synapse-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Fremantle <pzf...@gmail.com>
Subject Re: Synapse and WSRM
Date Tue, 27 Dec 2005 19:21:24 GMT
Geoffrey

I agree. The JMS was just one example.

There is certainly no requirement for JMS to implement

Paul

On 12/24/05, Geoffrey Fox <gcf@grids.ucs.indiana.edu> wrote:
>
> I am not certain it is fair to require systems using WSRM to use JMS
> Our implementation of WSRM uses JMS like capabilities to get
> required functionality i.e. using JMS and WSRM is somewhat overkill
> Again whereas .NET and Apache should be able to exchange
> messages using WSRM, that is is not so clear if you are required to use
> JMS
> on Java side
>
> We face some problems like this in implementing workflow for sensor nets
> and
> use publish-subscribe to queue messages between stages and cope with
> semantics like WSRM. I think its hard to get all this to work without
> queuing and asynchronous messaging.
>
> Davanum Srinivas wrote:
>
> Jaliya,
>
> Not if it is over JMS. right?
>
> -- dims
>
> On 12/24/05, Jaliya Ekanayake <jaliya.ekanayake@gmail.com> <jaliya.ekanayake@gmail.com>
wrote:
>
>  Hi Paul and Eran,
>
> Do you expect the messages to the other types of web services to go over a
> network after synapse or does synapse invoke the services by itself.
> IMHO if the message goes over a network then we break the RM and after that
> point we don't have any gurantee.
>
> Thanks,
>
> Jaliya
>
>
> ----- Original Message -----
> From: Paul Fremantle
> To: synapse-dev@ws.apache.org
> Sent: Saturday, December 24, 2005 6:41 AM
> Subject: Re: Synapse and WSRM
>
> Eran
>
> Exactly right. For example, someone may have a PHP web service or PERL web
> service that doesn't support WS-Sec, RM, and Addressing. We can put Synapse
> in front of the existing service and "terminate" the RM, Sec, and
> Addressing and then forward the simple SOAP to the Perl service.
>
> Paul
>
>
> On 12/24/05, Eran Chinthaka <chinthaka@opensource.lk> <chinthaka@opensource.lk>
wrote:
>
>  Hi Jaliya,
>
> Let me try to answer this.
>
> What if the service is not behind an RM Endpoint Manager. One of the
> usages of Synapse, as I think is, it provides the infra-structure for
> other web services engines. For example, you can have your Synapse
> engine as a proxy for all the web services engine that are inside your
> company. And this Synapse engine will provide you QoS services like RM,
> security, etc. so that the internal web services engines do not need to
> worry about them. To me, this is one of the great features of Synapse.
> So I believe Synapse should worry abt it.
>
> Chinthaka
>
> Jaliya Ekanayake wrote:
>
>        Hi Paul,
>
> Please explain why would synapse worry about in-order delivery. If the
> service is behind a RMEndpoint manager then it will handle the INORDER
> delivery.
> It does not matter whether we lose the order in the mediation level.
>
> Thanks,
>
> Jaliya
>
>
>
>
> Folks At the hackathon we started talking about RM. The problem with
> RM is where the messages have to be delivered inorder. In that case,
> it is quite possible for the current message to be held up. We do not
> have support in Synapse for the current processor to be paused and
> restarted. Its possible we have to add it in the future. But there is
> a simpler option we can try first: The message comes in. We process it
> and it hits the RM processor. The RM processor always finishes this
> "round" of processing (i.e. it returns false). At some point possibly
> now or in the future, the RM processor will re-inject the message. The
> processor will inject the messages inorder (assuming it is configured
> that way). The RM processor can also set some context flag to identify
> that the message has been processed by the reliability. The message
> now starts its second round of processing. It is up to the configurer
> to make the second round do something different and process the
> message correctly. This can be done using a rule that checks for the
> RM context flag we set earlier. So in other words, we are splitting
> the flow into two flows. The first flow is up until the RM processor
> and the second flow starts when the message is delivered inorder. I
> think we can probably hide some of this in future iterations, but I
> believe this is a good starting point. Paul -- Paul Fremantle
> VP/Technology, WSO2 and OASIS WS-RX TC Co-chair
> http://bloglines.com/blog/paulfremantlepaul@wso2.com
>
>  <https://webmail6.pair.com/src/compose.php?send_to=paul%40wso2.com> <https://webmail6.pair.com/src/compose.php?send_to=paul%40wso2.com>
>
>  "Oxygenating the Web Service Platform", www.wso2.com
>
>
>
>                 ---------------------------------------------------------------------
>
>  To unsubscribe, e-mail:
>
>  synapse-dev-unsubscribe@ws.apache.org
>
>  For additional commands, e-mail: synapse-dev-help@ws.apache.org
>
>
>        --
> Paul Fremantle
> VP/Technology, WSO2 and OASIS WS-RX TC Co-chair
>
> http://bloglines.com/blog/paulfremantlepaul@wso2.com
>
> "Oxygenating the Web Service Platform", www.wso2.com
>
>      --
> Davanum Srinivas : http://wso2.com/blogs/
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: synapse-dev-unsubscribe@ws.apache.org
> For additional commands, e-mail: synapse-dev-help@ws.apache.org
>
>
>
>
> --
> :
> : Geoffrey Fox  gcf@indiana.edu FAX 8128567972 http://www.infomall.org
> : Phones Cell 812-219-4643 Home 8123239196 Lab 8128567977
> : SkypeIn 812-669-0772 with voicemail, International cell 8123910207
>
>


--
Paul Fremantle
VP/Technology, WSO2 and OASIS WS-RX TC Co-chair

http://bloglines.com/blog/paulfremantle
paul@wso2.com

"Oxygenating the Web Service Platform", www.wso2.com

Mime
View raw message