Stubbing – (Not) A Solution
Some vendors propagate that their archiving solution supports so called "stubbing" to lower the amount of data stored on local systems. Stubbing breaks e-mail messages into individual parts, such as header, body, attachment, and so on. The actual message contents and attachments are relocated to a separate storage system; all that remains in the local e-mail client are the headers of the e-mail message.
While stubbing may sound like a great solution at first, it has several drawbacks in practice.
By splitting an e-mail message into its individual components, it inevitably loses its original condition. This makes it considerably more difficult to prove the authenticity of a message and that it was not changed. The problem is even more apparent when taking civil law into account, because e-mail messages that were split into components may lose their value as evidence.
Inefficient Search and Restore
Even if an archiving solution that uses stubbing has an integrated search feature, it may still be difficult to find e-mail messages inside the archive. The reason is that, because stubbing was used, a standard search from inside the client will only return the information that was stored in the header. Stubbing therefore makes searching for and restoring archived e-mail messages – one of the most important aspects of an e-mail archive – considerably more cumbersome than it has to be.
Reduced Database Performance
Because stubbing causes the number of search index entries to rise with each newly created stub, it may make systems that should be performing better (since they have less data stored locally) start slowing down even more. The result is that performance falls, and it might become necessary to continuously invest in faster hardware to keep performance at an acceptable level.