Browse Prior Art Database

Least Recently Used Methodologies within EDAM Disclosure Number: IPCOM000049320D
Original Publication Date: 1982-Apr-01
Included in the Prior Art Database: 2005-Feb-09
Document File: 4 page(s) / 56K

Publishing Venue


Related People

Cesa, LJ: AUTHOR [+2]


Disclosed are Least Recently Used methodologies used in an Extended Direct Access Method (EDAM), developed to overcome MVS (Multiple Virtual System) main storage limitations.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 47% of the total text.

Page 1 of 4

Least Recently Used Methodologies within EDAM

Disclosed are Least Recently Used methodologies used in an Extended Direct Access Method (EDAM), developed to overcome MVS (Multiple Virtual System) main storage limitations.

A set of programs, (EDAM) has been developed to manage large amounts of data spread over several direct-access data sets. These programs utilize Least- Recently-Used (LRU) methodologies to manage both page frames and the Data Control Blocks (DCBs) for the multiple data sets which may be associated with an EDAM file. These methodologies help to reduce the amount of paging I/O.

An EDAM file consists of many address spaces of 32 bits, for example. The address hierarchy uses two levels of directories to access a particular address space, and two levels of directories to access data within an address space. All of the directories involved, as well as all application data, are contained in 4096 byte pages, for example. This ensures that data and directories within an EDAM file will all fit in an EDAM main storage page frame as well as an EDAM disk slot, allowing them to coexist within the paging space.

EDAM has two resources to manage during its processing: buffer frames, and a limited amount of data set DCBs. In managing these two resources EDAM uses a Least Recently Used queue for each. This allows a mapping of the real data required for processing onto the limited amount of main storage and DCBs available.

The DCB queue, as shown in Fig. 1, is used to manage a collection of data sets whose I/O priority is constantly changing. The queue is initialized to a list of available DCBs and is updated whenever a data set is opened, or a read/write operation occurs. To help associate a disk address within an EDAM file to a data set and its attached DDNAME, there are special logical connectors called DDNODES. EDAM has a stack of available DDNODES from which one is taken when a date set needs to be opened, and returned when the EDAM file containing that data set is closed.

When an EDAM file is opened, a DDNODE is obtained to associate the data set containing the File Control Block (FCB) with a DDNAME. An LRU DCB is then obtained and the pointer to it is placed in the DDNODE. A DDNAME, numbered sequentially by EDAM, is assigned to the data set and the number of that DDNAME is also placed in the DDNODE. The DDNAME is placed in the DCB, and the data set is dynamically allocated and opened.

The PAGE-IN, PAGE-OUT routines are the only internal functions that interact with the physical data set. They are passed the disk address of the page to be read or written, and the pointer to the FCB for the file containing the page. This disk address is broken down into a data set number and a relative record number. The data set number is used as an index into the FCB to locate the DDNODE pertaining to the specified data set. If there is no DDNODE associated with that data set, an LRU DCB must be obtained from the queue and all the processing i...