Comments on: Event Models – Continued http://rulecore.com/CEPblog/?p=4 Marco writes about Complex Event Processing Fri, 27 Aug 2010 17:59:16 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Damir Sudarevic http://rulecore.com/CEPblog/?p=4&cpage=1#comment-21 Damir Sudarevic Mon, 15 Dec 2008 20:01:34 +0000 http://rulecore.com/CEPblog/?p=4#comment-21 Few more points to consider: Id ------- A remote sensor in the Arctic measures air temperature and pressure. Every minute, the sensor sends data via a satellite link to a government weather office server. Our CEP application has an input adapter for the weather office data. Who creates the EventId, and how? The Sensor, the weather office server or the CEP application? Entity ------- Joe works for XYZ Ltd. To enter the building Joe swipes his RFID card. The card has a fixed/unique ID number. A card reader at the door sends the event data to the CEP server. To whom or what does the event EntityId belong; Joe, the RFID card or the card reader? Time ------- Keeping in mind that small micro-controller based devices are notorious for not having synchronized clocks--or at least assuming this to be true--how could one obtain reliable TimeStamp data? Few more points to consider:

Id
——-
A remote sensor in the Arctic measures air temperature and pressure. Every minute, the sensor sends data via a satellite link to a government weather office server. Our CEP application has an input adapter for the weather office data.

Who creates the EventId, and how? The Sensor, the weather office server or the CEP application?

Entity
——-
Joe works for XYZ Ltd. To enter the building Joe swipes his RFID card. The card has a fixed/unique ID number. A card reader at the door sends the event data to the CEP server.

To whom or what does the event EntityId belong; Joe, the RFID card or the card reader?

Time
——-
Keeping in mind that small micro-controller based devices are notorious for not having synchronized clocks–or at least assuming this to be true–how could one obtain reliable TimeStamp data?

]]> By: marco http://rulecore.com/CEPblog/?p=4&cpage=1#comment-11 marco Sun, 14 Dec 2008 21:21:33 +0000 http://rulecore.com/CEPblog/?p=4#comment-11 Damir, I think you are on the right track with your idea. For an event model to work it must capture the essence on the event concept. So the same event format should work with a huge number of different systems. In order to do so it is very important that the semantics are well defined. In order to manage to do this we need to keep things as simple as possible. Hans, I think Id must go in there. I'll see if I can build a case for that... As for source, instinct says no... Senders are a technical term and might not be interesting when looking at an event, that needs explanation too... Damir, I think you are on the right track with your idea. For an event model to work it must capture the essence on the event concept. So the same event format should work with a huge number of different systems. In order to do so it is very important that the semantics are well defined. In order to manage to do this we need to keep things as simple as possible.

Hans, I think Id must go in there. I’ll see if I can build a case for that…

As for source, instinct says no… Senders are a technical term and might not be interesting when looking at an event, that needs explanation too…

]]> By: Hans http://rulecore.com/CEPblog/?p=4&cpage=1#comment-9 Hans Sun, 14 Dec 2008 18:41:57 +0000 http://rulecore.com/CEPblog/?p=4#comment-9 Would ID necessarily be be unique? It would be more convenient, but is it needed? Also, what about some kind of field about the sender of the event? Would ID necessarily be be unique? It would be more convenient, but is it needed? Also, what about some kind of field about the sender of the event?

]]> By: Dimitar Dimitrov http://rulecore.com/CEPblog/?p=4&cpage=1#comment-8 Dimitar Dimitrov Sun, 14 Dec 2008 06:52:00 +0000 http://rulecore.com/CEPblog/?p=4#comment-8 Forget my previous question, I should have read this post till the end first... Forget my previous question, I should have read this post till the end first…

]]> By: Dimitar Dimitrov http://rulecore.com/CEPblog/?p=4&cpage=1#comment-7 Dimitar Dimitrov Sun, 14 Dec 2008 06:49:11 +0000 http://rulecore.com/CEPblog/?p=4#comment-7 What happened to the old posts? What happened to the old posts?

]]> By: Damir Sudarevic http://rulecore.com/CEPblog/?p=4&cpage=1#comment-3 Damir Sudarevic Sat, 13 Dec 2008 16:21:10 +0000 http://rulecore.com/CEPblog/?p=4#comment-3 It seems to me that the simplest way to figure this out would be to envision existence of a global event-server/router. Myriad devices (or systems) send their events to the server in a "standard" format and the server delivers events to any system/application that may be interested in particular events. It would be like having a generic interface for any type of device (system). What information would the imaginary event-router need to accomplish this? How would an event object look like? It seems to me that the simplest way to figure this out would be to envision existence of a global event-server/router. Myriad devices (or systems) send their events to the server in a “standard” format and the server delivers events to any system/application that may be interested in particular events. It would be like having a generic interface for any type of device (system). What information would the imaginary event-router need to accomplish this? How would an event object look like?

]]> By: marco http://rulecore.com/CEPblog/?p=4&cpage=1#comment-2 marco Sat, 13 Dec 2008 14:42:56 +0000 http://rulecore.com/CEPblog/?p=4#comment-2 Just testing, does this work? Just testing, does this work?

]]>