White Papers

11 PS Series Asynchronous Replication Best Practices and Sizing Guide | BP1012
A volume’s replica set is identified by the volume name with a numbered extension. The number corresponds
to the number of replication partners, in the order they were added. For example, an inbound volume from the
first replication partner, with the name repl-vol1, will have a replica set with the name repl-vol1.1 as shown in
Figure 3. A volume from a second replication partner, also with a source volume named repl-vol2 will have a
replica set named repl-vol2.1. Each replica that makes up the replica set is identified with a date and time
stamp that correlates to the time that the replica was created.
Navigation tip: Group Manager GUI (secondary) > Replication > Inbound Replicas
Inbound Replicas
Typically a volume will be replicated across a network (SAN or WAN). However, Dell EMC also provides a
Manual Transfer Utility (MTU). This tool allows an administrator to create a local replica on a portable storage
system (such as an external USB disk). After transporting the portable storage to the secondary site, you can
transfer the replica data from the portable storage system to the secondary group. Once the volume transfers,
then asynchronous replication will synchronize only new changes that occurred after the copy onto portable
storage was made to the remote replica, conserving WAN bandwidth. If necessary, you can use the MTU for
both initial and subsequent delta replication events. The Manual Transfer Utility can also be used to conserve
WAN bandwidth if a large one-time update is applied to a volume.
For unusually large amounts of data, expedited shipping of external media (tape or disk) may be faster than
electronic transfer across a network. Also, depending on how often you need to replicate, shipping may be
more cost-effective than provisioning the connection path with the necessary bandwidth.
The design of the network link between primary and secondary groups will affect how fast changes replicate
between arrays. This in turn can also dictate how many replicas can be created in a given time period. For
example, if you have a constant volume data change rate and it takes four hours to replicate changes made
(since the last replica) then in practice you should be able to complete up to six replication processes within a
24-hour period.