Trends > What's New

Make Storage a Snap

Improving DFSMShsm performance with good management practices

Gain unprecedented access into the heart and soul of the IBM* DFSMS Hierarchical Storage Manager (DFSMShsm*) with a new storage-management solution offered by the Tivoli* Advanced Reporting for DFSMShsm and Tivoli Advanced Audit for DFSMShsm products. This solution provides visibility, automation and flexibility for storage administrators and includes the Tivoli Advanced Backup and Recovery Manager for z/OS*.

Now storage administrators can see which datasets waste valuable resources and degrade DFSMShsm’s performance. They can pinpoint by dataset name the offenders and immediately implement permanent resolutions. They can execute audits in a fraction of the time and with no impact on DFSMShsm’s performance while automating resolutions of audit errors so manual interaction is no longer necessary. They can archive inactive data out of the DFSMShsm management arena, reducing the size of the control datasets (CDS) and DFSMShsm’s resource utilization while maintaining complete access to the archived data. They can provide audit reporting and recovery services with full rename capabilities and automatic archive deletion.

Automatic Space Management

One of DFSMShsm’s main functions is Automatic Space Management, the migration of inactive data from primary storage to storage lower in the hierarchy. This function makes room on primary storage for new dataset allocations and data frequently used by applications. DFSMShsm uses management classes defined to DFSMS to provide lifecycle management policies about the data in the environment. There are typically several policies in place to manage the wide variety of data in the environment.

The management class policies are intended to maximize storage availability and keep frequently used data on primary storage. Over time, dataset access patterns change and new data is introduced into the environment, resulting in migrating data too soon or not at all.

Data that’s migrated and subsequently recalled in a relatively short amount of time is called thrashing. In some environments where hourly interval migration is used, data may be moved back and forth several times a day. DFSMShsm uses CPU and I/O resources to move data from primary storage to a lower tier and back again. Datasets that are thrashing use these resources unnecessarily and degrade DFSMShsm’s overall performance.

Conversely, datasets that are inactive but remain on primary storage waste valuable space. DFSMShsm doesn’t use any resources to move this data but the inactive data takes up space on primary disk, the redundant disk and, if remote replication is implemented, space on the remote storage as well. Often, data on primary storage is backed up daily using full-volume dumps or replication technologies to other storage.

Next page: >>