Release Notes

Optimize an SC Series array for an Oracle data warehouse
22 Optimizing Dell EMC SC Series Storage for Oracle OLAP Processing | 2009-M-BP-O
3.6.2 Configure I/O for bandwidth
Storage configurations for a data warehouse should be chosen based on the I/O bandwidth and not just on
the overall storage capacity. The following factors should be considered:
• The expected throughput of the data warehouse and all other applications that share the SC Series
array.
• The number of drives — more drives increase overall bandwidth
• Types of drives — SSDs offer higher transfer bandwidth per drive. Therefore, fewer SSDs in Tier 1
might be used to provide similar bandwidth as HDDs.
• Higher speed 12Gb SAS channels are available for back-end drive enclosure connectivity in some
SC Series arrays. Choose the highest speed wherever available.
• Account for all components along the I/O path and ensure that they are adequately configured to
handle the total aggregated bandwidth, for example, the array front-end FC/iSCSI/SAS channels, the
number and speed of server HBAs, and connectivity.
3.6.3 Capacity planning
The advanced features of SC Series storage such as thin provisioning, auto-tiering, multi-RAID levels, data
reduction features, and the snapshot capability make the best and most efficient use of storage space
possible. To figure out the capacity needed to support a data warehouse, consider the following factors that
have direct effects on the storage consumption. Also, storage capacity can be dynamically added or removed
from the SC Series array without affecting the applications. Thus, customers only invest in what they need
and still maintain the flexibility to meet any future growth. When new drives are added, data is redistributed
and rebalanced across all drives automatically to maximize performance.
Consider the following when planning for capacity needs:
• RAID types, such as RAID 10 or RAID 5/6
• Use of deduplication and compression
• Snapshot frequency and retention policy
• Number of data warehouse environments, full-size copy plus thin clones
• Data growth rate, factoring new data added each day, plus new archivelogs, plus new
customized/summarized data
• Data change rate, considering updates or refreshes of existing data and reports
• Oracle ASM redundancy level: 1x size for External, 2x size for Normal, and 3x size for High
• Storage profiles
• Amount of archive logs to keep
• Tier 1 storage should have enough storage to hold the active data and aging data until they are aged
out to lower tier. For accessible data pages, the aging period is 12 Data Progression cycles. For
inaccessible data pages, they are migrated to the lower tier in the next Data Progression cycle. See
the Dell Storage Manager Administrator’s Guide and Dell EMC SC Series Arrays and Oracle for more
information on snapshots and Data Progression.
• Estimate storage per tier. For example:
o Estimate tier 3 storage based on the amount of historical data, plus snapshots, plus data
growth.
o Estimate tier 1 storage based on the amount of customized/summarized data, plus frequently
accessed data, plus snapshots, plus new data.