traceability and auditing: which one is better - JMS flow or JdbcAuditor

["Rossmanith, Philipp" <ph...@t-systems.es>]

Hi,

I'm currently investigating which is the preferred way to achieve
message traceability and auditing with SM. The idea would be to persist
message exchanges and to give some kind of analysis interface to it.

I found that there are 2 main ways to achieve message persistence:
- Via JMS and clustering
(http://incubator.apache.org/servicemix/clustering.html): "...
durability (persist messages to disk/database) to ensure they survive
catastrophic hardware failure..."
- Via adding a JdbcAuditor to the JBIContainer
(http://www.nabble.com/Capture-any-message-off-the-NMR--tf1902644s12049.
html#a5552188)

Now I'm wondering: which is the preferred SM-way? Are there any
advantages of one procedure over the other?

At a first glance, the JMS way seems to be superior (ActiveMQ
auto-discovery and clustering
-http://activemq.apache.org/discovery.html,
http://activemq.apache.org/clustering.html- plus out of the box message
persistence -http://activemq.apache.org/persistence.html,
http://activemq.apache.org/replicated-message-store.html).

I should mention that in the end, we will have to implement a cluster,
anyway.

Thanks in advance for any comments,
Ciao, Philipp

This e-mail may contain confidential or privileged information. Any unauthorised
copying, use or distribution of this information is strictly prohibited.

Re: traceability and auditing: which one is better - JMS flow or JdbcAuditor

[Guillaume Nodet <gn...@gmail.com>]

The JMS flow is just about clustering.  You won't have any added
traceability or auditing with it, as messages will be consumed and
lost.  The JdbcAuditor is one way to add auditing features to ServiceMix
(it can be easily changed or enhanced), but this is a way to keep the
exchanges so that they can be re-read after they have been processed.

Even replicated message stores in AMQ won't help, as the point is that
JMS messages are removed from the store after being consumed.
Currently, there's no easy way to keep JMS messages for auditing purposes.

On 4/17/07, Rossmanith, Philipp <ph...@t-systems.es> wrote:
>
>
> Hi,
>
> I'm currently investigating which is the preferred way to achieve
> message traceability and auditing with SM. The idea would be to persist
> message exchanges and to give some kind of analysis interface to it.
>
> I found that there are 2 main ways to achieve message persistence:
> - Via JMS and clustering
> (http://incubator.apache.org/servicemix/clustering.html): "...
> durability (persist messages to disk/database) to ensure they survive
> catastrophic hardware failure..."
> - Via adding a JdbcAuditor to the JBIContainer
> (http://www.nabble.com/Capture-any-message-off-the-NMR--tf1902644s12049.
> html#a5552188)
>
> Now I'm wondering: which is the preferred SM-way? Are there any
> advantages of one procedure over the other?
>
> At a first glance, the JMS way seems to be superior (ActiveMQ
> auto-discovery and clustering
> -http://activemq.apache.org/discovery.html,
> http://activemq.apache.org/clustering.html- plus out of the box message
> persistence -http://activemq.apache.org/persistence.html,
> http://activemq.apache.org/replicated-message-store.html).
>
> I should mention that in the end, we will have to implement a cluster,
> anyway.
>
> Thanks in advance for any comments,
> Ciao, Philipp
>
> This e-mail may contain confidential or privileged information. Any
> unauthorised
> copying, use or distribution of this information is strictly prohibited.
>



-- 
Cheers,
Guillaume Nodet
------------------------
Principal Engineer, IONA
Blog: http://gnodet.blogspot.com/