Storage locations can potentially be very far apart, separated by
thousands of miles, even around the globe. The real consideration
in selecting distance is that of
latency.
@30656 Latency is the enemy of distance. If you're doing
synchronous remote
mirroring -- copying data in real time with
no loss of data -- you're striving for a
recovery point objective (RPO) and
recovery time objective (RTO) near zero. In
synchronous mode, you need to keep distances short to minimize
latency and use a network with enough bandwidth to handle your
real-time data load. This all changes if you run in asynchronous
mode. Since you're not operating in real time -- several minutes
or even several hours of delay between locations -- latency is
no longer a factor, allowing us to span hundreds or thousands of
miles.
Latency can exist in a data center, a campus or a metropolitan
area even where you have fast dedicated networks. If there is any
network bottleneck or congestion, or anything that causes delays,
delays are latency, and those work against distance. So, when
you're looking at long distances, don't just consider bandwidth;
also evaluate the effective latency and understand the
application's sensitivity to that delay.
Listen to the
SAN FAQ audiocast here.
Go back to the beginning of the
Storage Area Network FAQ Guide.