In the early 90’s while consulting at Sprint, I worked with Peter Lothberg, who is, without a doubt, one of the world’s top networking experts. I remember Peter as a super-genius who debugged Cisco router software live on the major Internet routing exchanges back then. Those days were quite controversial (and political), and some of the commercial routing polices of the good ole’ days were the motivation for my 1997 IEEE paper on exterior gateway routing protocols.
Today, I saw this article about Peter’s mother, Swedish woman gets superfast Internet – Peter has connected her to the Internet at 40 Gigabits per second, her first Internet connection! I know my Mom would like to connect to the Internet at this speed if she could filter out all the spam, pop-ups, and other malware that makes life on the net really exhausting at times. In the words of my dear sweet Mom, who connected to the Internet 14 years before Peter’s mother, “The Internet Ain’t No Fun Anymore!”
This is an area where CEP can help, and perhaps, someday, make the Internet a safer, cleaner, and fun place for folks! Bayesian filtering now dominates anti-spam technology; and I am amazed at the accuracy of Google’s GMail spam filters. Bayesian techniques are also very dominate in fraud detection and other CEP-related solutions domains, including Bayesian diagnostic applications.
The state-of-the-art (and promise) of CEP is realized when distributed applications are processed by high speed, low latency networks. Imagine the complex processing we can do when event processing agents (EPAs) are cooperating at 40 gigabits per second!
On the other hand, think of the challenges of trying to filter out all the spam, malware, and other malicious code flying at your Mom, or your children, at 40 Gigabits per second! This is one of the biggest challenges in cyberspace, a virtual world where information is transmitted globally near lightspeed, creating both opportunities and threats; and one of the reasons I remain excited about evolving CEP to solve some of our most challenging event processing problems in the future.