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

Enhancing Packet Communications Efficiency

IP.com Disclosure Number: IPCOM000123503D
Original Publication Date: 1998-Dec-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 2 page(s) / 61K

Publishing Venue

IBM

Related People

Aldred, BK: AUTHOR [+3]

Abstract

Most packet communication protocols rely on a single data packet being sent by a sender to a receiver and then being subsequently acknowledged by transmission from that receiver to the sender. Protocols can be developed where a collection of packets are sent (here termed a bunch) following by acknowledgement of their collection.

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

Enhancing Packet Communications Efficiency

   Most packet communication protocols rely on a single data
packet being sent by a sender to a receiver and then being
subsequently acknowledged by transmission from that receiver to the
sender.  Protocols can be developed where a collection of packets are
sent (here termed a bunch) following by acknowledgement of their
collection.

   The number of packets sent before an acknowledgement (ack)
is requested for that bunch, is called the bunch size.  This
parameter is important as it has a direct bearing on packet storage
requirements at the sending node.  If the protocol at the sender is
one that waits for the ack before proceeding to send more packets,
then the bunch size affects the ability of a slow processor at the
receiving end to cope with the arriving bunch of packets.
Alternatively if the sender is allowed to start sending the next
bunch before the previous bunch is acknowledged, then the sender
packet buffer might overflow.  Hence if the packet size is too big,
the storage requirements at the sender may be too high and the
ability of the receiving processor to keep up may be compromised.  As
both storage and processor usage change depending on process
activity, a fixed bunch size is likely to be a poor compromise.

   The improved protocol described here allows the packet
size to be renegotiated dynamically, i.e. at any time within the
conversation.  The receiver measures the time it takes to process
each received packet and keeps a running average, calculated across
multiple packets, e.g. 0.5* (current_average+current_packet_time).
This average is recorded in the ack returned to the sender.  Th...