Determining whether tiering is appropriate for your storage environment

Edit online

Before you tier data from disk storage on directory-container storage pools to cloud or tape storage, review the information about storage types and determine which type best meets your business requirements. In general, cloud and tape storage are appropriate for long-term storage of data that is infrequently accessed and that does not require quick retrieval.

Consider the advantages and disadvantages of each storage type:

Disk storage on directory-container storage pools

Disk storage on directory-container storage pools (subsequently referred to as disk storage) is useful for data that must be frequently accessed and quickly retrieved. However, disk storage is typically more expensive than cloud or tape storage. For this reason, some users tier their data to cloud or tape storage to save space and reduce costs. The amount of space savings depends on how well the data is deduplicated on disk. If the data is efficiently deduplicated on disk, you might not achieve the expected space savings by tiering the data.

Tip: To determine whether data is effectively deduplicated on disk, take one of the following actions:

Generate data deduplication statistics by using the GENERATE DEDUPSTATS command. Then, view the statistics by using the QUERY DEDUPSTATS command.
From the administrative client, run the SELECT * from SUMMARY command. For examples, see Viewing data deduplication statistics.

In general, if the data deduplication ratio is 4:1 or better, the data is efficiently deduplicated. For example, if you have 100 KB of data and reduce the size to 25 KB, the data deduplication is considered efficient. If the data on disk storage is not efficiently deduplicated, and there is no requirement to quickly retrieve the data, consider tiering the data to cloud or tape storage frequently with a brief tiering delay.

Cloud storage

Cloud storage can be more scalable and cost effective than disk storage. However, the time that is required to retrieve data from the cloud is typically longer than from disk. For cloud tiering, you must find a cloud provider that uses the Microsoft Azure cloud computing system or a cloud computing system that uses the Simple Storage Service (S3) protocol, such as IBM Cloud® Object Storage or the Amazon Simple Storage Service (Amazon S3). For the latest information about cloud object storage services that are supported, see technote 2000915.

Tape storage

Tape storage can be more scalable and cost effective than disk storage, but the time that is required to retrieve the data from tape is typically longer than from disk. For small file workloads (with an average file size of 50 KB or smaller), the process of tiering data to tape can take more time than a typical tiering window allows for. You must also consider the effort that is required to share tape drives and coordinate access to tape volumes.

Tip: To calculate average file size, use the IBM® Db2® Command Line Processor to run the following commands:

db2 "connect to tsmdb1"
db2 "set schema tsmdb1"
db2 "select avg(logical_size) from sc_all_objects for read only with ur

Tape tiering flowchart: Before you tier data to tape, you can review the How to determine whether tape tiering is appropriate flowchart in Determining whether tiering is appropriate for your storage environment.

The decision points in the flowchart are summarized here:

If the data must be retrieved quickly, tiering data to tape might not be appropriate.
If data deduplication on disk is highly successful, tiering data to tape might not help you achieve the expected space savings.
If the data must be accessed frequently, tiering data to tape might not be appropriate because accessing data from disk is typically faster.
If the data includes many small files, tiering data to tape might not be appropriate because tiering operations might exceed the allotted time.
If none of the previously listed conditions are true, tiering data to tape might help to optimize storage and reduce costs.

Tiering options

If you decide to tier data to cloud or tape, you must also select a tiering option. The following options are available:

All data: When data meets a specified age threshold, the data is tiered. (If you configure tiering by using administrative commands, this option is known as tiering by age.)
Inactive data: When data meets a specified age threshold, only inactive data is tiered. (If you configure tiering by using administrative commands, this option is known as tiering by state.)

Flowchart for selecting a tiering option: To select a tiering option, you can review the How to select a tiering option flowchart in Determining whether tiering is appropriate for your storage environment.

Guidelines for selecting a tiering option are available in the flowchart and in the following table:

Client type	How to tier data
IBM Storage Protect backup-archive client	For frequently accessed data, specify the Inactive data option. For infrequently accessed data, specify the All data option.
IBM Storage Protect for Databases: Data Protection for Microsoft SQL Server	For frequently accessed data, specify the Inactive data option. For infrequently accessed data, specify the All data option.
IBM Storage Protect for Databases: Data Protection for Oracle	For all data, specify the All data option.
IBM Storage Protect for Databases: Data Protection for SAP	For all data, specify either option; in both cases, all data that meets the age threshold is tiered.
IBM Storage Protect HSM for Windows	For all data, specify either option; in both cases, all data that meets the age threshold is tiered.
IBM Storage Protect for Mail: Data Protection for IBM Domino	For frequently accessed data, specify the Inactive data option. For infrequently accessed data, specify the All data option.
IBM Storage Protect for Mail: Data Protection for Microsoft Exchange Server	For frequently accessed data, specify the Inactive data option. For infrequently accessed data, specify the All data option.
IBM Storage Protect for Space Management	For all data, specify the All data option.
IBM Storage Protect for Virtual Environments	For production environments, do not tier data. For test and development environments, specify the All data option.
IBM Storage Protect Plus	For all data, specify the All data option.

Restriction: Tiering is not available for data from NetApp network-attached storage (NAS) file servers.

For an example of tiering data from different client types, see Tiering scenario: A step-by-step example.