6:26 p.m.
Where would I find information about the rationale of having multiple named
WM entry points?
Thanks
-W
6:55 p.m.
Wolfgang Laun wrote:
Where would I find information about the rationale of having multiple
named WM entry points?
Only the release notes. It's related mostly to CEP so
that you can have
dedicated partitions for an input stream, to avoid bottle necks, they
are also multi-threaded, if you turn on multi-threading mode.
In general you'll have one entry point to one named input stream.
Mark
Thanks
-W
------------------------------------------------------------------------
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
7:23 p.m.
I'm not sure about the rationale of the Drools devs, but I can give an example of that
design pattern from another rules package.
At my workplace (Southwest Airlines) we use Tibco BusinesEvents, and BE has the concept of
Channels with Destinations that produce Events. A Channel can be mapped to various
message producers (we mainly use JMS topics and queues) and our internal event types are
mapped to the message types inside the Destination.
This design pattern can be useful because often data with the same format can come from
different sources, and the application may want to react differently mattering on the
source. Also, there may be reasons you can't or don't want to include that source
information in the data itself.
For example, my current work deals with the airline's flight schedule, and reacting to
daily changes in that schedule. We get RouteChange messages from various sources.
(Sometimes due to a flight cancellation, and other times due to a flight addition.) We
react differently to a RouteChange mattering on the reason, but the reason isn't
encapsulated in the message, it's in the source. Thus it's nice to have multiple
WM entry points, as otherwise we'd need yet another translation layer to inject a
source marker into our RouteChangeEvent.
GreG
--- On Fri, 3/27/09, Wolfgang Laun <wolfgang.laun(a)gmail.com> wrote:
From: Wolfgang Laun <wolfgang.laun(a)gmail.com>
Subject: [rules-users] org.drools.runtime.rule.WorkingMemoryEntryPoint
To: "Rules Users List" <rules-users(a)lists.jboss.org>
Date: Friday, March 27, 2009, 12:26 PM
Where would I find information about the rationale of having
multiple named
WM entry points?
Thanks
-W
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
9:03 p.m.
On Fri, Mar 27, 2009 at 7:23 PM, Greg Barton <greg_barton(a)yahoo.com> wrote:
I'm not sure about the rationale of the Drools devs, but I can give an
example of that design pattern from another rules package.
At my workplace (Southwest Airlines) we use Tibco BusinesEvents, and BE has
the concept of Channels with Destinations that produce Events. A Channel
can be mapped to various message producers (we mainly use JMS topics and
queues) and our internal event types are mapped to the message types inside
the Destination.
This design pattern can be useful because often data with the same format
can come from different sources, and the application may want to react
differently mattering on the source. Also, there may be reasons you can't
or don't want to include that source information in the data itself.
For example, my current work deals with the airline's flight schedule, and
reacting to daily changes in that schedule. We get RouteChange messages
from various sources. (Sometimes due to a flight cancellation, and other
times due to a flight addition.) We react differently to a RouteChange
mattering on the reason, but the reason isn't encapsulated in the message,
it's in the source. Thus it's nice to have multiple WM entry points, as
otherwise we'd need yet another translation layer to inject a source marker
into our RouteChangeEvent.
This makes sense, and I was thinking that the Entry Point might leave its
tracks in some org-drools.event.rule.*Event, bu tI couldn't find any
reference to WorkingMemoryEntryPoint in any of these. So, is there a way of
knowing the WMEP thorugh wich a fact has arrived?
-W
GreG
--- On Fri, 3/27/09, Wolfgang Laun <wolfgang.laun(a)gmail.com> wrote:
> From: Wolfgang Laun <wolfgang.laun(a)gmail.com>
> Subject: [rules-users] org.drools.runtime.rule.WorkingMemoryEntryPoint
> To: "Rules Users List" <rules-users(a)lists.jboss.org>
> Date: Friday, March 27, 2009, 12:26 PM
> Where would I find information about the rationale of having
> multiple named
> WM entry points?
>
> Thanks
> -W
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
3:50 p.m.
Hi Wolfgang,
The answers so far were all correct. Just to add a bit more, try this:
https://hudson.jboss.org/hudson/job/drools/lastSuccessfulBuild/artifact/t...
Feedback most welcome.
Also, going on into the internals, as Mark mentioned, entry points are
an abstract way of partitioning (or better saying: scoping) parts of your
alpha network. This abstract concept has a few applications in very specific
fields, but the first and more common use is to support streams of events,
as described in the above link.
So, is there a way of knowing the WMEP thorugh wich a fact has
arrived?
The FactHandle implementation has this information, yes, but it is not
exposed through the interface. I don't think it would be good/safe to expose
it to users, but if you have a compelling use case we may discuss a way to
do that without compromising the consistency/safeness of the engine.
[]s
Edson
2009/3/27 Wolfgang Laun <wolfgang.laun(a)gmail.com>
On Fri, Mar 27, 2009 at 7:23 PM, Greg Barton
<greg_barton(a)yahoo.com>wrote:
>
> I'm not sure about the rationale of the Drools devs, but I can give an
> example of that design pattern from another rules package.
>
> At my workplace (Southwest Airlines) we use Tibco BusinesEvents, and BE
> has the concept of Channels with Destinations that produce Events. A
> Channel can be mapped to various message producers (we mainly use JMS topics
> and queues) and our internal event types are mapped to the message types
> inside the Destination.
>
> This design pattern can be useful because often data with the same format
> can come from different sources, and the application may want to react
> differently mattering on the source. Also, there may be reasons you can't
> or don't want to include that source information in the data itself.
>
> For example, my current work deals with the airline's flight schedule, and
> reacting to daily changes in that schedule. We get RouteChange messages
> from various sources. (Sometimes due to a flight cancellation, and other
> times due to a flight addition.) We react differently to a RouteChange
> mattering on the reason, but the reason isn't encapsulated in the message,
> it's in the source. Thus it's nice to have multiple WM entry points, as
> otherwise we'd need yet another translation layer to inject a source marker
> into our RouteChangeEvent.
>
This makes sense, and I was thinking that the Entry Point might leave its
tracks in some org-drools.event.rule.*Event, bu tI couldn't find any
reference to WorkingMemoryEntryPoint in any of these. So, is there a way of
knowing the WMEP thorugh wich a fact has arrived?
-W
>
> GreG
>
> --- On Fri, 3/27/09, Wolfgang Laun <wolfgang.laun(a)gmail.com> wrote:
>
> > From: Wolfgang Laun <wolfgang.laun(a)gmail.com>
> > Subject: [rules-users] org.drools.runtime.rule.WorkingMemoryEntryPoint
> > To: "Rules Users List" <rules-users(a)lists.jboss.org>
> > Date: Friday, March 27, 2009, 12:26 PM
> > Where would I find information about the rationale of having
> > multiple named
> > WM entry points?
> >
> > Thanks
> > -W
> > _______________________________________________
> > rules-users mailing list
> > rules-users(a)lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/rules-users
>
>
>
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
Edson Tirelli
JBoss Drools Core Development
JBoss, a division of Red Hat @ www.jboss.com
5500
days inactive
5503
days old
4 comments
4 participants
participants (4)
-
Edson Tirelli -
Greg Barton -
Mark Proctor -
Wolfgang Laun