Browse Prior Art Database

FAST Asynchronous Time Multiplexing Header Error Protection COMPUTATION

IP.com Disclosure Number: IPCOM000115491D
Original Publication Date: 1995-May-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 2 page(s) / 62K

Publishing Venue

IBM

Related People

Galand, C: AUTHOR [+2]

Abstract

Asynchronous Time Multiplexing (ATM) defines fixed length cells including a 48-byte data payload and a 5-byte header. The header includes a 4-byte field for routing and control, and a 1-byte field, called Header Error Protection (HEC), which is used to detect and correct transmission errors.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 54% of the total text.

FAST Asynchronous Time Multiplexing Header Error Protection COMPUTATION

      Asynchronous Time Multiplexing (ATM) defines fixed length cells
including a 48-byte data payload and a 5-byte header.  The header
includes a 4-byte field for routing and control, and a 1-byte field,
called Header Error Protection (HEC), which is used to detect and
correct transmission errors.

      The HEC is computed as the remainder of the polynomial division
of the polynomial of degree 32 representing the routing/control field
by the generating polynomial <x sup 8>+<x sup 2>+x+1 This computation
is usually performed by an implementation combining a shift-register
and three XOR gates.  The HEC field of the ATM cell is obtained by
XORing the so obtained remainder byte with the x'55' pattern, which
improves the delineation process on the line.

      Fig. 1 shows this circuit, where '+' represent XOR gates, 'D'
represent one clock cycle delay, and the input bits of the
routing/control field are denoted B(n), with n=1,...,32.

      It may be desirable to implement the computation of the HEC
function in software.  The straight implementation of the shift
register optimizes the program size at the expense of a poor cycle
performance.  A direct tabulation of all the possible HEC results
from any given 32 bit pattern would optimize the cycle performance,
but is not feasible for memory constraints.  The proposed method
allows to carry out the computation in a few cycles while requ...