Our performance test turned into a 10-horse photo-finish. In other words, all 10 products tested hit pretty much the same numbers.

Our test bed involved both measuring the time it took each product to archive an Exchange 2003 server with a (4.3GB) mailbox of more than 60,000 real messages dating back nearly 10 years, and, subsequently measuring the impact that running an archiving process on an Exchange 2007 server with nearly 10,000 mailboxes and a total of about 5 million messages had on overall messaging performance.

In our initial test, we had each product do a full archive of the Exchange 2003 store and all the products archived the messages at about the same rate, 30,000 messages per hour, plus or minus about 900 messages per hour. These numbers indicate that the bottleneck was the Exchange servers rather than the archiving solutions.

Archiving old messages is something that only needs to be done once – after that the archiving process only needs to deal with new incoming messages.

To measure the impact of the archiving process on the messaging server itself, we delivered e-mail to the Exchange 2007 server at a rate of about 2,500 messages per hour. Server loads were recorded first with archiving disabled and then with it enabled. Server loads rose between 8% and 11% across all products tested. We were not surprised by the proximity of the load test results because all products use one of three methods of getting the messages from the Exchange server: journaling, copying all incoming messages to an archiving mailbox; or using the SMTP connector in Exchange. These are all methods available through the Exchange server itself - no product uses a separate piece of software to intercept messages.

< Return to main story: Symantec's Enterprise Vault wins top spot in 10-vendor test >