MainFrames.Com

 Site Map

 Home SiteMap Email
 


Main Storage

IBM z/OS Storage Management Design (ASM, RSM, VSM) " "Storage: This time it's for Real - The Real Storage Manager's Evolution for Scalability and Performance"

The Real Storage Manager (RSM) is an integral part of the z/OS operating system. This session will cover how RSM has evolved to support up to 4T of real storage with an added focus on the performance aspects of RSM. We discuss topics such as WLM page replacement and UIC, RMF UIC changes in z/OS 1.8, elimination of physical swap, pageable storage shortage handling, z/OS migration and coexistence considerations, performance, and real storage scalability considerations.


Two decades of policy-based storage management for the IBM mainframe computer

Today, storage management vendors see the need and business opportunity for an enterprise-wide policy-based storage management solution for their customers. In the middle 1980s, IBM introduced the Data Facility Storage Management Subsystem (DFSMS) as a policy-based storage management solution for large mainframe computer systems. As an integral part of the operating systems OS/390® and z/OS™, DFSMS continues to be enhanced. This paper provides an overview of DFSMS and describes a few of its recent enhancements.


In September 2006, z/OS V1R8 introduced a new real storage management system that is able to manage single z/OS images with up to 4 terabytes (TB) of real storage. Changing the infrastructure of z/OS storage management to support this extra storage is like trying to convert the Guggenheim Museum into a skyscraper! This article explains the remodel of the z/OS foundation and its new features.


z/OS Systems Management (Session technical level: Intermediate)


Review the OS/390 MVS Initialization and Tuning Guide Document Number SC28-1751-05 for more details.


Storage Map

The common area contains system control programs and control blocks. The following storage areas are

located in the common area:

  • Prefixed storage area (PSA).
  • Common service area (CSA).
  • Pageable link pack area (PLPA).
  • Fixed link pack area (FLPA).
  • Modified link pack area (MLPA).
  • System queue area (SQA).
  • Nucleus, which is fixed and nonswappable.

Each storage area in the common area (below 16 megabytes) has a counterpart in the extended common area (above 16 megabytes) with the exception of the PSA.

Each address space uses the same common area. Portions of the common area are paged in and out as the demands of the system change and as new user jobs (batch or time-shared) start and old ones terminate.

The private area contains:

  • A local system queue area (LSQA).
  • A scheduler work area (SWA).
  • Subpools 229, 230, and 249 (the authorized user key area).
  • A 16K system region area.
  • Either a V=V (virtual = virtual) or V=R (virtual = real) private user region for running programs and storing data.

Except for the 16K system region area and V=R user regions, each storage area in the private area below 16 megabytes has a counterpart in the extended private area above 16 megabytes.

Each address space has its own unique private area allocation. The private area (except LSQA) is pageable unless a user specifies a V=R region. If assigned as V=R, the actual V=R region area (excluding SWA, the 16K system region area, and subpools 229, 230, and 249) is fixed and non swappable.

Disclaimer

Copyright / Trademarks