Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!


Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


THE EMERGENCE OF CELL-BASED ROUTING AND SWITCHING

The next section discusses the various ways ATM and routing can evolve outside of the wiring closet for campus and wide area connectivity and the emergence of a new class of functionality.

An important objective is to build networks that maximize application performance and enhance service predictability through the support of various classes of service, for managing delay and throughput variations. Most Ethernet switches boast latencies in the 10 to 100 microsecond range, with pure ATM switches at the low end of this range. With ATM switching exhibiting latencies at least a hundred times better than routers, it seems obvious that latency can be minimized by integrating router functionality into switching.

A number of approaches can be used to combine routing and switching in the wide area. Routers can be interconnected over an ATM network, but this does not eliminate the latency introduced by routers. This approach does not fully exploit the attributes of switching (i.e. ATM serves as a pipe), because there is only a loose coupling between routing and the management of ATM connections.

A second approach has been proposed by a number of vendors and it relies on routing to link switched networks. In this case, routing functionality is distributed between centralized route servers and workgroup level packet forwarders. Many VCs travel across the single high-speed interface to the route server, which establishes connections among virtual and emulated LANs and also takes care of conventional tasks like firewalls and segmentation. This approach establishes a single point of failure and a potential bottleneck in the route server and could have a negative impact on network latency by introducing extra hops. Backbone traffic will increase because interdepartmental traffic needs to travel down to the router and back up.

A third approach calls for a single device that integrates conventional routing and ATM switching. This device is sometimes called a cell-based or ATM router, or an ATM enterprise network switch (ENS) because adaptation and switching of non-LAN traffic is also incorporated. An ATM ENS offers low application delay by choosing optimal switched paths and eliminating multiple networking devices between end systems. It integrates routing, address resolution, packet forwarding, and ATM connection, traffic, and quality of service management. Furthermore, it can consolidate the full range of wide area connectivity needs including voice and video.

The remainder of this chapter discusses the interLAN switching capabilities of enterprise network switches in a number of key application environments. The key values of interLAN switching are:

  Multiple application-based qualities of service.
  High-performance cell-based routing and switching.
  Scalable networking.
  Network simplification.

ENTERPRISE NETWORK SWITCHING

An effective enterprise network switching tool supports interLAN switching and ATM adaptation and switching on a single, high-speed platform. InterLAN switching is an implementation of cell-based routing and switching in an ATM enterprise network switch. InterLAN switching is integrated with its capability of consolidating voice, data, video, and ATM on a single platform.

Multiple Application-based Classes of Service

An ENS supports multiple application-based qualities of service to provide more predictable application performance and to optimize the use of network resources. The full suite of ATM classes of service should be supported by an ENS, including constant bit rate (CBR), variable bit rate (both real-time and non-real- time VBR), unspecified bit rate (UBR) and available bit rate (ABR, initially using end-to-end signaling and forward-congestion signaling). Classes of service are based on a queuing structure that supports three hardware emission priorities and four discard priorities. In addition to these mechanisms, a logical network topology can be configured that rides on top of the physical wide area network topology. A multicast facility is also provided to broadcast packets to all members of a logical network.

This logical network capability provides the following benefits:

  Bandwidth effectiveness, because trunk bandwidth can be pre-allocated to certain applications to provide guaranteed minimum performance, while taking full advantage of statistical sharing for peak loads.
  Mapping of different classes of service to different logical networks.
  Automatic reconfiguration without affecting users because the physical structure of the WAN is decoupled from the logical structure.
  Provision of statistics for the bill back of end users and applications.

High-Performance Cell-based Routing and Switching

An appropriately configured cell-based routing and switching platform can deliver high-performance operation at the switch and network levels. In addition to the standard LAN and interLAN services, such a network optionally provides a distributed switching and routing system that supports the low-latency transmission of interLAN switched packets across a network. An internal address structure is used to support very high-speed, low-latency forwarding at transit nodes. A particular network protocol only has to be supported at the edge of the network.

Scalable Networking

Cell-based routing and switching offers scalable networking, providing an integrated solution from the campus to potentially very large global networks operating over narrowband to broadband facilities. Cell-based routing and switching platform vendor offerings are available in five and 16-slot versions to economically serve different site sizes. In addition, interfaces operating at speeds from 56K bps to 155M bps can be provisioned on either the user interface or network sides on a plug and play basis. At speeds up to T3, such a network supports frame and frame/cell trunking, which dramatically reduces the overheads associated with ATM adaptation.

InterLAN switching can offer switch capacity, simplified internal addressing, OSPF-based routing, and hierarchical network support, which are all enablers for building very large networks. It can also provide an evolution path to ATM by supporting a separation between routing and high-speed packet forwarding functionality. Routing is the process of finding paths through a network made up of interconnected nodes and links, and forwarding is the process of physically passing user data through a network node based on the topology information gathered by the routing protocol.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.