174 IBM TotalStorage DS6000 Series: Performance Monitoring and Tuning
The advantages of SAN Volume Controller remote copy is that we can implement such
relationships between two SAN Volume Controller clusters with different backend disk
subsystems. in this case, you reduce the overall cost of the disaster recovery infrastructure
implemented on the production site (first site) high performance backend disk subsystems
and on the recovery site low-cost backend disk subsystems, even if backend disk subsystems
copy services functions are not compatible (for example, different models, different
constructors...). Figure 6-4 represents an example of a remote copy relationship between two
SAN Volume Controller clusters. This relationship is established at the vDisk level and doesn’t
depend of the backend disk storage subsystem.
Figure 6-4 Synchronous remote copy relationship between 2 SAN Volume Controller clusters
6.1.4 SAN Volume Controller performance considerations
The SAN Volume Controller cluster is scalable up to eight I/O groups (eight pairs of SAN
Volume Controller nodes), and the performance is almost linear when adding more I/O groups
to a SAN Volume Controller cluster, until it becomes limited by other components in the
storage infrastructure. While virtualization with the SAN Volume Controller provides a great
deal of flexibility, it does not diminish the necessary to have a SAN and disk subsystems
which can deliver the desired performance.
In the following section, we present the IBM SAN Volume Controller concepts and discuss the
performance of the SAN Volume Controller. In this section, we assume there are no
bottlenecks in the SAN or on the disk subsystem.
Determine the number of I/O groups
The SAN Volume Controller cluster consists of I/O groups, where each I/O group consists of a
pair of SAN Volume Controller nodes (also called storage engine). Since the SAN Volume
Controller cluster is linearly scalable, we will first focus on performance for a single I/O group.