Browse Prior Art Database

Architecture for High Performance Transparent Bridges

IP.com Disclosure Number: IPCOM000108987D
Original Publication Date: 1992-Jul-01
Included in the Prior Art Database: 2005-Mar-23
Document File: 7 page(s) / 358K

Publishing Venue

IBM

Related People

Serpanos, D: AUTHOR [+3]

Abstract

The High Performance Transparent Bridge (HPTB) is an architecture for implementing multiport bridges interconnecting gigabit speed networks. The provision of specialized hardware support, coupled with the proper partitioning of bridging protocol entities, enable the processing of frames at very high rates. The architecture allows the concurrent bridging of asynchronous, synchronous and isochronous traffic among heterogeneous networks. Both traditional frame-based networks, such as LANs, and cell-based networks, using ATM technology, are supported and can be interconnected transparently through the HPTB.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 18% of the total text.

Architecture for High Performance Transparent Bridges

       The  High Performance Transparent Bridge (HPTB) is an
architecture for implementing multiport bridges interconnecting
gigabit speed networks.  The provision of specialized hardware
support, coupled with the proper partitioning of bridging protocol
entities, enable the processing of frames at very high rates.  The
architecture allows the concurrent bridging of asynchronous,
synchronous and isochronous traffic among heterogeneous networks.
Both traditional frame-based networks, such as LANs, and cell-based
networks, using ATM technology, are supported and can be
interconnected transparently through the HPTB.

      The current standard procedure for LAN bridging is referred to
as transparent bridging (1).  According to this procedure, the bridge
should examine the header of every frame flowing through every
network attached to it. Based on the destination address of the
frame, the bridge should be able to decide whether that frame is
destined to a node residing on the same side through which it is
received (in which case, it will be ignored) or the destination is
reachable through another port (in which case, it is received
(copied) by the bridge and forwarded to the proper output port to
continue its journey to its final destination).  In this procedure
also, a bridge has to learn the ports through which a certain node
can be reached.  This is done by monitoring the source addresses of
the frames received through each port.

      Moreover, emerging standards and application requirements are
introducing new types of traffic that need to be relayed by future
bridges.  The isochronous service is the most crucial among these
because of its very little tolerance to delay and, especially,
jitter.  The HPTB described in this article is an attempt to address
these issues.  A more detailed description of the components of the
HPTB is contained in (2).  The example described there is based on
the design of a bridge capable of interconnecting up to seven high
bandwidth networks.  Each network  is  expected to run at rates
reaching 800 Mbps (622 Mbps for ATM cell-based networks).  Slower
speed networks, such as FDDI, can be grouped and attached to a single
port of the HPTB through special network attachment units, thus
forming a tree structure within the bridge.  Isochronous traffic from
several ports can be handled with less than 3 microsecond delay, if
its aggregate rate does not exceed 622 Mbps.
Basic Components

      The HPTB can be decomposed into four main functional
components:
1.   Network Attachment Units (NAUs):
      Each NAU provides connectivity to the network attached to it
and contains the MAC interface to that network. Therefore, an NAU has
to be developed for each network type to which the HPTB has to be
connected.  On the other hand, each NAU has to interface with the
rest of the HPTB through a common NAU interface.  This interface is...