The complex communications infrastructure of the Internet consists of its hardware components and a system of software layers that control various aspects of the architecture. While the hardware can often be used to support other software systems, it is the design and the rigorous standardization process of the software architecture that characterizes the Internet and provides the foundation for its scalability and success. The responsibility for the architectural design of the Internet software systems has been delegated to the Internet Engineering Task Force (IETF).[8] The IETF conducts standard-setting work groups, open to any individual, about the various aspects of Internet architecture. Resulting discussions and final standards are published in a series of publications, called Request for Comments (RFCs), freely available on the IETF web site. The principal methods of networking that enable the Internet are contained in specially designated RFCs that constitute the Internet Standards.
These standards describe a framework known as the Internet Protocol Suite. This is a model architecture that divides methods into a layered system of protocols (RFC 1122, RFC 1123). The layers correspond to the environment or scope in which their services operate. At the top is the Application Layer, the space for the application-specific networking methods used in software applications, e.g., a web browser program. Below this top layer, the Transport Layer connects applications on different hosts via the network (e.g., client-server model) with appropriate data exchange methods. Underlying these layers are the core networking technologies, consisting of two layers. The Internet Layer enables computers to identify and locate each other via Internet Protocol (IP) addresses, and allows them to connect to one-another via intermediate (transit) networks. Lastly, at the bottom of the architecture, is a software layer, the Link Layer, that provides connectivity between hosts on the same local network link, such as a local area network (LAN) or a dial-up connection. The model, also known as TCP/IP, is designed to be independent of the underlying hardware which the model therefore does not concern itself with in any detail. Other models have been developed, such as the Open Systems Interconnection (OSI) model, but they are not compatible in the details of description, nor implementation, but many similarities exist and the TCP/IP protocols are usually included in the discussion of OSI networking.
The most prominent component of the Internet model is the Internet Protocol (IP) which provides addressing systems (IP addresses) for computers on the Internet. IP enables internetworking and essentially establishes the Internet itself. IP Version 4 (IPv4) is the initial version used on the first generation of the today's Internet and is still in dominant use. It was designed to address up to ~4.3 billion (109) Internet hosts. However, the explosive growth of the Internet has led to IPv4 address exhaustion which is estimated to enter its final stage in approximately 2011.[9] A new protocol version,IPv6, was developed in the mid 1990s which provides vastly larger addressing capabilities and more efficient routing of Internet traffic. IPv6 is currently in commercial deployment phase around the world and Internet address registries (RIRs) have begun to urge all resource managers to plan rapid adoption and conversion.[10]
IPv6 is not interoperable with IPv4. It essentially establishes a "parallel" version of the Internet not directly accessible with IPv4 software. This means software upgrades or translator facilities are necessary for every networking device that needs to communicate on the IPv6 Internet. Most modern computer operating systems are already converted to operate with both versions of the Internet Protocol. Network infrastructures, however, are still lagging in this development. Aside from the complex physical connections that make up its infrastructure, the Internet is facilitated by bi- or multi-lateral commercial contracts (e.g., peering agreements), and by technical specifications or protocols that describe how to exchange data over the network. Indeed, the Internet is defined by its interconnections and routing policies.
Structure
The Internet structure and its usage characteristics have been studied extensively. It has been determined that both the Internet IP routing structure and hypertext links of the World Wide Web are examples of scale-free networks. Similar to the way the commercial Internet providers connect via Internet exchange points, research networks tend to interconnect into large subnetworks such asGEANT, GLORIAD, Internet2 (successor of the Abilene Network), and the UK's national research and education network JANET. These in turn are built around smaller networks (see also the list ofacademic computer network organizations).
Many computer scientists describe the Internet as a "prime example of a large-scale, highly engineered, yet highly complex system".[11] The Internet is extremely heterogeneous; for instance, data transfer rates and physical characteristics of connections vary widely. The Internet exhibits "emergent phenomena" that depend on its large-scale organization. For example, data transfer rates exhibit temporal self-similarity. The principles of the routing and addressing methods for traffic in the Internet reach back to their origins the 1960s when the eventual scale and popularity of the network could not be anticipated. Thus, the possibility of developing alternative structures is investigated.[12]
