On Elephants and Analytics

June 26, 2008

In On EP and Analytics, good friend and respected colleague Opher Etzion applies the well known metaphor of the big elephant to describe how, if you are observing certain specific domains of a subject, like fraud detection, then your view of the whole elephant is biased by your lack of perspective of the entire big elephant.

I am pleased that dear Opher continues to use this metaphor in counterpoint because the same metaphor can be used to describe the carefully selected group of vendors that have banded together to called themselves CEP Vendors.  This group, many founding members of the EPTS, have formed a merry band of well-intended event processing “specialists” and the same lovely elephant causes this group of bonded colleagues to make elephant-blinded statements, as Opher has made in his quoted post:

“Currently most CEP applications do not require analytics.” 

The reason, I believe, that Opher makes the statement above is because the group of software vendors calling themselves “CEP vendors” represent a very small part of the overall event processing elephant;  and hence, since these self-described CEP applications appear to require very little or no analytics, then, by the same logic, CEP requires no analytics. 

(I should outline the boolean logic in a future post!)

For example, one friend and colleague in Thailand is the CTO of True Internet, a leading telecommunications, voice, Video and Internet service provider in Thailand.   True processes myriad events on their network using a dynamic, self-learning neural networking technology.    The US company providing this very clever and highly recommended event processing application does not call themselves a “CEP vendor”; however, they process complex events better and more interesting than the band of merry self-described “CEP players”.

Again,  visualize the gentle giant elephant metaphor that Opher likes to use as a basis for his comments in CEP counterpoint.

When folks define the term “complex event processing” to match a technology marketing campaign that is primarily driven by software running rules against time-series data streaming in a sliding-time windows, and then go on to take the same software capabilities and apply these capabilities to problems that are suitable for that domain, then you match Opher’s elegant description of “a small view of the overall elephant”.

The fact of the matter is that the overall domain of event processing is at least two orders of magnitude larger (maybe more) than the combined annual revenue of the self-described companies marketing what they call “CEP engines.”  The very large “rest of the big elephant” is doing what is also “complex event processing” in everyday operations that are somehow overlooked in “other” analysis and counterplay.

Therefore,  I kindly remain unmoved from my view  that the self-described CEP community, as currently organized, is not immune to counterpoint using the same gentle giant elephant metaphor.  I like this metaphor and hope well-respected colleagues will continue to use this metaphor; because we can easily apply this elegant manner of discussion to explain why the current group of self-described CEP vendors are, in a manner of speaking, selling Capital Market Snake Oil because they are making outrageous claims about the capabilities of their products, as if they can solve the entire “elephant” of event processing problems.   Recently, in this article, CEP was positioned as a technology to mitigate against corporate megadisasters like the subprime meltdown.

Advice:  Tone down the hype.

Furthermore, the noise in the counter arguments marginalize most of the real event processing challenges faced by customers.

In consistant and well respected rebuttal, Opher likes to use the “glass half-full, half-empty” metaphor.   Opher’s point is a valid attempt to paint my operational realism as “half empty” negativism; while at the same time positioning the promotion of the (narrow) event processing capabilities of the self-described CEP rules community as “half-full” thinking. 

For the record, I do see my worldview as “half full” or “half empty”; but an unbiased pragmatic view based on day-to-day interaction with customers with what they would call “complex event processing” problems. 

These same customers would fall over laughing if we tried to bolt one of these rule-based, time-series streaming data processing engines on their network and told them they can detect anything other than trival business events, business opportunities and threats, in near real-time. 

Is it “half empty” thinking to caution people that a “glass” of software that is being touted as the answer to a wide range of complex (even going so far in a recent news article to imply CEP would have magically stopped the subprime crisis!) tangible business problems is not really as that it is hyped to be?  

If so, then I plead guilty to honesty and realism, with the added offense of a sense of fiscal responsibility to customers and end users.

Advertisements

More on CEP Maturity: Capability Versus Reliability

June 3, 2008

Louis Lovas of Progress Apama wrote a complimentary blog entry on the topic at hand, CEP Maturity Models.   In his post, Louis says:

“What a CEP platform has tracks independently of what it is capable of doing. ….. What CEP does, is likely what Tim is referring to when he states we’re in the Technology Trigger phase.”

Peter Lin’s comment, in reply to Louis, concurs:

“Given that COTS CEP has only been around a few years, I think it is safe to say it’s still in the early phase. If we compare it to messaging middleware, which has been around for more than 15 years, CEP isn’t as mature. Another comparison is business rule engines and expert systems. The earliest business rule engines date back to late 80’s. All things considered, I would agree with Tim. COTS CEP still has a lot of time to mature.”

Louis was spot on when he said that I was focused on overall CEP functionality; not individual product reliability.

Independent of how reliable a particular CEP-type application might appear; the overall state-of-the-art of CEP is really quite immature.

 

 


Open Service Event Management

May 17, 2008

One of the benefits of working in different countries is to get the perspectives of various client’s event processing problems.    Of interest to event processing professionals, companies are moving away from expensive software solutions and increasingly moving toward experimenting with economical and open software packages to solve complex problems.   

Recently, I was talking with a client about their experience with commercial security event management (SEM) solutions, for example ArcSight.   In his opinion, ArcSight was not an economically viable solution for his company, so he recommended I take a look at Open Service Event Management (OSEM). 
 
OSEM helps organizations collect, filter, and send problem reports for supported systems (ProLiant and Integrity) running compatible agents.   OSEM automatically sends service event notifications when system problems are detected.

I have not had a chance to look under the hood of OSEM and see how it can be used to collect and send events to emerging rule-based event processing engines.    However, this looks like an interesting lab project and I would like to hear from readers who have experimented with this systems architecture.


A Bitter Pill To Swallow: First Generation CEP Software Needs To Evolve

February 8, 2008

Frankly speaking, the CEP market is now saturated with hype about all the great things CEP can do, detecting opportunities and threats in real time and supporting the decision cycle.  However, in my opinion, it is time for the software vendors and analysts to move beyond the marketing hype and demonstrate real operational value with strong end user success, something seriously lacking today.

I have advocated this evolution for two years, including the notion of expanding CEP capabilities with proven techniques for event processing that have worked well long before current “Not yet CEP but called CEP” software hit the marketplace and airwaves.

For example, in my first CEP/EP presentation in New York in 1Q 2006, I presented Processing Patterns for Predictive Business and talked about how the US military has implemented high performance detection-oriented systems for many years (in the art-and-science of multisensor data fusion, MSDF), and how every day, when we sit at home (or at work or in transit), we are comforted to know we are safe from missile attacks because of what I would also call “complex event processing.”   There is a very rich history of “CEP but not called CEP” behind the scenes keeping people safe and warm. (The same thing can be said with many similar examples of complex event processing in use today, but not called “CEP” by CEP software vendors.)

This is one reason, when I read the “CEP history lessons,” I am amused at how, at times, the lessons appear self-serving, not end user serving.  There is so much rich event processing history and proven architectures in “CEP but not called CEP” (CEP that actually works, in practice everyday, long before it was called CEP).  It continues to puzzle me that a few people the CEP/EP community continue to take the “we invented EP” view.  Quite frankly, the history we read is missing most, if not all, of the history and practice of MSDF.

When we take the current CEP COTS software offerings and apply it to these working “CEP but not called CEP” applications, the folks with real operational “CEP but not called CEP” detection-oriented experience quickly cut through the hype because they are, based on their state-of-the-practice, now seeking self-learning, self-healing “real CEP type” systems.  They are not so excited about first generation technologies full of promises from software vendors with only a few years of experience in solving detection-oriented problems and very few real success stories.

The same is true for advanced fraud detection and other state-of-the-art detection-oriented processing of “complex events” and situations.  The state-of-the-art of complex event processing, in practice, is far beyond the first generation CEP engines on the market today. 

This is one of the reasons I have agreed with the IBM folks who are calling these first generation “CEP orchestration engines” BEP engines, because that view is closer to fact than fiction.  Frankly speaking again, process orchestration is much easier than complex detection with high situation detection confidence and also low false alarms.

Customers who are detection-savvy also know this, and I have blogged about a few of these meetings and customer concerns.  For example, please read my blog entry about a banker who was very sceptical in a recent wealth management conference in Bangkok.  I see this reaction all the time, in practice. 

Complex problems are not new and they still cry out for solutions.  Furthermore, many current-generation event processing solutions are already more advanced that the first generation CEP engines on the “call it CEP” market today.  This is a very real inhibitor, in my opinion, to growth in the “call it CEP” software space today – and credibility may ultimately be “at risk.”  Caution is advised.

Candidly speaking again, there are too many red-herring CEP-related discussions and not enough solid results given the time software vendors have been promoting CEP/EP (again, this is simply my opinion).  The market is in danger of eventually losing credibility, at least in the circles I travel and complex problems I enjoy solving, because the capabilities of the (so called) CEP technologies by software vendors in the (so called) CEP space have been over sold; and, frankly speaking, I have yet to see tangible proof of “real CEP capabilities” in the road maps and plans of the current CEP software vendors.  This is dissappointing.

This pill is bitter and difficult to swallow, but most of my life’s work has been advising, formulating and architecting real-time solutions for the end user (the C-level executives and the operational experts with the complex problems to solve).   CEP software must evolve and there needs to be more tangible results, not more marketing hype.


An Overture to the 2007 CEP Blog Awards

January 9, 2008

Before announcing the winners of the 2007 CEP Blog Awards I thought it would be helpful to introduce the award categories to our readers.

I have given considerable thought to how to structure The CEP Blog Awards. This was not an easy task, as you might imagine, given the confusion in the event processing marketspace. So here goes.

For the 2007 CEP Blog Awards I have created three event processing categories. Here are the categories and a brief description of each one:

The CEP Blog Award for Rule-Based Event Processing

Preface: I was also inclined to call this category “process-based event processing” or “control-based event processing” and might actually do so in the future. As always, your comments and feedback are important and appreciated.

Rule-based (or process-based) event processing is a major subcategory of event processing. Rule-based approaches to event processing are very useful for stateful event-driven process control, track and trace, dynamic resource management and basic pattern detection (see slide 12 of this presentation). Rule-based approaches are optimal for a wide-range of production-related event processing systems.

However, just like any system, there are engineering trade-offs using this approach. Rule-based systems tend not to scale well when the number of rules (facts) are large. Rule-based approaches can also be difficult to manage in a distributed multi-designer environment. Moreover, rule-based approaches are suboptimal for self-learning and tend not to process uncertainty very well. Never the less, rule-based event processing is a very important CEP category.

The CEP Blog Award for Event Stream Processing

Stream-centric approaches to event processing are also a very important overall category of event processing. Unlike a stateful, process-driven rule-based approach, event stream processing optimizes high performance continuous queries over sliding time windows. High performance, low latency event processing is one of the main design goals for many stream processing engines.

Continuous queries over event streams are genenerally designed to be executed in milliseconds, seconds and perhaps a bit longer time intervals. Process-driven event processing, on the other hand, can manage processes, resources, states and patterns over long time intervals, for example, hours and days, not just milliseconds and seconds.

Therefore, event stream processing tends to be optimized for a different set of problems than process-based (which I am calling rule-based this year) event processing. Similar to rule or process-based approaches, most current stream processing engines do not manage or deal with probability, likelihood and uncertainty very well (if at all).

The CEP Blog Award for Advanced Event Processing

For a lack of a better term, I call this category advanced event processing. Advanced event processing will more-than-likely have a rule-based and/or a stream-based event processing component. However, to be categorized as advanced event processing software the software platform must also be able to perform more advanced event processing that can deal with probability, fuzzy logic and/or uncertainty. Event processing software in this category should also have the capability to automatically learn, or be trained, similar to artificial neural networks (ANNs).

Some of my good colleagues might prefer to call this category AI-capable event processing (or intelligent event processing), but I prefer to call this award category advanced event processing for the 2007 awards. If you like the term intelligent event processing, let’s talk about this in 2008!

Ideally, advanced event processing software should have plug-in modules that permit the event processing architect, or systems programmer, to select and configure one or more different analytical methods at design-time. The results from one method should be available to other methods, for example the output of a stream processing module might be the input to a neural network (NN) or Bayesian Belief (BN) module. In another example pipeline operation, the output of a Bayesian classifier could be the input to a process or rule-based event processing module within the same run-time environment.

For all three categories for 2007, there should be a graphical user interface for design-time construction and modeling. There should also be a robust run-time environment and most, if not all, of the other “goodies” that we expect from event processing platforms.

Most importantly, there should be reference customers for the software and the company. The CEP Blog Awards will be only given to companies with a proven and public customer base.

In my next post on this topic, I’ll name the Awardees for 2007. Thank you for standing by. If you have any questions or comments, please contact me directly.


Motor Vehicle Crashes and Complex Event Processing

December 30, 2007

The Research and Innovative Technology Administration (RITA) coordinates Department of Transportation’s (DOT) research programs.  RITA’s mission is to advance the deployment of multi-disciplinary technologies to improve transportation system in the U.S.

Shaw-Pin Miaou, Joon Jin Song and Bani K. Mallick wrote a detailed paper, Roadway Traffic Crash Mapping: A Space-Time Modeling Approach, in RITA’s Journal of Transportation and Statistics.    In their paper, the authors state that, “motor vehicle crashes are complex events involving the interactions of five major factors: drivers, traffic, roads, vehicles, and the environment.”

Maiou, Song and Mallick go on to say that “studies have shown that risk estimation using hierarchical Bayes models has several advantages over estimation using classical methods.”    They also point out that “the overall strength of the Bayesian approach is its ability to structure complicated models, inferential goals, and analyses. Among the hierarchical Bayes methods, three are most popular in disease mapping studies: empirical Bayes (EB), linear Bayes (LB), and full Bayes methods.”

Maiou, Song and Mallick directly reference two important problems that David Luckham recently mentioned during his keynote presentation at the 2007 Gartner Event Processing Symposium, traffic congestion management and global epidemic warning systems.  In addition, Jean Bacon, professor of distributed systems in Cambridge University’s computer laboratory, was recently mentioned in the article Fusing data to manage traffic.

As Maiou, Song and Mallick point out, motor vehicle crashes are complex events requiring the correlation of five situational objects (drivers, traffic, roads, vehicles, and the environment).  Each one of these five situational objects may also be a complex event.  The representation of each object requires complex event processing.

We often see these discussions and articles across the wire, for example, “Is BAM (Business Activity Monitoring) Dead?”  or “Is it CEP or Operational BI (Business Intelligence)?” or “Is it Event-Driven SOA or Just Plain Old SOA?”  Frankly speaking, these debates and discussions are red-herrings.

What is important to solving real problems, indicated by the complex event processing paper by Maiou, Song and Mallick, are real solutions not buzzwords and three letter acronyms.  Please keep this in mind when using the term “complex event processing.”  


End Users Should Define the CEP Market.

December 17, 2007

My friend Opher mistakenly thought I was thinking of him when I related the story of the fish, as he replied, CEP and the Story of the Captured Traveller.

I must not have related the fish story very well, because to understood the story of the fish, is to know that we are all like the fish, in certain aspects of life, and there is nothing negative to be gleaned from the story.

However, to Opher’s point on CEP, I disagree.   Just because the marketing people (not the market) has misdefined CEP and therefore the vendors are drifting from the technology described in Dr. Luckham’s original CEP work, including his CEP book, we should not change the context of CEP.    Therefore, I don’t agree we should redefine CEP, as David envisioned, as Intelligent Event Processing (IEP) because CEP, as today’s software vendors sell it, is really SEP (or whatever!)  Please recall that David’s background at Stanford was AI and he did not define CEP as the software vendors have defined it either!

The fact of the matter is that the software marketing folks have decided they are going to use Dr. Luckham’s book to sell software that does not perform as Dr. Luckham described or envisioned!   I make no apologies for being on the side of end users who actually need to solve complex problems, not sell software that underperforms.

As I mentioned, this positioning and repositioning does not help solve complex problems.   At the end of the day, we have problems to solve and the software community is not very helpful when they place form over substance, consistently. 

Furthermore, as most customers are saying, time and time again, “so what?” … “these COTS event processing platforms with simple joins, selects and rules do not solve my complex event processing problems.”  “We already have similar approaches, where we have spent millions of dollars, and they do not work well.”

In other words, the market is crying out for true COTS CEP solutions, but the software community is not yet delivering.  OBTW, this is nothing new.  In my first briefing to the EP community in January of 2006, I mentioned that CEP required stating the business problem, or domain problem, and then selecting the method or methods that best solve the problem or problems.

To date, the CEP community has not done this because they have no COTS tool set other than SEP engines (marketed as either ESP engines or CEP engines – and at least ESP was closer to being technically accurate.) 

Experienced end users are very intelligent. 

These end users know the complex event processing problems they need to solve; and they know the limitations of the current COTS approaches marketed by the CEP community.  Even in Thailand, a country many of you might mistakenly think is not very advanced technologically, there are experts in telecommunications (who run large networks) who are working on very difficult fraud detection applications, and they use neural networks and say the results are very good.   However, there is not one CEP vendor, that I know of, who offers true CEP capability in the form of neural nets.  

Almost every major bank, telco, etc. has the same opinion, and the same problem. They need much more capability than streaming joins, selects and rules to solve their complex event processing problems that Dr. Luckham outlined in his book.   The software vendors are attempting to define the CEP market to match their capability; unfortunately, their capabilities do not meet the requirements of the vast majority of end users who have CEP problems to solve.

If the current CEP platforms were truely solving complex event processing problems, annual sales would be orders of magnitudes higher.  Hence, the users have already voted.   The problem is that the CEP community is not listening.