Business Transformation in Action

Joe McKendrick

Transcript: Event Processing Roundtable

user-pic
Vote 0 Votes

The following is a full transcript from ebizQ's recent "2009 Event Processing Roundtable: Managing Events in an Eventful Year," featuring Brenda Michelson, principal with Elemental Links and also contributor here at ebizQ, David Olson, director of CEP product marketing for Progress Software's Apama unit, and myself.

Listen to the 45-minute interactive panel discussion here.

Joe McKendrick: My name is Joe McKendrick, contributing analyst to ebizQ, and I'll be your moderator for today's panel discussion. We have a couple of highly regarded experts and commentators in the event processing space. Let me introduce: Brenda Michelson, principal with Elemental Links and also contributor here at ebizQ, and David Olson, director of CEP product marketing for Progress Software's Apama unit.

We've really got a lot to talk about today, and we're going to try to explore as many avenues of this subject as we can within the next hour.  In recent years, there's been a lot of work around the concept of event processing, and what that means to organizations trying to get a better handle on managing their businesses in a more real-time fashion. A lot of our understanding of event processing comes by way of the work of Dr. David Luckham of Stanford, long considered the father of complex event processing, as well as Dr. K. Mani Chandy of Cal Tech, and Roy Schulte of Gartner, all of whom have participated here at previous ebizQ Webcasts on the topic, and have done a tremendous job of framing the discussion on what CEP is, and the opportunities CEP can provide for today's organizations.

In fact, when asked for an example of a "complex event," Dr. Luckham cited World War I as a classic example of a complex event, with many inputs, many underlying relationships, and many impacts, some of which continue to this day.

Brenda and David, I don't know if we want to attempt to grapple with something on the magnitude of World War I in our analysis of complex event processing, but I know you both have been doing a lot of work in identifying the events and their business relevance on an organizational level. Every organization has a fair share of events that impact it every day, from transactions to purchases to employee changes to all those imaginable and unimaginable external events that could rock the boat, such as the weather or as we've seen recently, economic conditions.

When Roy Schulte was our guest here last year, he compared the ability of business to survive and thrive in any environment as being similar to that of what we observe in the animal kingdom. There's a sense-and-respond capability that enables animals to be aware of opportunities and threats that appear at unpredictable times around them. For example, when a zebra senses the presence of a lion, the zebra responds by quickly running away. Or when a zebra comes across a watering hole or some food, they know to stay, ands that's where they should hang out for a while.

The challenge is to provide our organizations with that same situational awareness. And to change our stripes when necessary, right?

And, as we suggested in the title of this Webcast, there's been a whole stew of unpredictable external factors rocking the boats of companies over the past several months, from the economic and financial tsunami that swept through at the end of 2008. to the repercussions, the budgetary cutbacks, and topsy-turvy movements in the market. Organizations need to understand what has been happening, and what's being affected, and how to Dodge these lions. Plus, and this is where it gets interesting, they also need to see where the new watering holes are - the new food sources -- where the new opportunities exist in what has been a crazy and confusing and sometimes even terrifying economic environment.

And, members of our audience, we want this session to be as interactive as possible, and we welcome your questions and comments....

Okay Brenda, David, let's get started. Just to kind of define and clarify what we mean for our audience. We have event processing, or complex event processing, and event-driven architecture. Let's just kind of go over, to clarify for our audience, what these terms mean, exactly.

Brenda Michelson: I do sit on a committee with Dr. Luckham and Roy Schulte, and we have some very in-depth conversations on the CEP language. But what I really want to drill into is much more, very quickly what CEP and events mean to the average person that would be using event processing.

So very simply I talk about an event, it's a thing that happened. It might be an order was placed. It might be a stock price changed.  It might be the zebra sensing danger.

Event processing is what discerns is if that thing that happened is notable. Is it important to me? Is it important to my business? Do I need to act on it?  And is that notability by itself... or perhaps its notable because of a couple of other events that are going on. Is it just my business that's bad, or is it the entire credit market that's tanked?

And then once you discern that notability, you decide what am I going to do with that event next?  Am I going to forward it along into an event channel? am I going to trigger some kind of downstream action?

Event driven architecture, or EDA, is more the architectural style, which features an asynchronous trigger action type paradigm. The really important thing to understand about EDA is its extremely loosely coupled, so you cant always trace back the origin of the event. That's kind of Dr. Luckham's point on World War I. You don't really know the origins, the contributors. You can't always trace it back, which means you cant always put in compensating transactions. You know what you observed, you know there was this war. So architecturally, you need to think about that.

Joe: David, anything to add?

David Olson: Brenda's got it absolutely right.  I think when people come into the world of 'how can things like complex event processing benefit what I do my business,' is we all come in the IT world especially -- business applications from 30 to 40 years of building applications -- that are reactive to the business.

We capture data, we store data, we analyze it, and we make decisions hopefully that are going to determine what we should be doing in the future. That concept, we focus ourselves internally, and particularly around business transactions. How do we map a traditional paper transaction in the business to something that can be automated?

When you start expanding your knowledge as to all the hundreds of thousands of events that are flowing in and out of the business daily, you broaden your horizons beyond just simply events that are triggered by applications, but events that could be coming from external sources that we still -- from a business decision perspective takes that in as an input -- and makes decision with it..

When we move into CEP, how can that be done through software, through an application that's working with traditional business applications, alongside these event feeds that could be coming in from anywhere, and employ this very proactive situational awareness into the kinds of decision making practices that we have to make in the business?

Joe: I've heard it said that some organizations may have hundreds of thousands of events in a day, maybe millions of events in a day. If you look at everything going on across an organization, but it's essential to be able to filter that out and look at the few events, maybe four or five events that are actually impacting the business, it seems like a huge task.

David: That's sort of one the things that the CEP community in the early stages recognized, that its not sufficient just to be able to capture the events and perhaps store them for future use. But since the volume and the velocity can be quite great in certain circumstances, actually, one of the triggering factors in CEP is that there should be enough volume there that warrants specialized software to be able to deal with it.

But you've got to be able to effectively filter and find those patterns that are meaningful, and Brenda certainly specified this correctly. You got to take those events and from any and all inputs that make sense to the kinds of decisions you need to make, and identify the patterns by which you can take actions, that give you situational awareness...

And we kind of view it on two fronts that need to occur here in the processing of these events. One is visibility. You probably want to see what's going on, make to sure that you are actually going in the right direction with the decisions that you need to make...

But also be able to take those proactive notifications to pass an event along to some kind of thing that's going to take an action to drive some degree of automation. Certainly there's a tremendous volume of events in any particular business. But finding those meaningful patterns is the critical key.

Joe: We have a lot of data available -- more than five to ten years ago. There's not only the databases that have proliferated, but there's RFID, information coming in from a lot of different sources. Is event processing the optimal way to begin to make sense of this data or filter this data?  or is there a BI and analytics component that need to be part of this equation?

Brenda:  Yes...  

Joe: I threw a lot of things into that question.

David: Yes, you did.

Brenda: At the edge, where the events are being generated, and specifically, with RFID, if you think every time the can of soda moves on the pallet in the warehouse, events are generated. So you do need to filter some of those at the edge. And there will be business rules as well applied at the edge. So there is some type of analytic processing in there. Maybe it says I only really care when a high-volume item leaves the warehouse. I want to be notified when the pallet of TVs goes out, make sure I know about it. The pallet of soda I don't care about as much...

As things progress, along the event channels and the processing, you're definitely going to get to your complex event processing by looking for patterns. The business analyst has identified here are the patterns that were looking for, into that engine or that tool. And then downstream too you might be doing further analytics, because often in the event processing portion. What you're really doing is you're trying to make informed actions for your business right now. And analytics and business intelligence is often about making informed decisions for your business tomorrow..

It's all one large flow that works together with analytic components, and event-processing components embedded along the line.

David:  There are capabilities in business intelligence solutions today that provide you some degree of feedback in a business intelligence manner. Here's where we are, here's the current situation that we're into, where CEP starts to work on top of systems like that we mentioned its ability to find those complicated patterns across multiple event streams, and not just from a single source, could be market data feeds...

They can also use the rules that are used to identify patterns to go back at that data you've stored in the past into data warehouses or other databases to help influence actions that you take in the future.

One of the things that people should recognize. Along the decision-making continuum, you've got the past and the future. And traditionally we've used our decision-making tools, particularly in the BI front to predict.  When you talk about situational awareness, you have three aspects of  that continuum. We've got the past, the future and the present. What CEP does is add a significant amount of intelligence in the present, so that the business can act in the moment, and improve decision making in the future...

That's where I think CEP can provide that quantum leap, so business can start making those decisions on what they should be doing, and not what happened. That's where that transition needs to occur.

Joe: Looking back over the past year... the economy has been full of disruptive events for a lot of businesses... let's look at the industry that's been in the middle of the storm...  capital markets, financial services.. ... David I know your company has Done a lot of work in the capital markets area... what kind of events did they miss..  that kind of got us into this mess... if they had CEP technology or methoDologies... in place...  would that have helped these firms better understand what was happening... for example with subprime mortgages.. or derivatives...  can you tell how CEP may have headed off some of these problems that we saw in financial services?

David: Certainly there's an influence of the past.. to help you figure out what kind of rules you should make working in the future...  there are patterns... that people can go back and analyze now... there are patterns...  to help us refine rules..

One of the exciting areas in CEP, especially in capital markets is the whole notion of surveillance...  where there are a fair amount of rules.. that are in play right now, perhaps maybe not in CEP..  but maybe on paper play... that look for abnormal patterns of activity... for example one of our customers uses it to monitor actions that could indicate insider trading, or joint trading opportunities.. where two traders.. trade large volumes of a symbol, but not large enough... that it tips anybody's radar to abnormal activity...

But when these joint trading activities... occur either ahead or behind a news feed... that could indicate to us there's unusual activity going on here...  and we should Do something with this particular symbol that's being traded...

In the past trying to catch those types of activities...could take months of auditing... in order to figure out that something bad happened..

And in the capital markets space, minutes... of something bad happening could be billions of Dollars... and so by applying the capabilities of CEP.. upon all kinds of inputs... coming into the capital markets space..  from market feeds to ...business applications that are processing order transactions.. and things like that...

...these business analysts putting rules on top of these sensitive event feeds... can look for these pretty nasty potentially fraudulent activities under the notion of surveillance..  ... actually were seeing the idea of surveillance sort of bleed out... into other markets too..

..could trigger somebody to say we've got shrinkage issues here... or other revenue loss capabilities... but the capital markers space has served as a wonderful proving ground... for CEP... to people be able to recognize that CEP can handle the volume and the velocity ... and the complicated nature of the kinds of business rules necessary...  to find strange behavior... and perhaps capture fraudulent activity...

So, yes, I think CEP could have been used in some situations that got us into this financial crisis...  perhaps we should have employed CEP a few more years ago... in order to help stem the tide of what has happened... ...but there's certainly a lot of learning that's been going on...  as people go back and research what has happened... ...to figure out what kind of rules.. they should put in place in the future to make it happen again.

Joe: Brenda, would CEP have helped us to avoid some of the worse of the 2008 financial meltdown?

Brenda: I went to school for economics.. but then I switched to computer science... not really sure. But what CEP offers is a broader visibility... so you can have visibility into information from a broad number of sources that perhaps with a traditional monitoring type solution... that you're not going to see as many... ..your view isn't going to be as wide... and you wont have the ability as well to really detect the patterns and find anomalies..

And you're see those results in a moving real-time window... if you thinking about...  ..softgen(?).. you had one bad trader... lose billions and billions of dollars within a short time window...

...CEP for surveillance and risk management.. they would have seen it a lot sooner, and it wouldn't have been such an issue to the bank..

..that was probably September. So we've forgotten about that.. because since then the capital markers have gotten even worse...

...I was reading a really interesting piece in MIT Sloan yesterday...  and it was about kind of detecting  ...weak signals... how Do you make sense of weak signals?  One of the things that these researchers pointed out that having all the data in the world is the first step..  then the next thing you have to Do is you have to open your mind to expect the unexpected.. to test your scenarios that go against your plan......

You need to look at results...  your CEP gets.. then ...you can go back into history and test for another pattern or another outcome... and you really have to test yourself to see what might happen... what might go wrong, and what are the implications. So there's the technology perspective, then the human perspective..  ...analytics...  and being brave enough to expect something bad might happen...

David: We end to use words, scenarios and strategies.. and things like that to help better clarify...  the types of analytics and types of rules you can place on the event patterns you're trying to detect..

Joe: David and Brenda, we've been talking about the financial services sector meltDown.. some interesting thoughts...  on how another disaster could be avoided there... a lot of lessons and absorbed by the financial services sector... I'd love to discuss how other industries can benefit from these lessons..  David, what industries leading the way with CEP...  beyond capital markets?

David: It's a subject near and dear to my heart my primary focus on our efforts is outside of cap markets. We've proven the technology, we've implemented it in a such a way that mortals in IT can certainly think about implementing it as a an overlay technology.

The verticals that were seeing traction and success and what I would call the get-it factor...  where these industries recognize that they have events... coming from lots of places, lots of different angles, lots of different sizes and shapes...  ...and if they could have the ability to correlate these events... then there's much that they can do...

A couple of areas where we've seen interesting usage is in transport and logistics...  these people are managing inventory in motion... .and some of the third party and fourth party logistics companies that are around the world.. are talking about  ...millions of events that are moving through their systems daily... that they're trying to keep track of things... as they go from point A to point B, especially if you look at the events that contribute to appropriate decision making across the supply chain... whether you own it manage it or contribute to it.. it can be a complicated Domain in order to get the visibility you need... in order to make these particular decisions...

One interesting case that we have is a company out of the Netherlands.. called Royal Dirkswager, who manage about 90,000 vessel or ship movements a year. And they're ..capturing and correlating events coming in from.. GPS..  GIS feeds.. other satellite  and telematic feeds off.. ships.. alongside port operations events in order to manage berth space... in loading and outloading of containers in particular... ...and a mode of operation that they had before.. was a ship would sail from some eastern location... and they would go full steam to port.. and they would do this on purpose.. because  ...you could never actually predict what activity would be going on in the harbors...

And if you're late for your berth space, there are some penalties... and some traumatic downstream problems if you're late... but if you're early... you would sit and wait out in the harbor...  ..and so the shipping would do steaming from...  not very green...  ...not an optimal way of handling ... things...

They never really had a cohesive view of their entire network... from ships leaving and arriving.. to movement to berth activity in the ports...   ...what Royal Dirkswager does now... and feed back to the shipping companies what I call rightsteaming.. to ensure that the ships are moving at the right speed...  given the conditions of where they're headed...  weather being an influencing factor... so they're optimizing the flow of goods and services... saving money... saving energy and doing quite well... ..so there are lots of events coming from lots of different sources... to be able to make those proactive decisions...

The other front is retail, RFID especially. We have a company that does item-level RFID tagging...  they Do it not only for inventory management and control, but also product placement in their stores t make sure that their ...products in the right place.. ..because their shelves.. actually contain the readers to make sure they're in the right spot or not in the right spot...

You've got not only innovation...  from an inventory management situation, but also ..customer experience viewpoint...

On a third front, telco... telcos a great domain where lots of events... ..from the devices that you carry from... that are in play in the telco, managing operations from customer service...  ..ordering the entire management... infrastructure...

There's so much going there that... ...the visibility and control that they can get from monitoring and managing the events from their entire Domain..  customer experience is going to be enhanced... billing... convergent billing...  there are lots of..  ways they can use events throughout their entire Domain...  to change the way they Do business...

Joe: Great examples.. rightsteaming that's a pretty interesting term...  I can see there's a lot of money on the line there... with moving these ships...

David: And being green.. is a good thing... it's a great way to facilitate that aspect.

Joe: With telco.. you can see with tens of millions of subscribers... the ability to see what's happening out there... if a subscriber is having trouble Downloading certain features... ...

David: Another aspect that we thought was interesting... telco's depending on what part of the world you're from...  some of them, marketing organizations are continuously offering promotions... service increases, service changes, things like that... in some cases they're initiating these promotions faster than their business systems can react... in some cases they get significant revenue leakage... because they're not actually able to correlate all the events across the business at the time the new service is initiated... will lead to the kinds of checks and balances.. to make sure were not getting revenue leakages...

in other words ..somebody could accept a promotion... and they Don't have credit to Do so... and they're certainly looking at not enabling the service ... to make sure their customer experience is enhanced..

...pretty interesting world when we look at the complicated event streams  that we run across..

Joe: Brenda are you seeing other industries..  ..that are applying the CEP approach... in interesting ways?

Brenda: I am. More recently the people that I speak with... that are reading a paper that I write a while back... are ...people in healthcare... ..so it's really about the medical equipment..  medical equipment generates a lot of events, there are a lot of sensors... ...even implanted in medical devices... that can now give readings remotely... so if you implant a defibrillator, that defibrillator can talk about... the status of itself and the patient and send it to the Doctor...  ..Doctor can send adjustments back to the defibrillator.. previously the patient would have to be opened up every few years to get the batteries... changed and stuff...

...a lot of advances there... I see in insurance... claims processing scenario... ...especially in something like long-term disability.. where it's not a one and Done interaction... it's more how is the patient progressing?  How is their physical therapy going? How many hours are they working any given week at the employer?

...all that information... while its an insurance claims processing... process...  that the inputs into it are all event driven...  ...and then the next action that happens and how much that employee gets paid is based on the events that come in...  so I see a lot of interest there...

...and then of course governments, in trying to be efficient with roads, the smarter cities... managing water.. and really interesting stuff about  ..sensors on bridges... to understand how well a bridge is performing over time, Does it need repair... ...a lot of interesting activities... coming up...

Joe: So you're seeing activity in public sector organizations... as well... are looking into managing complex events...

Brenda: And certainly, if you think of the US government as an organization.. and the federated agencies and the importance of sharing information across those agencies... because its information about a state or some type of event, maybe a health event... .or its all the information about a particular citizen that needs to be shared...  best way of sharing that is through... an event processing type app...

Joe: Example -- the intelligence community... if anyone watches 24 for example, every hour there's an event that needs to be handled...

David:  that's why I use the word surveillance... because I think that helps people think differently about applying technology...

Brenda: It definitely is. There are government agencies... funding the development of CEP engines for specifically threat analysis... and airline terrorism...

Joe: Brenda and David, you talk about some fascinating applications across a range of industries... it sounds like we're beyond perhaps the early stages, we're moving into the next stage of maturity in the market.. is that a fair statement?  beyond pilots.. and test projects... theoretical stage..  and we're in the actionable stage... of the CEP market?

David: From our perspective.. seeing the kinds of customers that are coming to us, who have that 'get-it factor.'  Yes, that is increasing... but what people should recognize is that CEP Does in some cases require a certain volume of events in order for the product to be appropriate for the things it needs to do...

The other things is that ...people need to recognize that there probably are... approaches or strategies that may ...not be conducive to CEP...

So when you're thinking of the volume an velocity of events... that you need to correlate and make sense of..  ..if you're taking hundreds of thousands of events per second.. having them ride through your esb.. and into your event correlator.. is probably not the right approach to take it.

People have to look at ...from a maturity perspective, the technology is mature, it has been tested... and tried.. there are lots of use cases. in other markets we can learn from in order to apply...  .. but the events need to be there in order for this technology to be a fit... for what they need to do..

Joe: We're getting some questions in from our audience....  I want to jump to those...  We have a question about artificial intelligence...  that's a concept that has been around for some time... and perhaps we're getting close to the realization... the audience member wants to know if artificial intelligence techniques... such as pattern recognition and semantic networks ..are being used in a CEP context?  

Brenda: I recently asked an event-processing vendor -- not Progress -- if they saw where the event engine would eventually learn from what it sees to suggest new patterns to then look for, and that individual told me I was living in science fiction.  I said, 'oh that's nice.'

But then in further conversations with people...  ..in the event processing technical society... that's definitely on the horizon..  that would be the next step of things that could go on...  maybe its being used in government or scientific applications.. but I know that's not prevalent in commercial use yet...

David:  If you think about it, the pieces and parts in order to help facilitate that, not the algorithms necessary to do that kind of automated predictive analysis. The infrastructure is there to do it in a number of systems...

We have an event store, that will store events you care to store in time order sequence. So anything that you'd want to do in an event stream, going back and applying an existing strategy or a new strategy you could put human intelligence behind refining your strategies moving forward. Those key components need to be there, in order for any those automated methods. The structure's there. I think people have to work on what are the ways to implement a system.

And then I think we run into a similar thing that AI has had in the past, some skepticism in the past. Do people trust it? There's going to be a level of, a business analyst wants to go, let me look one more time, before I press this okay button..

Brenda: We saw that since the early days of using neural networking in business intelligence tools.

Joe: David talked about establishing the structure for CEP. What does it take to put a CEP system in place?  For most companies out there, and I'm talking about small to medium size businesses as well? What are the necessary components of a CEP system? It seems it could be a fairly resource-intensive undertaking to put such a system in place...  and an audience member says they hear that CEP vendors tend to use adaptors to integrate into event driven architecture, which is difficult and costly and intrusive development approach...  David, do you provide an agent or an intermediary based approach? How Do you get around integration work with adapters?

David: I'll take care of the first one, which is implementation. Obviously this is a server-oriented process. We have to go somewhere to capture and correlate those events, and put whatever rules you want on them  ..in order to turn them into something meaningful..  I needed to be enlightened too... In crossing over to the Apama world... and to wrap my head around semantically... what Does it mean to implement CEP. And I was actually quite surprised that it isn't as complicated as people think it may be.  Yes there's an implementation of a server to capture events...

One of the more complicated endeavors...  that people have to figure out...
Is how are they are going...  ..what events make sense to them, and what types of patterns Do they need to look for?

So the implementation side of things.. really revolves around the scenarios or rules.. you create around those patterns...  we look at it from multiple levels... from a person's capability in implementing this, from a code level....  Hand-writing code... ..put appropriate rules in there...  ..all the way up to drag-and-drop... state-oriented scenario management...

People of a developer caliber, or people of an analyst caliber, can create the kinds of rules necessary for patterns that make sense to you. People want some form of visualization...  of the events that are in flight.... A real-time dashboard to tell them what direction they're going in..  not necessary to help them make the direction, but at least give them the visibility necessary this is what's going on in the business... case in point is..

we have a telco, 3italia.. recently announced their usage of Apama.. in the course of just a couple of months, ...they were able to implement 60 dashboards.. of how they were handling not only the revenue assurance but the customer assurance...

On top of that, I think there were over 200 specific rules... tied to certain KPIs.. so in a short matter of time...  these people were able to implement decision-making power ..on top of their existing systems

Once the light came on, ...you don't tackle this thing like  a 18-month order entry upgrade.. CEP applications can be done in a quick manner. we often help people through that process.. to see how simply people can get involved in it... of course they get the bug, they start thinking of 101 ways to using the product...

so ...these aren't multi, multi mufti year engagements... to implement these things...  ...overlay solution...

That gets me into the whole adapter notion... I Don't know how you can live without adapting... an event stream into a CEP solution....  well get to  ...instrumentation in just a moment..

Without an adaptor... you're assuming there some regulatory... or standardization in event streams... but were not at that point yet... ...there's talk of Doing it... ...there are some standards bodies... talking about the appropriate size shape and color...  

At the moment, if you think about the universe of events... that you could capture in your scenarios...

They could  be processing control... devices or scata systems.  there's a lot of people that have to get together... this is how an event should be formatted in order to accept them without some kind of adapter...

There's got to be some form of transformation...  acceptable to get it into the CEP engine... standardization should help us there...  but like any other methoDologies.. standards aren't there yet...

We've got a fairly rich adaptor framework..  from all the use in capital markets...  from the API level to adaptors...  to various understood data feeds, plugging in becomes a very simple and very rhythmic method... of getting events in to the CEP system...

Follow-on to that is instrumentation.. that's coming.. instrumentation of events from major application servers is a way of life form the Web services world.. especially the management and monitoring... we see a convergence of some of those techniques.. and offering up that data into CEP engines...  it's a methoDology that's coming, and we'll take it on wholeheartedly...

Joe: Another question, the WS-*, Web services standards... ws-eventing for example... can play a role in CEP?  Brenda and David, are standards off in the horizon, or are there current standards that can be employed?

Brenda: I'd say its still off on the horizon. WS-Eventing... it has a competitor in the Web services world..  WS-Notification,, there had started an effort to bring together with WS-EventNotification... but that quietly died at the end of last year... There's definitely a place for standards... ..but it's still a little too early in the evolution in the CEP space.. and some of these standards that we'd like to base these things on... aren't really solidified either... in the case of some of the ws-* standards, which have their own conflicts...

David: I agree. The maturity in the standards front is evolving.... But if the business imperative is there I Don't think it can wait..  ...just architect your solution in such as a way that when a standard Does emerge, you can adapt it, you're not looking at a lot of changes to your physical infrastructure... in order to be able to manage it..

Joe: People talk about three levels to event processing -- the .dashboard level then attach BAM, business activity monitoring to it; then move on to CEP. We have an audience  member asking about the difference between business activity monitoring and complex event processing.  When do you step up from BAM into CEP? What's the continuum here?

David: For me, it's when the whole pattern capability starts to come into it. Where it's not enough to know just what's going on currently. And have some form of analysis of what's going on currently...  But being able to correlate events across multiple streams, and identify patterns across multiple different streams...  Where that actionable side of CEP starts to play a significant role..  if we want the needles and dials to move and say, yep, looks pretty good... ....that's a simplistic answer but that's where we've seen an implementation of BAM... but being able to are the complexity... of what's happening across multiple event streams...  with business rules that make sense to my business, and ...make them actionable.. situational awareness... that also gives me some control... is where we start to see the power of CEP start to play into this...

Joe: Unfortunately we're running out of time, and we need to wrap up. I want to get some final thoughts from Brenda and David. For folks looking into CEP, or starting to investigate solutions or approaches for your business, I'd like to get a couple of nuggets of advice from our panelists on what they should consider.  Brenda, what's your advice for companies starting on the CEP journey?

Brenda: First, you need to understand that event processing, events themselves, are really a new class or tier of enterprise information flow. You really shouldn't restrict your thinking to an application or solution point of view. It's not like your just trying to implement a dashboard or algorithmic trading. You really need to thing about it in the context of information flow.

The other thing I would advice people along the same line is, don't define your events with a third normal form that you would use in a database table. Because an event can flow anywhere, and that's the power of it. But whoever is receiving or observing that event is not going to have access to that original source or even knowledge of that original source. The event as you send it has to be fully described so somebody can derive meaning from it and act on their own.

David: Look at your entire domain from where your organization is to every, if it's a supply chain, where you fit in that supply chain. If there are events coming out of your trading partners, if there are events coming out of your application partners, look at the domain of all those events. And if detecting patterns of interesting activities of all those events can help you improve your decision-making, this is the place to be.

Joe: As David said, it's about surveillance. That's what's happening in our capabilities.

Listen to the 45-minute interactive panel discussion here.

.

Leave a comment

In this blog (formerly known as "SOA in Action"), Joe McKendrick examines how BPM and related business and IT approaches can promote business transformation.

Joe McKendrick

Joe McKendrick is an author and independent analyst who tracks the impact of information technology on management and markets. View more

Subscribe



Subscribe in Bloglines
Subscribe in NewsGator Online
Add ebizQ's SOA in Action Blog to Newsburst from CNET News.com
Add to Google

Recently Commented On

Monthly Archives

Blogs

ADVERTISEMENT