Comments on: IBM’s Conceptual Model for Event Processing Systems http://tibcoblogs.com/cep/2010/02/22/ibms-conceptual-model-for-event-processing-systems/ Complex Event Processing (CEP) Sat, 11 Feb 2012 22:04:06 +0000 http://wordpress.org/?v=2.7.1 hourly 1 By: Paul Vincent http://tibcoblogs.com/cep/2010/02/22/ibms-conceptual-model-for-event-processing-systems/comment-page-1/#comment-1057 Paul Vincent Fri, 02 Apr 2010 15:56:08 +0000 http://tibcoblogs.com/cep/?p=1128#comment-1057 I see IBM has issued this paper now as a <a href="http://www.redbooks.ibm.com/abstracts/redp4642.html?Open" rel="nofollow">RedBook</a>... I see IBM has issued this paper now as a RedBook…

]]> By: Paul Vincent http://tibcoblogs.com/cep/2010/02/22/ibms-conceptual-model-for-event-processing-systems/comment-page-1/#comment-1023 Paul Vincent Tue, 02 Mar 2010 10:51:19 +0000 http://tibcoblogs.com/cep/?p=1128#comment-1023 Hi Anil - good points, but probably these are too detailed for a "Conceptual Model". There are 2 aspects to "guaranteed event delivery": - source event distribution failover support: this is typically a middleware issue, and where TIBCO EMS extensions to the JMS spec come into play - event processing failover support: this is where the event processing component failover features are required. In effect there are 2 sub aspects here: -- event store failover support: this is where tools like TIBCO ActiveSpaces, the TIBCO BusinessEvents cache, and your Gemstone cache provide relevant features -- event processing agent failover support: this is where many of the TIBCO BusinessEvents reliability features hit (things like hot standby agents, etc) Naturally all the above has to be coordinated: the event processing agent failure modes have to be coordinated with the middleware failure modes etc. For example, it could be considered that one of the "special advantages" of TIBCO BusinessEvents is that it can exploit the reliability and resilience features of TIBCO EMS messaging... Cheers Hi Anil - good points, but probably these are too detailed for a “Conceptual Model”.

There are 2 aspects to “guaranteed event delivery”:
- source event distribution failover support: this is typically a middleware issue, and where TIBCO EMS extensions to the JMS spec come into play
- event processing failover support: this is where the event processing component failover features are required. In effect there are 2 sub aspects here:
– event store failover support: this is where tools like TIBCO ActiveSpaces, the TIBCO BusinessEvents cache, and your Gemstone cache provide relevant features
– event processing agent failover support: this is where many of the TIBCO BusinessEvents reliability features hit (things like hot standby agents, etc)

Naturally all the above has to be coordinated: the event processing agent failure modes have to be coordinated with the middleware failure modes etc. For example, it could be considered that one of the “special advantages” of TIBCO BusinessEvents is that it can exploit the reliability and resilience features of TIBCO EMS messaging…

Cheers

]]> By: anilkumar http://tibcoblogs.com/cep/2010/02/22/ibms-conceptual-model-for-event-processing-systems/comment-page-1/#comment-1022 anilkumar Mon, 01 Mar 2010 23:09:21 +0000 http://tibcoblogs.com/cep/?p=1128#comment-1022 We here at gemstone have extensively worked on the CEP system, the challenging part of the CEP mechanism is highly available event delivery system; guaranteeing the event delivery in case of the system failure (failure in the node processing the event). Which are critical in financial sectors that track the event matching high risk criteria (Risk management systems). I would have liked to see more comments from author on this subject. We here at gemstone have extensively worked on the CEP system, the challenging part of the CEP mechanism is highly available event delivery system; guaranteeing the event delivery in case of the system failure (failure in the node processing the event). Which are critical in financial sectors that track the event matching high risk criteria (Risk management systems). I would have liked to see more comments from author on this subject.

]]>