IBM Will Acquire AptSoft

I was wondering when IBM would actually jump into the event processing market.  

Well, it was announced today that IBM will acquire Aptsoft, adding an event processing platform to the IBM WebSphere porfolio.  IBM will also gain AptSoft’s event processing reference customers.  This was a smart move by IBM.

Oracle is acquiring BEA which uses Esper under the hood, another stream processing engine. 

However, AptSoft has a more advanced user interface and graphical design-time environment when compared to the Oracle/BEA/Esper platform,; so the IBM/WebSphere/AptSoft offering will propel IBM to the top of the event processing market.

Now, it’s TIBCO’s turn to acquire an event stream (time series) processing platform to compliment their process-driven event processing offering.

12 Responses to IBM Will Acquire AptSoft

  1. Opher Etzion says:

    Hello Tim. I have never understood the difference between CEP and ESP in your terminology, but it seems to me that AptSoft will be CEP even according to your definition – I think that it does support event clouds… About Apama’s classification – I’ll let them react on that (maybe they prefer to take the CEP Blog award and go for ESP -).



  2. Tim Bass says:

    Hi Opher,

    My apologies for the confusion. I agree with you about AptSoft and have reworded the blog post.

    Thank you.

    Yours faithfully, Tim

    PS: Congrats to IBM on the very clever acquisition.

  3. Francis Marion says:

    Mr. Tim Bass,
    I am new to the CEP blogosphoere and just read your post and previous things you wrote. You seem very knowledgable about technology and the market, and I am confused and hope you can expound on your comments in the same way mr . ophjer etzion was confused. Is this same Aptsoft at that lists customers firms including TotalMove, New Plan Excel Realty Trust, and Vent-Axia? I look at the firms and applications and this contradicts things people have written about event procesing, but maybe I have misundertood. Isn’t Aptsoft a rules engine or BPM software, not CEP or ESP? Also, is there a difference between BPM and CEP/ESP?

    Vent-Axia makes heating and air conditioning equipment. I did some google searches and according to press release, Aptsoft replaced “ineffective paper-based system used to manage customer contact and executive reporting. ” With AptSoft, a provider of outsourced services for sales effectiveness automates the orchestration of sales activity between Salesnet’s on- demand CRM, and Vent-Axia’s in-house ERP system. Does anyone belive automating a manual lead entry and reporting process equals “CEP”? How “complex” could the event processing and pattern-detection around orders for HVAC equipment possibly be?

    TotalMove is a consumer real estate and moving services compny. According to google search press release, they have 500K visitors/month to website, and sales leads are routed to 750 brokers. If 10% of visitors turn into leads that means 50000 visitors/month or 1700/day and ~3 per minute in a 10-hour day. What is the big deal that calls for eSP or CEP? Business have been routing website leads to thousands of reps since mid-1990s without app servers even, let along CEP/ESP? Is it not simple rules engine application?

    New Plan Realty Trust owns and operates 470 community and neighborhood shopping centers, like the kind that have CVS and Kinkos I think. Press release says they have a whopping 12 steps in their paperwork process for signing new lease or renewing a lease, and Aptsoft helps streamline this, becasue legal omments and revisions from a number of sources can happen at any point in each step of the process. WE all buy condos or homes, and merchants rent storefronts for past 100 years without cEP. What kind of event pattern detection could there possibly be? Why isn’t this just BPM?

    Maybe I do not understand terminology. Looking at customers and applications, it does not sound to me like IBM is now suddenly the market leader in cEP/ESP–they are big fish that happens to acquire small fish with dozen random accounts doing different kinds of rules processing and BPM. I would appreciate any clarification you make about why this is considered CEP or ESP by anyone, and difference between BPM and what Aptsoft does? Also, do you know if IBM was shopping for a CEP/ESP vendor and paid lots of $$, or was Aptsoft not doing well and was up for sale?

    Thank you for your explanation to help me better understand what Aptsoft did/does and taxonomy of terms.


  4. Tim Bass says:

    Hi Francis,

    Yes, I agree with you that process-oriented rules-engines are more process-management / orchestration integration tools than true CEP engines.

    In fact, some may argue that they (AptSoft) are not doing CEP at all, because they are not doing anything “truly complex”. I happen to agree with your logic, for the most part.

    AptSoft’s closest competitor in this event-driven orchestration space is TIBCO. Neither TIBCO nor AptSoft, as I recall the AptSoft product, are event stream processing (ESP) engines, where continuous queries are performed across a sliding time window of streaming data.

    Opher does not like seem to like to draw this distinction. He considers event-based process orchestration and also time-series calculations as “CEP” (I think!) and he seems to consider the “real complex” event processing (what you and I would refer to as CEP) as “Intelligent Event Processing”.

    I do not necessary agree with Opher, technically. However, the market is calling these various products “CEP engines” and Opher tends to follow the marketing use of the terms, whereas I tend to focus on what I consider the notion of “complex” and I agree with you that that AptSoft is doing is not really CEP.

    Furthermore, If we raise the bar high enough to actually define “true CEP” there will be no real “CEP players” and Dr. Luckham will not be pleased with that 🙂

    For this reason, among others, I’m heading back into consulting where I advise end users, not software companies. My time working closely with software companies was like a trip to the Land of Oz . Truth seems to be one of the first casualities in the quest to sell software along the Yellow Brick Road.

    On the other hand, since AptSoft has more event processing customers than most of their competitors, IBM jumps to the top when they acquire both the technology and the customers.

    I hope this was useful. Also, search the blog for more ESP and CEP related posts.

    Yours faithfully, Tim

  5. Tim Bass says:

    Hi (Again) Frank,

    FYI, Sandy Carter, IBM’s vice president of SOA and WebSphere strategies, says something I completely agree with:

    “In the marketplace today, everybody talks about complex event processing,” Carter said. “We actually are trying to rename that category, because we believe the real value is in business event processing, with a focus on the business.”


    Yours faithfully, Tim

    PS: Opher, I hope you will find Sandy Carter’s thought leadership with regard to not (wrongly) calling business event processing, CEP.


  6. Maureen Fleming says:

    Since I began looking at this market, I have been amazed at how much discussion — some might call it bickering — there is about the definition of CEP, ESP, etc. There is little effort among competitors to discuss how their technology is different, looking at it from an application perspective.

    AptSoft is very interesting in usage scenarios where the subject matter expert is on the business side and that expert does not want to cede control to a programmer. It works really well in (1) long-lived scenarios such as chronic treatment monitoring in health care, (2) continuous measurement of SLAs, (3) where there is a need to do some orchestration based on what gets triggered when the condition is matched and (4) will be very good in fraud detection and other types of exception detection. It is not used nor optimized for stream processing — most of which have use cases similar to math calculators on steroids.

    Under any stretch of definition, AptSoft Director is not a BPM solution but you can definitely see the value in combining this with process automation tools.

    Apama and AgentLogic are two othervendor with products that can be used by trained business-side subject matter experts.

    Pretty much everything else I see is based on the classic model of the business side sending requirements to IT, and IT implementing and handling change cycles. That doesn’t make these products bad at all, just points to different cultures that will use the products.

    So, then the question is what is the difference in the types of applications that can executed using stream processing versus TIBCO BE vs the business-side products? In addition, when changes need to be made, which of the products have the best technology for rapidly and accurately making the change.

    That would be a more interesting discussion that what to name this technology.

    Would love to hear what people are seeing on this issue.

    Tim, I wish the best with your end user consulting. The market needs people like you in that role.


  7. Tim Bass says:

    Hi Maureen,

    The problem, as I see it, is that the CEP/EP “bickering” is directly related to marketing positioning, as software companies seek to be seen as a leader in an “emerging technology;” however and in fact, event processing is not an “emerging” technology.

    The art-and-science of combining distributed events to identify opportunity and threats in real-time did not begin with the publication of David Luckham’s excellent book. “The Power of Events”.

    As I like to point out in my days of CEP public speaking and customer engagements, every time an airplane flies, a missile is launched, a major weather pattern changes, a major news story hits the market, and so forth, event processing technologies are used.

    Also, every time a neural network is used to identify rogue network activity and fraud, or a Bayesian classifer is used to filter and route RSS news, event processing technology is used.

    Simple Network Management Protocol (SNMP) traps are definately “events” and processing these network management related events, looking for threats, opportunities and root cause analysis in real-time, has been ongoing for years.

    The domain of event processing is very broad, and the CEP/EP community represents a small fraction of the overall domain – and they seem, at times, content to keep it that way. The reason, I think, is because they are looking for analyst coverage which identifies them as leaders or visionaries in an emerging technology, but the fact is that they are just the opposite in most domain specific areas.

    One area for improvement, as I see it, as mentioned to Opher in this blog reply:

    … is that the community is overshadowed by CEP/EP market positioning as “an emerging technology” but in fact, event processing has a very rich history and myriad use cases not in the warp-and-wolf of the software companies, and the analysts covering these companies.

    Yours faithfully, Tim

  8. Maureen Fleming says:

    Hi Tim,
    There is a big difference between an emerging technology and an emerging market. You are correct in pointing out that event processing has been around for a long time and really is pretty common. And I don’t think anyone would argue that many of the things that are trying to be solved with CEP are currently solved with custom code and point solutions.

    At the same time, I don’t think anyone could argue that CEP as commercial off-the-shelf software is anything but emerging and the market immature.

    It won’t be until the technology has a richer product flavor to it that we will start seeing really interesting things happen with all the variants of event processing technology because it will become much more affordable.

    To your point on software companies seeking leadership, I don’t think any vendor wants to be a leader of an emerging market — they want to be the leader of a mature market. So, all of the scappling is an effort to gain market share as this market goes through its cycle. This sort of thing is always a bit of a soap opera.


  9. Tim Bass says:

    Hi Maureen,

    Excellent points. You are right on target with your views and perspectives.

    Also, to elaborate, there is only minimal value to consumers in COTS CEP/ESP/EP engines because where the real expenses, and expertise, exist will be in developing and maintained the Key Indicators (patterns, rules, analytics, domain knowledge, intellectual property) that makes the COTS software actually perform a useful task for the user; and at the same time, making it so these KIs are reusable both within the organization across various platfoms and LOBs. The same is true, of course, when we talk about sharing KIs across organizations.

    To your point, reusable COTS engines have only limited appeal, because the real value is in the reusable KIs (the knowledge). This is a long standing problems with expert systems, of course; and the same problem is already negative impacting customers who are tuned into the modern mantra of “reuse” but do not get “reuse” with current COT CEP/EP software.

    Almost every end-user (expert) I talk to in Thailand is keen to point out that the value proposition of CEP/EP is diluted without reusable KIs across various CEP/EP platforms.

    The CEP/EP market will have a hard time “emerging” without standards according to just about every end user I talk to. There is very little value proposition in developing code (KIs) for proprietary CEP/EP platforms that can’t be reused.

    Yours sincerely, Tim

  10. PatternStorm says:

    Hi Tim, you ‘ve said: “there is little value proposition in developing code (KIs) for proprietary CEP/EP platforms that can’t be reused”, and I absolutely agree, this is why a standard Event Processing Meta-language, i.e. a high level language (and associated semantics) to express situations and reactions to these situations in a platform-independent way that can be automatically translated to different platform-specific EP languages, a concept that Opher has been pushing for a while now, is more than an academic nicety but something with real business value. Against this idea it has been argued many times that it is too early to look for standards, that no one size fits all, that there is too much variation to be captured by a single meta-language, etc…These arguments are all very well, however, as you pointed out, the real value for the business is not in the COTS EP engines themselves, but in the knowledge that’s coded in their EP languages: without an standard platform-independent notation (and associated semantics) to express situations and reactions to these situations we are locking in this value into often, obscure, proprietary EP languages. I can perefectly understand that customers complain about this, I would also 😉

    So IMHO the EP community and vendors would be wise to join efforts to develop a standard platform-independent EP meta-language and associated semantics. I know of two (separate?) initiatives pursuing this, the EPTS’s one, led by Opher, and the OMG’s one. Any more? Any papers, also?

  11. Tim Bass says:

    Hi Claudi,

    I have not seen any tangible work on this in the EPTS, and agree with you that this should be one of the highest priorities, I also recall that Opher also is keen on this; and he has mentioned this topic a number of times, and the community has pushed back, just as you well described.

    Most customers are not interested in proprietary engines where they cannot share the KIs across platforms. This is a very real Achilles’ Heel in the CEP/EP market.

    This Achilles’ Heel need to be addressed sooner than later, as you mention.

    Yours sincerely, Tim

  12. Maureen Fleming says:

    There is an interesting book written in the 80s called “Danger in the Comfort Zone,” which describes how people fail to progress because their thinking stays where it is most comfortable.

    This discussion about a standard CEP language is comfort zone thinking. If you look at growth in the middleware market, it is driven by (1) the use of modeling tools and other types of abstractions that remove the need to code and (2) high performance, non-standard or quasi-standard middleware. Growth is not in code.

    If you look at other markets, growth and traction comes from the best approach and not a collective group of vendors determining before the market even develops that they know best about what should be the market’s standards. Think VMWare and not BPEL.

    It will be interesting to monitor the progress of standards discussions, but my guess is that the products that will gain traction will either (1) have a monster services organization behind them and will be deployed as embedded technology or (2) the products offer the best user experience and domain expertise.

    In scenario 1, vendors should be focused on performance In the second, I’d put my money on the modeling- and graphically-oriented products, particularly those aligned with BPM and business. In this case, what types of standards would be most useful?


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: