Using Alfresco together with LifeRay/Jackrabbit CMIS support

[kidahl]

Hi, I need some advice on how/if the CMIS standard changes the way we could use Alfresco together with Liferay/Jackrabbit.

The goal of my client is to have Alfresco as a primary CMS source. I am working on a Liferay Portal project that has a lot of "out of the box" collaboration functionality based on JSR 170 going against a Liferay provided Jackrabbit store. Given the known issues of accessing Alfresco(and probably other CMS systems) remotely using the JSR 170 remote access serialization wrapper (it is basically described as a quick hack with performance issues), we have decided not to use JSR 170 to access Alfresco remotely with these "out of the box" portlets.

That means that the content in "out of the box" jsr 170 portlets will be bound to the local Jackrabbit store, which is really not what our client want (Everything should be accessible from Alfresco, as it is the main CMS).

I had just about given up on this until I saw that Jackrabbit now has a sandbox implementation of CMIS:
http://wiki.apache.org/jackrabbit/SandboxCMIS

Can this be used in a similar manner to OpenSearch so that Alfresco can register CMIS sources and then aggregate content from those CMS'es?

I am interested in how this could solve our issues in practice.
-Can we use the "out of the box" Liferay portlets storing content in JackRabbit and then seamlessly access this information from Alfresco as well?
-Can we search in Alfresco, and it will return and display hits from JackRabbit using CMIS?
-Can for example one Alfresco document refer/link to a document available through CMIS and show it to an Alfresco user?

..or are we still talking about yet another API that needs custom code both in the Liferay Portal and Alfresco to seem "unified?"

I am aware that CMIS is not finalized, I am looking for the "strategic" implications of opening JackRabbit up with CMIS in relation to Alfresco.

Thanks,
Karl Ivar Dahl

[paolomoz]

Hi,
Regarding the following:

–I am interested in how this could solve our issues in practice.
—Can we use the "out of the box" Liferay portlets storing content in JackRabbit and then seamlessly access this information from Alfresco as well?
—Can we search in Alfresco, and it will return and display hits from JackRabbit using CMIS?
—Can for example one Alfresco document refer/link to a document available through CMIS and show it to an Alfresco user?

I have to add some more details on the Jackrabbir CMIS implementation. It is a great initiative inside the Jackrabbit project and unforstunately it is not usable yet. Sure, it promises, as CMIS does, to solve all the three points that you mention.
What you should figure out is how to build your interface; CMIS will be able to provide a common layer of comunication between the two repositories, then you have to manage those contents.

To summarize, CMIS could be one of the best solutions, but you have to evaluate your effort from an higher perspective and see if you want to build an interface for this.
Hope this helps,

[kidahl]

Hi paolomoz, thank you very much for your reply, and sorry for my late follow-op (i've been away).

So if I understand correctly, I will not be able to "register" a Jackrabbit CMIS source in Alfresco, and then access its content as if it was stored locally in Alfresco using existing Alfresco functionality.

However, I will be able to access LifeRay Jackrabbit content and display it in the separate Alfresco server if I write custom code for it on the Alfresco side. In my view this would put CMIS on the same architecture abstraction level as SQL.

This would still be an important improvement, as this remote access would not be possible without the JackRabbit CMIS support, right?

To make it clearer what kind of scenarios I was considering:
-We use Liferay/Jackrabbit to store content, but configure JackRabbit to actually store contents using CMIS on the Alfresco Server, completely transparent to the LifeRay portlets still using JSR 170 (transparent remote storage).
-We create custom portlets that use JSR 170 against JackRabbit, but the contents are actually retrieved from Alfresco (transparent remote access)
-Other "enterprise" clients only access Alfresco as a single content repository, even when content actually are stored across many repositories (aggregation).

Any thoughts are appreciated.
-Karl Ivar