Hi Shankar, This is still an open research, people are figuring out how best do it.
In the released version, we set the actual time for new event, which is not perfect. General requirement is if events new event n1 is derived from events a,b,c, then n1 should be given a time stamp greater than a,b,c and if n2 is derived with events older than a,b,c .. then n2 should be larger than n1. Have to figure out how to do it. It is bit like in distributed systems. --Srinath On Thu, Apr 30, 2015 at 5:51 AM, Selvaratnam Uthaiyashankar < [email protected]> wrote: > > > On Wed, Apr 8, 2015 at 12:01 PM, Sriskandarajah Suhothayan <[email protected]> > wrote: > >> Yes we have thought about this during the Siddhi rewrite and we have made >> necessary architectural changes to support this. >> Here our approach is to update Siddhi system time based on event arrival >> > > So, the time in events created by Siddi will also based this "system > time"? in that case, lets say we have to create an event now, and last > event received by siddi was 10 s back, new event will be created with old > time? > > > > >> and >> also support event ordering at the input event handler level. >> >> We have not done a full implementation on this for the upcoming release, >> but expecting to add this features for the next release. >> >> Suho >> >> On Wed, Apr 8, 2015 at 11:55 AM, Srinath Perera <[email protected]> wrote: >> >>> Hi Suho, Sachini, >>> >>> Lot of IoT usecases are coming, and we will soon have usecases where we >>> get lot events from devices. >>> >>> Currently we put a timestamp once we received an event. However, that is >>> only a approximation and might be problematic. >>> >>> For IoT usecases, events from same devices monotonically increasing has >>> time stamps but they will not be in sync. I think we need to figure out how >>> to handle this scenario. >>> >>> I think we should >>> >>> 1. We should support external time windows for all usecases ( IMO >>> best approach is to make Siddhi work that way always and deriving current >>> time from the events we have received, and current model we can place on >>> that model easily) >>> 2. When multiple clients send events to the same data stream, we >>> should figure out a way to order events before processing them and >>> support >>> it out of the box. >>> >>> WDYT? obviously I am not talking about release that goes next month. >>> >>> --Srinath >>> >>> -- >>> ============================ >>> Blog: http://srinathsview.blogspot.com twitter:@srinath_perera >>> Site: http://people.apache.org/~hemapani/ >>> Photos: http://www.flickr.com/photos/hemapani/ >>> Phone: 0772360902 >>> >> >> >> >> -- >> >> *S. Suhothayan* >> Technical Lead & Team Lead of WSO2 Complex Event Processor >> *WSO2 Inc. *http://wso2.com >> * <http://wso2.com/>* >> lean . enterprise . middleware >> >> >> *cell: (+94) 779 756 757 <%28%2B94%29%20779%20756%20757> | blog: >> http://suhothayan.blogspot.com/ <http://suhothayan.blogspot.com/>twitter: >> http://twitter.com/suhothayan <http://twitter.com/suhothayan> | linked-in: >> http://lk.linkedin.com/in/suhothayan <http://lk.linkedin.com/in/suhothayan>* >> >> _______________________________________________ >> Architecture mailing list >> [email protected] >> https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture >> >> > > > -- > S.Uthaiyashankar > VP Engineering > WSO2 Inc. > http://wso2.com/ - "lean . enterprise . middleware" > > Phone: +94 714897591 > -- ============================ Blog: http://srinathsview.blogspot.com twitter:@srinath_perera Site: http://people.apache.org/~hemapani/ Photos: http://www.flickr.com/photos/hemapani/ Phone: 0772360902
_______________________________________________ Architecture mailing list [email protected] https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture
