While this sounds simple enough, it is a very complicated (and usually extremely expensive). The complicated part is the software and/or firmware needed to handle everything. On the surface, the controller or computer writes data to the local disk and somehow creates a separate network packet for the remote system. Then everyone waits until an affirmative response comes back. If you use a modem, then expect performance to be 1000:1 slower (a 1 second update to the database may now take 15 minutes).
But what if the link goes down or becomes erratic? Without some sohpisticated code, your primary system will be hung or at least the integrity of the remote data may be in question.
And finally, can you afford a fast enough link to the remote location? The FC60's are much too fast for a T1 or DS3 line so you'll likely need a fiber link, perhaps costing $25-50,000 a month. Again, the massive impact of a slow link on your local system will determine how much you'd be willing to spend.
Now if the remote system is just in another building, it becomes a more reasonable design but now the question of how safe the remote location might be if a huge fire takes out everything. Distance becomes an important criteria in this case.
Bottom line is that the FC60's are not designed for any data replication capability. You'll need a much more sophisticated box such as the XP arrays and lots of extra cash to spend to reduce the impact on the local system. A good rule of thumb is: how much does each hour of downtime cost the company? Is it possible to survive a one day outage? If so, tapes in briefcase are pretty darn fast compared to the cost of transmitting 500Gb over a long distance.
Bill Hassell, sysadmin
