Browse Prior Art Database

Mechanism for Implementing Preemptive Priorities in Dual Bus Networks

IP.com Disclosure Number: IPCOM000119589D
Original Publication Date: 1991-Feb-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 5 page(s) / 218K

Publishing Venue

IBM

Related People

Bisdikian, C: AUTHOR [+2]

Abstract

Disclosed is a mechanism for implementing priorities in a dual bus network. The mechanism allows the priorities to be preemptively multiplexed among each other, thus, always guaranteeing maximum bandwidth to the highest priority traffic whenever it is requested. The disclosed mechanism takes full effect in less than one round trip delay between two neighboring active stations with different highest internal priority levels. Moreover, only a few gates are needed for the implementation and no specific additional hardware (e.g., counters, timers, etc.) is needed. It also uses the same OR-writing principle followed throughout the conception of (1). Introduction: Priorities in (1)

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 28% of the total text.

Mechanism for Implementing Preemptive Priorities in Dual Bus Networks

      Disclosed is a mechanism for implementing priorities in a
dual bus network.  The mechanism allows the priorities to be
preemptively multiplexed among each other, thus, always guaranteeing
maximum bandwidth to the highest priority traffic whenever it is
requested.  The disclosed mechanism takes full effect in less than
one round trip delay between two neighboring active stations with
different highest internal priority levels.  Moreover, only a few
gates are needed for the implementation and no specific additional
hardware (e.g., counters, timers, etc.) is needed.  It also uses the
same OR-writing principle followed throughout the conception of (1).
Introduction: Priorities in (1)

      Dual Bus topologies are becoming the basis for various proposed
high-speed Local Area Networks (LANs) and Metropolitan Area Networks
(MANs).  Implementing effective priorities in such networks has been
of special concern to the technical and standardization communities.
In this article, we present a new mechanism to implement effective
priorities in dual bus networks, such as the subnetworks using the
Distributed Queue Dual Bus (DQDB) protocol (1).

      It is stated in (1) that the specified priority mechanism "...
allows higher priority segments to claim access ahead of already
queued segments. ...". Unfortunately, this is correct only on a per
station basis and also only according to the perception that the
station has about the rest of the network (a perception that
sometimes may not reflect the real situation).

      According to the DQDB protocol, a station cannot transmit a
request on Bus B for a newly arrived segment of priority i, until it
has transmitted (on Bus A) all the previously arrived segments (of
priority i) in this station, that are intended to be transmitted on
Bus A.  It is this feature and also the natural ordering of the
stations in the network, that can result in unfair treatment of a
station downstream the network (2,3), even when the downstream
station has high priority traffic to transmit.
The Disclosed Priority Mechanism

      Let us start by stating the objectives that an effective
priority mechanism should achieve (4, 5) and, then, use these
objectives as guidelines to deduce and justify our priority
mechanism.
1.   The average access delay of the high priority traffic must be
substantially less than that of the low priority traffic.
2.   The delay characteristics of any given and fixed distribution of
high priority traffic, should be (almost) independent of the low
priority traffic.

      We agree with 5 that, in particular, objective 2 imposes more
stringent requirements on any proposed mechanism, but this objective
is also more useful, "... as it gives a more clearly defined Quality
of Service, ...".

      We have identified three factors that result in the unfair
treatment of the high priority traffic in netw...