Complex Event Processing
Viva La BAM!
By Ross Altman, CTO for SOA and Business Integration, SUN Microsystems
Business Activity Monitoring (BAM), Business Events and the Event Driven Architecture (EDA) are all becoming hot topics in the trade press and in industry analyst publications. But, the marketplace for this technology is still very soft. Very few of the stand-alone BAM vendors have grown to any significant size and the BAM offerings from the multi-product integration middleware vendors are being used by a relatively small percentage of their customers.
But just look what you get from BAM Ė real time or nearly real time visibility into the operations of the enterprise.
- Want to know the call centerís average ďtime on holdĒ RIGHT NOW? BAM can do that Ė just look at the ďspeedometerĒ on the dashboard.
- Need to know the current average time required to acknowledge orders submitted through your B2B gateway? Again, itís on the dashboard.
- Is it useful to be able to alert operations managers when the average time to pick orders in the warehouse goes past four hours? Just set a threshold in your BAM product and it will send an SMS message to the operations managerís cell phone.
But, with all of the good stuff thatís available with BAM, itís no big surprise that some folks are a little hazy on just what you can do with BAMÖand what you canít (or shouldnít) do.
BAM captures data from ongoing business operations in real time and then uses that data for two purposes. Most visibly, BAM products produce a dashboard-like user interface where gauges and meters are used to show the current rate for order picking in a warehouse or time-on-hold in a call center. To make these graphical representations even easier to read at a glance, the results may be color-coded. For example, a graphical stoplight might shine red, yellow or green, depending on the data thatís been aggregated by the BAM event processing engine.
While the dashboard is the more visible aspect of a BAM product, itís another BAM capability that really generates value for users. Letís say that youíve used your BPM product to automate car loan processing at a bank. Itís a complex process, but youíve automated every step, so you promise all of your customers that youíll provide approvals within ten seconds.
Implementation of a system like this is quite straightforward. With a time stamp applied to each message when it arrives at the bank and another time stamp applied when the loan approval is sent to the dealer, you can determine the response time for each loan application.
Finally, your BAM software has an alert engine that can continuously monitor this aggregated data. This means that whenever the average response time for the past five minutes exceeds five seconds, you can have the alerting mechanism automatically send a message to tell the operations manager to correct the problem.
Itís this, the ability to aggregate data in real time, which sets BAM apart from both data warehousing and BPM monitoring. Data warehousing involves too much of a time lag and itís too difficult to aggregate data in BPM monitoring.
So, given this specialized tool, what do you use it for?
The key here is recognizing that not all of your metrics are best tracked in real time. For example, you could try to calculate earning per share on a minute-to-minute basis. But, apart from that calculation being far to complex to undertake more frequently than once a quarter, what would be the point? We already properly castigate modern corporate managers for focusing too much on short-term financial outcomes to the detriment of longer-term business development. Why would we possibly want to collect earnings information more frequently than we do today?
The bottom line is that using BAM to track metrics that canít be used to trigger real time corrective actions is useless. There are certain business activity measurements that cry out for real time or nearly real time monitoring Ė those measurements that can be impacted by comparably real-time remediation. For those metrics that can only be impacted by initiatives that required a lengthy gestation period, business intelligence solutions based on data warehousing technology are more than adequate.
About the Author
Ross Altman, CTO for SOA and Business Integration at Sun Microsystems, leads the direction, development, and communication of Sun's business integration technology vision and strategy. Ross joined Sun through Sun's acquisition of SeeBeyond, the application integration and SOA development
tools vendor. Prior positions include director of integrated technologies at EDS and vice president and research director at Gartner, where his research focused on the use of integration middleware to support composite applications, straight-through processing and business process management. Ross holds bachelor's and master's degrees from Rutgers University. More by Ross Altman
About SUN Microsystems
Since its inception in 1982, a singular vision -- "The Network Is The Computer" -- has propelled Sun Microsystems, Inc. (Nasdaq: SUNW) to its position as a leading provider of industrial-strength hardware, software and services that make the Net work. Sun can be found in more than 100 countries and on the World Wide Web at http://sun.com