Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Fragment Recovery When Implementing a Datagram Fragmentation Scheme

IP.com Disclosure Number: IPCOM000113697D
Original Publication Date: 1994-Sep-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 2 page(s) / 56K

Publishing Venue

IBM

Related People

McIntyer, MS: AUTHOR [+2]

Abstract

When a communications protocol needs to send connectionless data across a network it usually sets an initial size packet size that fits within the physical limitations of the underlying network. The protocol must also devise a method to break up larger messages on one side and reassemble the pieces on the other side. This process is known as datagram Fragmentation/Reassembly. Due to the non-guaranteed nature of datagrams, when one fragmented piece of a datagram is lost the entire datagram must be discarded. Hence, the communication protocol must deal with fragmented pieces of datagrams. This includes freeing fragments that have arrived and belong to a datagram that has missing fragments.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 57% of the total text.

Fragment Recovery When Implementing a Datagram Fragmentation Scheme

      When a communications protocol needs to send connectionless
data across a network it usually sets an initial size packet size
that fits within the physical limitations of the underlying network.
The protocol must also devise a method to break up larger messages on
one side and reassemble the pieces on the other side.  This process
is known as datagram Fragmentation/Reassembly.  Due to the
non-guaranteed nature of datagrams, when one fragmented piece of a
datagram is lost the entire datagram must be discarded.  Hence, the
communication protocol must deal with fragmented pieces of datagrams.
This includes freeing fragments that have arrived and belong to a
datagram that has missing fragments.  The problem addressed is - How
does the protocol efficiently decide when to check for potentially
freeable memory associated with datagrams? It is efficient to
minimize this type of checking for memory that can be returned to the
system.  Overchecking just wastes CPU time and system resources.

      A solution to this problem involves timestamping each
fragmented piece of a datagram as it arrives.  It is necessary to
compare a correlator value contained in each incoming fragment with
the correlator of each group of fragments already received to
determine in which datagram the incoming fragment fits.  Since we
have already accessed the fragment comparing correlators, a simple
check of the timestamp...