Dell Compellent Series 30 - Manual (Page 338)

330 Storage Center 5.5 System Manager User Guide
Remote Instant Replay
• Define a remote system to replicate to.
• Replications are mapped to or from the remote system.
• There is at least one volume on the remote system to replicate to, of equal or greater
size than the volume you are replicating.
Estimating Bandwidth
Replication bandwidth cost, capacity, availability, and usability are key considerations when
developing a replication plan. Consider the time and cost of replicating the initial data load
from the Replicating System to the Remote Connection system. Consider how far behind
a replication is allowed to become after it has been established. The larger the tolerance
for missing data, the less you have used peak utilization time, requiring less overall
bandwidth that may go under-utilized during off hours. There are two basic strategies in
determining the amount of bandwidth required for replication. Have a good idea of the
volumes you want to replicate from your Disaster Recovery Plan. It is very possible you will
use the first strategy to get started and the second once you are in production.
• Derive the Required Bandwidth based on modeling the list of volumes to be replicated
and the recovery requirements.
• Replicate data as required. Add volumes to the replication as bandwidth allows. Tune
Replay schedules to meet Disaster Recovery requirements with the available
bandwidth.
There are other considerations outside the realm of the Storage Center that can affect
replications. Nearly all of these considerations relate to using iSCSI connectivity for
replication.
• Quality of the link (dropped packets, fragmented packets, resends, link down)
• Competition on the link (other traffic)
• Ability of the link to handle bursts of traffic
• Ability of the link to handle larger packet sizes
• Latency on the link
• Security on the link (if encryption is required)
• Cyclical business cycles affecting bandwidth requirements