White Papers

17 Dell Storage PS Series Architecture: Load Balancers | TR1070
9 Tiering with the APLB
When provided with tiered resources in a pool, such as arrays with different spindle speeds or set to different
RAID types, the APLB is able to use them to tier the workload. This is not limited to any particular RAID type,
interface I/O type, spindle speed, number of disks, or PS Series controller generation, since all of these
factors are abstracted by the use of latency, which is the primary factor when deciding when to rebalance the
workload. The ability to tier gives the customer greater flexibility in selecting products that provide the correct
combination of performance and capacity for their environment, since any of the factors mentioned previously
could be relevant to creating differences in latency between PS Series members in a pool. For example,
combining large-capacity PS66x0 class arrays with lower capacity PS62x0 arrays using disks that provide
higher I/O to get better total system ROI may be the appropriate design for some customers. Others might
choose to combine members with 10K SAS and members with SSD to meet their application workload
requirements. Many other configurations are possible, these are simply examples. For one example of tiering,
see Figure 2. For additional examples, see appendix A.
The ability of the APLB to tier also means that PS Series products are capable of placing the inactive data on
lower-cost media, allowing more efficient use of resources to provide higher I/O to customer applications at a
lower overall cost. This includes snapshot data, which may end up on lower cost media compared to the
active data that they are protecting. Similarly, applications like OLTP databases or those that leverage PS
Series thin clones or similar technologies, (such as VDI) that concentrate the majority of the read activity into
a small portion of the total capacity can take advantage of the APLB. In these cases, when provided with
heterogeneous resources (those with different random I/O capabilities), the APLB optimizes these workloads
by migrating the frequently used data to members with greater random I/O capabilities, such as SSDs, while
utilizing capacity-optimized members with lower I/O capabilities to host the bulk of the content.
When setting up a tiered environment, the capacity load balancer typically chooses a maximum of three pool
members to distribute the slices for a given volume; if the ultimate goal of a particular SAN design is to
provide an automatically tiered storage environment, then the pool should be made up of heterogeneous
storage arrays and be limited to a maximum of three members for the most predictable results. Creativity in
design, such as using a combination of tiered resources at the same RAID level and RAID preference
settings, can provide predictable behavior with larger pools containing tiered resources. This is generally not a
consideration when tiering is not a primary goal, in which case homogeneous resources would be deployed in
the pool, up to the limits of the pool.