1.3 The Internet Architecture


The Internet as it stands today is composed of more than 60,000 constituent networks. Each network is an autonomous network, and the only things that are common across these networks is the use of a common protocol, adherence to a common addressing scheme, and a common name structure. Hence it is hard to visualize the Internet architecture as beginning somewhere or having a center. Over the past 25 years since the ARPANET became operational, the Internet has increased by factors of thousands with respect to backbone speed and by factors of millions with respect to the number of hosts connected to the network. The network can be extended at any point and offer further connectivity. In a way the network is akin to the universe itself, which is ever expanding.

Is there an Internet architecture at all? Many in the Internet community believe that there does not exist any real architecture but instead there exists a set of guidelines. The architectural goal of the Internet is connectivity; the tool for achieving this is the Internet Protocol, and the intelligence is end to end rather than being embedded in the network [RFC 1958].

Important features of the original Internet architecture are as follows :

  1. A connectionless packet-forwarding infrastructure (dumb network) that positions higher-level functionality at the edge of the network for robustness (fate sharing)

  2. Addresses that are fixed-size numerical with a simple (net, host) hierarchy

The evolution of this architecture over the last 25 years has led to the following architectural features:

  1. Subnetting, autonomous systems (AS), and the Domain Name System (DNS). These features were introduced as a direct result of the understanding obtained from deployments and as a measure to address scalability and growth.

  2. Congestion control mechanisms were introduced in the late 1980s.

  3. IP multicasting developed as a means to disseminate packets from a single node to a number of receiving nodes via an enhanced addressing scheme.

The Internet is structured in a hierarchical manner or at least can be visualized as such. However, it is not a clear hierarchy, but rather a mixed-up one. At the top there exist multiple backbone networks that connect to each other at network access points (NAPs) or exchange points (EPs). Backbone networks of large Internet service providers such as NTT/Verio, Sprint, and UUNet, which span the globe, carry the majority of the Internet traffic. Figure 1-2 shows a network map of UUNet's global network (http://uunet.com/network/maps/) as of June 2001.

Figure 1-2. An example global backbone network.

graphics/01fig02.gif

Regional networks and ISPs connect to the larger backbones, and smaller ISPs, in turn , are connected to these regional networks. End users obtain connectivity to the Internet via ISPs in general. Peering between ISPs and service agreements between ISPs allow inter-ISP traffic. A model of this hierarchical model of the Internet is shown in Figure 1-3.

Figure 1-3. Hierarchical view of the Internet.

graphics/01fig03.gif

In reality, the Internet is not organized as a strict hierarchy. Peering networks could connect to each other and at the same time have connectivity to a larger network. The large backbone ISPs could also be the ones providing end- user connectivity. This type of a model is feasible when you have a relatively small number of ISPs, but as the Internet continues to grow in size and the number of users increases , so do usage and the number of Internet service providers. As a result, ISPs will need to continue to exchange increasing amounts of data traffic. One of the basic underlying mechanisms in the Internet today is the points where various ISPs exchange traffic. These interconnection points are called NAPs or EPs. Exchange points can be defined as multiple-access networks allowing ISPs to exchange traffic and routing information with other ISPs. There are many exchange points across continents today. Some of the larger ones are Sprint NAP, MAE (metropolitan area exchange) East and West, and LINX (London Internet Exchange). A view of Internet architecture taking NAPs/EPs into consideration is shown in Figure 1-4.

Figure 1-4. Current view with exchange points.

graphics/01fig04.gif

The underlying telecommunications and transmission infrastructure that forms the Internet has grown significantly in the last few years. As the processing power of central processing unit (CPUs) of computers and handheld devices keeps increasing, so does the bandwidth of the networks.



IP in Wireless Networks
IP in Wireless Networks
ISBN: 0130666483
EAN: 2147483647
Year: 2003
Pages: 164

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net