alfresco environment and production deployment

Recently started evaluating Alfresco. We have to deploy Alfresco (Community Edition) in our client environment. The client has 3 main offices and 10 operating divisions spread geographically.

They process a whopping 50 lakh(5 million) documents an year. Please provide tips/best practices to handle such a heavy load.

How should we setup our environment..Should we use a app server like JBoss or Tomcat? How do we go about the deployment? Does it need开发者_如何学Go file system replication along with the server clusters..

Confused about this.

Regards

Vish

---

The bare number of documents is not enough to draw a deployment architecture. Usually you should also ask yourself:

• how many casual vs concurrent users will access the system?

• what's the read/write ratio going to be?

• which interfaces are you allowing your users to access (HTTP/WebDAV/CIFS)?

• are you doing anything on content, e.g. transformations, renditions...?

For an average case, Alfresco provides some guidelines based on a Transactions Per Second estimation. Depending on the list above, you might need up to two servers, which would also ensure HA and failover.

In case you're building such a cluster, the deployment scenario for the Alfresco repository will most likely involve no content replication, but rather shared DB and FileSystem.

Be aware that Alfresco Share, being an almost stateless separate web application that accesses the repository via HTTP, can scale independently.

The application server you're going to use is not so meaningful here, it depends mostly on your personal preferences. Tomcat is generally enough, even though YMMV.