Browse Prior Art Database

String Control Byte Reduced Systems Network Architecture Data Compression

IP.com Disclosure Number: IPCOM000107568D
Original Publication Date: 1992-Mar-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 4 page(s) / 165K

Publishing Venue

IBM

Related People

Christianson, MD: AUTHOR

Abstract

SCB (String Control Byte)-reduced SNA (Systems Network Architecture) data compression improves upon the original SNA data compression algorithm in situations involving prime data strings of 189 or more bytes or non-prime data strings of 126 or more bytes, reducing the quantity of compressed data by amounts approaching 100 percent.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

String Control Byte Reduced Systems Network Architecture Data Compression

       SCB (String Control Byte)-reduced SNA (Systems Network
Architecture) data compression improves upon the original SNA data
compression algorithm in situations involving prime data strings of
189 or more bytes or non-prime data strings of 126 or more bytes,
reducing the quantity of compressed data by amounts approaching 100
percent.

      The SNA compression algorithm eliminates redundant bytes within
blocks of data.  Strings of two to 63 bytes of a predefined
character, called a prime character, are reduced to a single byte
called a prime SCB (String Control Byte).  The SCB consists of a
two-bit identification code and a six-bit byte count.  Strings of
three to 63 non-prime characters are compressed to two bytes
consisting of an SCB followed by the non-prime character.  Strings of
one to 63 non-compressible, or mixed, data bytes are actually
expanded by one byte, the SCB, which is added to the front end of the
data string.

      Whenever a string of prime or non-prime data exceeds 125 bytes
in length, any 63-byte SCBs following the string's initial SCB can be
recompressed into an appropriate number of 'fully-extended SCBs'
('fully-extended' refers to the fact that only 63-byte, or full, SCBs
are being recompressed and that they are an extension of the string's
initial SCB). In other words, following an initial 63-byte prime or
non-prime SCB, up to 63 subsequent similar 63-byte SCBs can be
compressed into one 'fully-extended SCB'.  In cases involving very
long strings of the same prime or non-prime character, the amount of
compressed data bytes can be decreased by nearly 100 percent.

      Each SCB has a two-bit identification code:  '00'B corresponds
to mixed data, '10'B corresponds to prime data, '11'B corresponds to
non-prime data, and '01'B is usually considered an invalid SCB
identifier (or '01'B may indicate compaction, a function which
generally is not implemented, or it could indicate use of the
'extended non-prime SCB'). This last identifier, '01'B, could be used
as the identifier for the fully-extended SCB.  Its presence would
indicate that the data character which was associated with the
previous prime or non-prime SCB applies to it, the fully-extended
SCB, as well.  The remaining six bits of the fully-extended SCB would
be used to indicate the number of 63-byte SCBs it is replacing (or,
in other words, the 'number of 63- byte SCBs' times 63 bytes/SCB =
total number of bytes represented by a fully-extended SCB).

      The table in Fig. 1 displays compressed prime data for strings
of various minimal sizes.

      Examination of the table reveals that use of 'SCB-Reduced Data
Compression,' hereafter referred to as SRDC, provides no reduction in
the amount of compressed data for prime strings of less than 189
bytes; however, the resultant compression ratio is no worse than it
w...