gives the interleaving error:
Getting rid of interleaving errors requires making duplicate copies of some of the diagram nodes so that BPEL’s block structure is happy. In my tool this has to be done by hand, but Antoine sort of suggests that Intalio does this automatically by turning the BPMN into a “directed acyclic graph” and generating BPEL from that. More mathematical gobbledygook to me, but I think he is saying “we fix the interleaving thing automatically” by turning BPMN into a “tree”… I assume under certain conditions. Here is the result from his document:
As you can see, ‘merge’ and ‘do more things’ have been duplicated, but apparently it’s automatic. Sweet. I’m curious, Antoine, does Intalio automatically handle my mixed token diagram from yesterday? And, since those exclusive merge gateways are really optional in BPMN - uncontrolled merge means the same thing in this case - does Intalio require you to draw them?
April 5th, 2008
Previous Posts