Table of Contents
Advertisement
TAPE FORMATS
o
The method of recording a group
o
The tape layout (partitioning, Device Area, System Area, Data Area, and EOD Area)
o
The rules for recording mUltiple occurrences of a group in contiguous sequence
Basically, the only differences between the two formats are in the contents of groups.
This chapter describes the DDS and DDS-DC formats. The information is divided into six major topics:
o
Entities (DDS-DC format only, Section 6.2)
o
Tape Layout (Sections 6.3, 6.4, and 6.5)
o
Data Storage Organization (Sections 6.6, 6.7, and 6.8)
o
Indexing (Section 6.9)
o
Vendor Group (Section 6.10)
o
Read-After-Write (Section 6.11)
Information describing the DDS-DC extensions to the DDS format are tagged as DDS-DC only.
6.2 Entities (DDS-DC Format Only)
An entity is a recorded object that consists of an entity header followed by one or more compressed records.
The compressed records in an entity must be based on uncompressed records of equal size. All of the
compressed records in an entity depend on the same access point for decompression. An access point is the
beginning of a sequence of compressed records at which point codewords to a decompression algorithm must
start---whether or not the required data is at the beginning of the sequence. Thus, the access point mayor
may not exist at the start of the first compressed record in the entity; however, access points cannot appear
elsewhere in an entity.
Access points are required in certain cases and are important when performing SCSI SEEK or LOCATE
commands to a record within a entity. The first entity in a group must have an access point. Also, an access
point is required at the beginning of the first compressed record of an entity following a fIlemark, a setmark,
or an uncompressed record.
As long as all of the entity header and the first 8 bits of the first compressed record in the entity are in the
same group, an entity may cross groups. However, the requirement that an access point occur at the first
entity in a group means that all span points must occur within the last compressed record of an entity.
The entity header is made up of 8 uncompressed bytes as outlined in the following points:
o
Byte 1 contains the length of the entity header (08h)
o
Byte 2 specifies the compression algorithm identifier number (when an access point occurs at the
beginning of the entity) or specifies that the compressed records in the entity depend on a the access
point in the previous entity (all bits zero).
6-3
Advertisement
Table of Contents
