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


Internet Service Providers

In the late 1980s, the Internet evolved into a collection of Internet Service Providers (or just providers) and Internet Service Subscribers (or just subscribers). A subscriber is connected to one or more providers. Providers are interconnected among themselves to provide Internet-wide connectivity among all their subscribers.

To accommodate routing in the presence of multiple interconnected organizations (i.e., providers and subscribers), the IP routing architecture models an Internet (e.g., the Internet) as a collection of interconnected routing domains. Routing within each domain is provided by means of intradomain (interior) routing protocols. The architecture allows different domains to use different intradomain routing protocols. Routing among domains is provided by means of interdomain (exterior) routing.

The IP routing architecture assumes that usually routing domains are formed along the organizational boundaries, so that different organizations (e.g., providers, subscribers) form different domains. From a routing perspective, routing information exchanged among the providers, and among the providers and their subscribers is accomplished by using interdomain routing. Within each provider or subscriber routing information among the routers is exchanged by using intradomain routing.

DEMAND FOR IMPROVEMENTS IN ADDRESSING AND ROUTING

In the early 1990s, growth of the Internet demanded significant improvements in both the scalability of the Internet routing system, as well as in the IP address space utilization. Class B network numbers were the most popular types of network numbers, because the address space provided by a single class B network seemed to be sufficient for most of the organizations that were using TCP/IP, but had more hosts that could be accommodated by a single Class C network. A combination of its popularity, and the limited number of Class B network numbers (there are only 16,384 such network numbers), however, was leading to an imminent exhaustion of the available Class B network numbers within few years.

In principle, the problem of Class B network numbers exhaustion could have been solved by assigning an organization that requires more than a single Class C network number multiple Class C network numbers. This strategy, however, while certainly providing organizations with the needed address space, was impractical for another reason—increase in the routing overhead. Because the addressing and routing architecture at that time did not support aggregation of addressing information above the level of a single network number, using multiple Class C network numbers instead of a single Class B network number would have lead to a significant growth of the addressing information in the Internet routing system.

To make things worse, the Internet routing system has been already heavily stressed at that time. Adding more overhead to the system (by assigning organizations with more than 255 hosts multiple Class C network numbers, instead of a single Class B network number) would have lead to its imminent and inevitable collapse.

Even if it would be possible to aggregate multiple Class C network numbers assigned to a particular organization into a single entry in a forwarding table, the size of the forwarding tables on routers in the Internet (and specifically on the routers that do not use default route) would have scale linearly with the number of organizations (subscribers) connected to the Internet. Although such scaling was acceptable in the Internet of late 1980s-early 1990s, it became quite clear that it would become more and more of a problem with the growth of the Internet. To accommodate the growth of the Internet, it was necessary to take steps that would make the Internet routing system scale less than linearly with the number of the subscribers connected to the Internet.

Potential exhaustion of the available Class B network numbers also raised another issue—the potential exhaustion of all available IP address space. Assigning an organization with more than 255 hosts a single Class B network number quite often lead to a fairly low address space utilization. In an extreme case, an organization with 256 hosts would have been given 65,536 addresses (a single Class B network number), resulting in less than 0.4% address space utilization.

Use of multiple Class C networks would have certainly improve address space utilization, but the amount of the IP address space associated with all the Class C network numbers is just 1/8 of the total IP unicast address space. The amount of the available IP address space associated with all the Class A network numbers is 1/2 of the total IP address space, however tapping into this portion of the address space required changes to the routing protocols and address allocation procedures, as at that time interdomain routing protocols (e.g., EGP-2, BGP2, or BGP3) did not support routing to subnets, and address allocation was done at the level of network numbers.

To summarize, in the early 1990s the Internet routing and addressing was facing three interrelated problems: potential exhaustion of Class B network numbers, potential overload of the Internet routing system, and potential exhaustion of the available IP address space.

INTRODUCTION OF CLASSLESS INTER-DOMAIN ROUTING (CIDR)

In the early 1990s, classful structure of IP address space and associated with it classful routing turned out to be no longer adequate to deal with the address allocation and routing in the Internet. During the 1992-1993 period, the Internet adopted Classless Inter-Domain Routing (CIDR). The most immediate objectives of CIDR were to solve the problem of potential Class B network number exhaustion, and to significantly improve scaling properties of the Internet routing system. At the same time CIDR also aimed at improving IP address space utilization, thereby deferring the potential exhaustion of the available IP address space. CIDR encompasses new structure of IP addresses, new address allocation architecture, and new routing protocols.


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.