You should not encounter any difficulty backing up and restoring to similar hardware. Simply perform a full backup of the server, system state and Exchange databases. Restoring to dissimilar hardware presents considerable challenges. Here are some useful URLs: <a href="http://www.microsoft.com/technet/technetmag/issues/2006/08/ExchangeRecovery/default.aspx">http://www.microsoft.com/technet/technetmag/issues/2006/08/ExchangeRecovery/default.aspx</a> <a href="http://www.msexchange.org/tutorials/Restoring-Exchange-Server2003-Alternate-Hardware.html">http://www.msexchange.org/tutorials/Restoring-Exchange-Server2003-Alternate-Hardware.html</a> <a href="http://support.microsoft.com/kb/249694">http://support.microsoft.com/kb/249694</a> One of the SLA requirements for the new Exchange 2003 clustered solution was for the nightly backup jobs (disk to disk and then to tape) to take no longer than 4 hours using NTBACKUP. This was so that the Exchange online maintenance had sufficient time to run its course prior to staff coming into work in the early morning and to begin using the system. This target proved to be quite a challenge at first, with throughput of 640MB/min being achieved. This wasn’t sufficient. To rectify this, some registry modifications were required, along with a revised version of NTBACKUP, which included a new switch - the FU switch(!) – which stands for ‘file unbuffered’ in case you were wondering. Throughput was then increased to a more than acceptable 1.2GB/min which enabled the backup job to be completed within the SLA prescribed timeframe.

You should not encounter any difficulty backing up and restoring to similar hardware. Simply perform a full backup of the server, system state and Exchange databases. Restoring to dissimilar hardware presents considerable challenges. Here are some useful URLs: <a href="http://www.microsoft.com/technet/technetmag/issues/2006/08/ExchangeRecovery/default.aspx">http://www.microsoft.com/technet/technetmag/issues/2006/08/ExchangeRecovery/default.aspx</a> <a href="http://www.msexchange.org/tutorials/Restoring-Exchange-Server2003-Alternate-Hardware.html">http://www.msexchange.org/tutorials/Restoring-Exchange-Server2003-Alternate-Hardware.html</a> <a href="http://support.microsoft.com/kb/249694">http://support.microsoft.com/kb/249694</a> One of the SLA requirements for the new Exchange 2003 clustered solution was for the nightly backup jobs (disk to disk and then to tape) to take no longer than 4 hours using NTBACKUP. This was so that the Exchange online maintenance had sufficient time to run its course prior to staff coming into work in the early morning and to begin using the system. This target proved to be quite a challenge at first, with throughput of 640MB/min being achieved. This wasn’t sufficient. To rectify this, some registry modifications were required, along with a revised version of NTBACKUP, which included a new switch - the FU switch(!) – which stands for ‘file unbuffered’ in case you were wondering. Throughput was then increased to a more than acceptable 1.2GB/min which enabled the backup job to be completed within the SLA prescribed timeframe.