Advertisement
meet business priorities, z/OS and System z can handle unexpected workload spikes
and help improve your system's efficiency and availability.
• The z/OS Workload Manager is a cornerstone to z/OS leadership in on demand
computing. With workload management, you define performance goals and assign
a business importance to each goal. You define the goals for work in business
terms, and the system decides how much resource, such as CPU and storage,
should be given to it to meet each goal. Workload Manager will constantly monitor
the system and adapt processing to meet the goals. The scope of the Workload
Manager extends from helping the management of incoming TCP/IP and SNA
traffic, to managing requests for I/O. z/OS middleware like DB2, CICS, IMS,
WebSphere MQ, and other WebSphere products can take advantage of WLM
to manage the priority and execution of transaction requests across the z/OS
system. For z/OS V1.12, WLM is planned to be updated with enhancements to
improve batch management.
• Batch processing windows can be shortened and optimized through several other
applications. For example, in z/OS V1.11, job streams using the IEFBR14 program
during the batch window can be run faster by enabling Allocation to delete data
sets without first recalling them. z/OS V1.12 is planned to be updated so IDCAMS
can avoid DFSMShsm recalls when deleting generation data groups (GDGs).
• Extending the scope of z/OS and System z management, the Capacity
Provisioning Manager for z/OS enables z/OS and the System z10 server to add
temporary capacity automatically when necessary, with or without operator
intervention. Capacity Provisioning for z/OS V1.12 is planned to use CICS and IMS
monitoring data to determine if additional resources are needed to meet service
level requirements for these workloads. What has taken minutes or hours to
discover, identify, decide, and resolve, now can happen automatically in seconds.
Details on the optimization improvements planned for z/OS V1.12 include:
• Initiator address spaces consume processor time on behalf of starting and ending
job steps that in prior releases are not associated with a particular batch job.
There can be considerable variation in the processor time consumed by an initiator
for different jobs. To help you better understand the resources consumed by
batch jobs and improve the accuracy of chargeback programs, z/OS V1.12 will
be designed to record the CPU time consumed for job steps in initiator address
spaces using new fields in SMF Type 30 records.
• The creation of new VSAM data sets with IMBED and REPLICATE attributes has
been unsupported since z/OS V1.3. These attributes, originally introduced to
improve performance on older DASD, typically act only to occupy additional space
and slow performance on modern cached DASD. In z/OS V1.12, the system will be
designed to remove these attributes automatically from VSAM data sets logically
dumped using DFSMSdss and migrated using DFSMShsm when DFSMSdss is
used as the data mover during restore and recall processing. An informational
message is planned to confirm that newly restored data sets no longer retain
these attributes.
• DFSMSdss and DFSMShsm are planned to exploit the Fast Reverse Restore feature
of the IBM System Storage
allow recovery to be performed from an active, original FlashCopy target volume
to its original source volume without having to wait for the background copy to
finish when the volume pair is in a full-volume FlashCopy relationship. DFSMSdss
will be enhanced to create full-volume copies using a new keyword in order to
support a Fast Reverse Restore function. DFSMShsm FlashCopy backup and
recovery operations will be designed to create full-volume FlashCopy relationships
when the devices support it. The Fast Reverse Restore function will support
the recovery of volumes associated with copy pool backups including Space
Efficient and Incremental FlashCopy targets. A new DFSMShsm SETSYS parameter
is planned to allow you to specify whether extent or full-volume FlashCopy
relationships are to be established between volume pairs when DFSMShsm
invokes DFSMSdss to perform fast replication backup and recovery.
• It is planned that DFSMSrmm will help with reporting of data sets and logical
volumes which are copy exported from a TS7700 virtualization engine.
• WLM will be designed to consider resource group maximums and the projected
increase in system or sysplex demand before starting initiators during resource
IBM United States Software Announcement
210-008
DS8000
Series. This function will be designed to
TM
TM
IBM is a registered trademark of International Business Machines Corporation
22
Advertisement
