In exploring the role of events is it possible to achieve predictive analysis to provide rapid response and take proactive action?
One possibility is by tracking how humans handle event exceptions and locking their processing logic and turning this into business logic. This allows one to perform event correlations and to automate exception handling. Here event handling can take the form of rapid response or proactive action. Further, analysis of precursor events (i.e. events that occurred just prior to the exception) could lead to predictive alerts to be raised to circumvent exception situations and thus enable proactive actions to be taken.
If sensors and RF ID technology are the first steps to event capturing and event processing then addition of event analysis and event composition (Complex Event Processing style) is the next step in the evolution with exception based learning and proactive action based event emission may be considered a more advanced step in the process of EDA.
Many transportation companies and carriers and just in time supply chain providers could adopt EDA for rapid response or even proactive action. For example, combining weather based events, traffic flow patterns etc can be used to insure quality of the goods being transported to minimize wastage in transport. Furthermore, containers that transport organic food that does not use preservatives could use special types of "sensors" that detect the emission of gases and chemicals within the shipping container chambers to assess the freshness and the ripeness of the produce. If these events indicate rapid ripening proactive action based events can be sent to these shipping containers to lower temperature etc. to retain the freshness of the produce for transportation with minimum damage. (This example is only illustrative as I am not a expert on this subject.)
It must be noted that traffic and weather based events are combined with information about product preservation rules, correlated and processed to preserve sensitive consumer products to safely and preserve the high quality after this type of behavior has been observed in the human actor and this exception processing logic has been codified for future automation. EDA in this case is utilized for the purpose of tracking human exception processing and then automating this behavior albeit all the while depending on incoming current state events and outgoing proactive actionable events.
It seems very much a plausible use of EDA and so I am curious how many of you are using EDA for solving similar use cases.
As always your input is very valuable.
surekha -
One possibility is by tracking how humans handle event exceptions and locking their processing logic and turning this into business logic. This allows one to perform event correlations and to automate exception handling. Here event handling can take the form of rapid response or proactive action. Further, analysis of precursor events (i.e. events that occurred just prior to the exception) could lead to predictive alerts to be raised to circumvent exception situations and thus enable proactive actions to be taken.
If sensors and RF ID technology are the first steps to event capturing and event processing then addition of event analysis and event composition (Complex Event Processing style) is the next step in the evolution with exception based learning and proactive action based event emission may be considered a more advanced step in the process of EDA.
Many transportation companies and carriers and just in time supply chain providers could adopt EDA for rapid response or even proactive action. For example, combining weather based events, traffic flow patterns etc can be used to insure quality of the goods being transported to minimize wastage in transport. Furthermore, containers that transport organic food that does not use preservatives could use special types of "sensors" that detect the emission of gases and chemicals within the shipping container chambers to assess the freshness and the ripeness of the produce. If these events indicate rapid ripening proactive action based events can be sent to these shipping containers to lower temperature etc. to retain the freshness of the produce for transportation with minimum damage. (This example is only illustrative as I am not a expert on this subject.)
It must be noted that traffic and weather based events are combined with information about product preservation rules, correlated and processed to preserve sensitive consumer products to safely and preserve the high quality after this type of behavior has been observed in the human actor and this exception processing logic has been codified for future automation. EDA in this case is utilized for the purpose of tracking human exception processing and then automating this behavior albeit all the while depending on incoming current state events and outgoing proactive actionable events.
It seems very much a plausible use of EDA and so I am curious how many of you are using EDA for solving similar use cases.
As always your input is very valuable.
surekha -
2 comments:
In the financial sector EDA is used largely for fraud detection though at SunTrust we are also exploring revenue generating opportunities as well which involves a better predictive analysis concerning cross channel sales and services. We are just exploring this from a strategic perspective. As I’m sure you are all too aware of what is typically the case in large organizations where many different LOB’s spin up their own little projects to handle specific cases. The key in our case will hinge on the maturity of our SOA initiative. This only makes sense as we believe SOA and EDA do compliment each other well. We are in the midst of an enterprise SOA implementation involving an ESB so our EDA strategy will take a course of leveraging the SOA infrastructure and methodology extending it to handle events.
If you think about it what’s the difference between a message and an event. Not much other than the eventing information such as event id, type, timestamp, who created it, when it expires, occurrence, correlation, etc. This stuff ends up in the message header which is what distinguishes it as an event. So if our Event Generators are required to send their events to the bus just like our consumers send SOA messages then we have a consistent architecture, and naturally the bus can process the events and leverage CEP engines in a pluggable way to handle the complex processing that provide the real value in EDA.
So this is some of the things we are doing and thinking about concerning EDA.
Tom
Tom - I am glad to see that you are looking to not only leveraging "complex business events" but are looking to tie this in with your SOA infrastructure as in the ESB for all routing and transformation.
In fact, I had a blog on this very topic on how these two architecture paradigms fit. It takes a slightly different perspective but the fact remains that an enterprise does have to look at both forms/ styles of architecture to fit their needs.
http://entarch.blogspot.com/2007/10/key-learnings-using-eda-to-implement.html
Thanks a ton for your post.
surekha -
Post a Comment