Cloud Event Processing vs Event Cloud Processing?
Posted by Paul Vincent
Interesting to see the huge interest in Cloud Computing. David Luckham’s complexevents.com just referenced one of several recent Infoworld articles and blogs that attempt to define the term. Another compares the buzz to past excitements - remember Application Service Provider and Software As A Service? But Clouds are of direct interest to the Complex Event Processing community: CEP is, afterall, about processing clouds of events.
1. Complex Event Processing is another term for Event Cloud Processing: processing clouds of events, which may include multiple event streams of course.
2. Cloud-based Event Processing is where some event processing application is put into a closed / corporate, or public / distributed, cloud computing environment. Event latency and event store latency will be major considerations here.
It’s interesting to compare “cloud computing” with more conventional advances in distributed corporate IT. TIBCO BusinessEvents, for example, provides an “event cloud processing” (aka CEP!) platform that includes both the event processing and data cache agents. Will this ever be requested, or offered, as a cloud service by TIBCO or a TIBCO partner? Possibly. Application domains for this might be the interface between the social network and commercial IT worlds - consider things like medical drug programs, healthcare data collection, and so forth. We’ll let you know if and when it happens…
Agent-based, Complex Event Processing (CEP), Trends | 20 January 2009 |
7 Comments
Other Links to this Post
RSS feed for comments on this post. TrackBack URI
By Hans, January 20, 2009 @ 06:11
FWIW the terms “cloud of events” and “cloud computing” have some subtleties.
A cloud of events is really a bunch of streams, maybe multiplexed and maybe distributed. You can’t use an event cloud, you always use one or more specific streams that are known a priori. And there are architecture considerations for every single stream that’s added to the cloud.
Contrast that to cloud computing where the idea is to have many identical computers, where you do not need to know a priori which one you will use. And adding one computer is the same as adding another (in theory).
And then the event cloud might be delivered by a computing cloud where you do not know in advance which computer will deliver events to a subscriber, but each subscriber knows which events they are looking for.
By Tim Bass, January 20, 2009 @ 20:17
Hi Paul,
Han’s continues to make the mistake of stating that a clould of events (or event cloud) is simply “a bunch of streams”. This is technically incorrect and an oversimplification of complex systems.
Yours sincerely, Tim
By vincent, January 22, 2009 @ 04:19
Hi Hans - the point I was trying to lead to was that event cloud processing in a cloud computing environment probably is OK for high-latency applications…
Hi Tim - good catch on terminology… per EPTS…
Event cloud: a partially ordered set of events (poset), either bounded or unbounded, where the partial orderings are imposed by the causal, timing and other relationships between the events.
Notes:
Typically an event cloud is created by the events produced by one or more distributed systems.
An event cloud may contain many event types, event streams and event channels.
The difference between a cloud and a stream is that there is no event relationship that totally orders the events in a cloud. A stream is a cloud, but the converse is not necessarily true.
CEP usually refers to event processing that assumes an event cloud as input, and thereby can make no assumptions about the arrival order of events.
By Hans, January 22, 2009 @ 09:02
> the point I was trying to lead to was that event cloud processing in a cloud computing environment probably is OK for high-latency applications
Oh that most certainly depends on many factors. Certainly cloud computing has its place in processing events, but I would not be so quick to classify it as low or high latency.
By Hans, January 22, 2009 @ 09:54
On the terminology question, it is a mathematical fact that every element POSET is composed of chains (which some call TOSETs). If a cloud is a POSET and a stream is a TOSET then every cloud is composed of streams. Ask a mathematician.
But that fact is not useful in any way and the EPTS should have long ago removed POSET and TOSET from their definitions.
By vincent, January 25, 2009 @ 16:05
Hi Hans - probably one for the discussion on the EPTS glossary at complexevents.com or ep-ts.com … I wonder if a better definition for an “event stream” is as a particular type (i.e. time-ordered) of TOSET; the “event cloud” may or may not be made up of such streams. I guess its somewhat of an academic discussion, although the need to handle unordered vs ordered events is certainly a valid CEP technology characteristic.
Cheers
By Hans, January 26, 2009 @ 06:36
I posted to the forum a few days ago.
Any set can be associated with many algebraic relations. You’re never going to find one that is “the” overall relation.
Who cares if you get events in a time order? Computers talk in sequences of bytes - so every set of events will come with an implicit ordering. Again, who cares? What matters is how the software uses the events; the use case imposes the set relations.
My advice: There are already terms for POSET (”POSET”) and TOSET (”TOSET”, “chain”). Let anyone who wants to use those terms, use them. Let stream and cloud revert to the IT based definitions that make sense.