Basic direct access method

From Wikipedia, the free encyclopedia
  (Redirected from Basic Direct Access Method)
Jump to: navigation, search

Basic Direct Access Method, or BDAM is an access method for IBM's OS/360 and successors computer operating systems on System/360 and later mainframes. BDAM "consists of routines used in retrieving data from, and storing data onto, direct access devices."[1]:p.5 BDAM is available on OS/360, OS/VS2, MVS, z/OS, and related high-end operating systems.

Description[edit]

Basic, in IBM terminology, indicates a lower-level access method that deals with data sets in terms of physical blocks and does not provide either read-ahead, called anticipatory buffering, or synchronization — that is, the user program has to explicitly wait for completion of each input/output event.[2]:p.31

With BDAM, "the programmer can directly store and retrieve a block by specifying either its actual device address, its relative position within a data set (relative block number), or the relative track within a data set at which the system is to begin a search.[2]:p.63

BDAM provides no index or structure to the file except as programmed by the application. In many applications, a hashing or randomizing function may be used to assign the block address based on a key in the data. If physical keys are used, the key of the last record within the block must be written as the key for that block.

The BDAM application program interface can be compared with the interface offered by open, read, write and close calls (using file handles) in other operating systems such as Unix and Windows.

BDAM is still supported by IBM as of 2012. Because of its dependence on physical device geometry new IBM direct-access devices used with z/OS emulate IBM 3390 devices regardless of their real physical characteristics.

Details[edit]

BDAM requires that the file be allocated with DCB=DSORG=DA (direct access dataset organization) and SPACE=(,,CONTIG) (the file must be of fixed size with contiguous space). The file is opened output and all blocks are written sequentially to preformat the disk and optionally to load the data before the file can be used.[3]

The programmer used the READ or WRITE macro instructions to access data, providing the block address and/or key. It was possible to start multiple input/output operations to run concurrently. The CHECK macro had to be used to wait for completion of a specific operation before the data could be accessed or the data buffer reused.

Disk addresses[edit]

BDAM accepts disk addresses as either actual device addresses, as relative track addresses, or as relative block numbers.

Device addresses are eight byte fields in the form MBBCCHHR. All subfields are binary numbers.[4]

  • M is the extent number for files with more than one extent — non-contiguous files.
  • BB is zero. Previously this indicated the bin on an IBM 2321 Data Cell.
  • CC is the absolute cylinder number on the device.
  • HH is the track (head) number.
  • R is the record number on the track. Record zero contained control information for the track and could not be used.

Specifying device addresses would make the dataset unmovable.[2]:p.106

Relative track addresses are three byte fields in the form TTR. All subfields are binary numbers.

  • TT is the relative track number in the file, starting with zero.
  • R is the record number on the track.

Relative block numbers are four byte (fullword) binary numbers indicating the block number in the file. The first block is block 0. This form of addressing can only be used with fixed length blocks (RECFM=F).[2]:p.106

See also[edit]

References[edit]

  1. ^ IBM Corporation (1967). IBM System/360 Operating System Basic Direct Access Method. 
  2. ^ a b c d IBM Corporation (1965). IBM System/360 Operating System Data Management. 
  3. ^ CICS and BDAM, IBM infocenter
  4. ^ IBM Corporation. "Executing Direct Access Programs". Retrieved September 23, 2012. 

External links[edit]