This chapter is from the book
This chapter is from the book
This chapter is from the book
The Network Layer
No less critical than the server infrastructure is the local and wide area network infrastructure that is deployed by a business organization—the next layer in the simplified diagram. Networks, as previously discussed, enable the interconnection and interoperation of distributed servers, and provide access to applications and data for distributed users.
LANs began to appear in the 1970s as an enabler of distributed computing. Their wide area networking (WAN) counterparts, it can be argued, date back to the earliest days of telegraphy. The world's largest WAN, the Internet, continues to make use of the core networks established by telecommunications companies for telephony services.
Whether a LAN or a WAN, networks serve the function of interconnecting "nodes"—computers, peripheral devices, storage arrays, and so forth—so that they form an organized, manageable, and secure whole. Data traverses networks in the form of analog or digital packages represented by audible tones or electrical, infrared, microwave, radio, or laser emissions. Many standardized techniques, or protocols, exist for encoding and decoding data to facilitate its transmission across the network media—the generic name given to the wire, cable, or wireless carrier used to link nodes together.
LANs are usually networks that are deployed within a corporate premise, whereas WANs are used to move data between geographically remote locations. Increasingly popular is the acronym, MAN, which refers to a metropolitan area network—a variety of WAN used to interconnect nodes within the boundaries of a city or other regional area supported by a MAN vendor.
Within the corporate premise, LANs provide the equivalent of a roadway for moving information. Servers and other devices interconnected via the LAN may exchange information in the form of file transfers, share each other's resources on a peer-to-peer basis, or they may share access to server-hosted applications. Based on the characteristics of data traffic traversing the LAN, media bandwidth (capacity) and quality of service requirements may be deduced.
Many organizations deploy high-capacity networks, such as Gigabit Ethernet (with its 1,000 bits per second throughput), as "corporate backbone" networks, and utilize lesser capacity networks (Fast Ethernet at 100 bits per second or 10BaseT Ethernet at 10 bits per second) to interconnect nodes within a department or work group. This is roughly analogous to the idea of providing a six-lane superhighway (the enterprise backbone network) to which four-lane highways (departmental networks) and/or two-lane streets (work group networks) interconnect.
Of course, there are other networks besides Ethernet, but Ethernet—developed by Xerox Corporation, Digital Equipment Corporation, and Intel Corporation in 1976 and ratified as a standard by the Institute of Electrical and Electronic Engineers (IEEE)—enjoys the most widespread adoption within modern corporations. Ethernet so eclipsed its most popular LAN rival, Token Ring, that numerous analysts have discontinued the tracking and analysis of the Token Ring market completely. One analyst's rationale for this move was that Ethernet had become ubiquitous. In 1999, according to analysts, Ethernet accounted for 98% of the switch and hub ports shipped by vendors and 90% of the revenues within the LAN.
Ethernet supports a broad range of media options, including numerous wire and cabling standards and radio frequencies. It can be used to enable both peer-to-peer data and resource sharing, when deployed in conjunction with protocols such as NetBIOS Extended User Interface (NetBEUI) from IBM, and also more robust Internet working communications, when used with protocols such as the Novell-inspired Internetwork Packet Exchange (IPX) or the Internet Engineering Task Force's (IETF) Internet protocol suite. This suite of 20-odd protocols is typically known by the acronyms for the two best-known core protocols: Transmission Control Protocol/Internet Protocol or TCP/IP.
TCP/IP, which has the distinction of being the official protocol of the Internet, is a set of communications protocols specifically designed for Internet-working heterogeneous networks. The protocols—which were developed under the aegis of the U.S. Department of Defense (DOD) Advanced Research Projects Agency (DARPA) to interconnect DOD's diverse collection of systems and networks—took hold in the business world as a means to stitch together LANs and WANs based on different network technologies, including Ethernet, Token Ring, Fiber Distributed Data Interface (FDDI), X.25, Frame Relay, Switched Multimegabit Data Service (SMDS), Integrated Services Digital Network (ISDN), and, most recently, Asynchronous Transfer Mode (ATM). The Internet protocols are widely regarded as the most proven approach to Internet working a broad range of LAN and WAN technologies.
Network and communications protocol selection are only part of network infrastructure design. Implementing the modern network typically entails the deployment of networking devices that will facilitate the efficient movement of traffic among and between departments, work groups, data centers, and external networks. Building blocks of modern networks are numerous, but four general categories of network devices are commonly used, as summarized in Table 1–2.
Table 1–2 Network Devices*
|
Device |
Description |
|
Hubs (concentrators) |
Hubs (concentrators) are used to connect multiple users to a single physical device, which connects to the network. Hubs and concentrators act as repeaters by regenerating the communications signals as they pass through them. |
|
Bridges |
Bridges are used to logically separate network segments within the same network. |
|
Switches |
Switches are similar to bridges but usually have more ports. Switches provide a unique network segment on each port. Today, network designers are replacing hubs in their wiring closets with switches to increase their network performance and bandwidth while protecting their existing wiring investments. |
|
Routers |
Routers are used to connect different networks, directing network traffic based on network addresses rather than machine identifiers. They are protocol dependent. |
According to Cisco Systems, a leading vendor of switches and routers, network designers are moving away from bridges and concentrators and primarily using switches and routers to build "Internetworks." The design of a modern network utilizes the capabilities of routers and switches to forward data efficiently to its destination, prevent data loss, avoid bottlenecks and chokepoints that might impede performance and throughput, ensure necessary quality of service requirements as dictated by the application, and provide security and management.
LANs are increasingly connected to WANs to facilitate communications between a company headquarters and its branch offices, customers, partners, and/or suppliers. WAN gateways may be employed to establish these interconnections (TCP/IP provides an external gateway protocol for connecting to non-IP nets), or internal networks may be simply attached to the Internet or to WAN services offered by public carriers, most of which support TCP/IP connections directly.
In supporting LAN-WAN communications, the network infrastructure needs to attend to two practical matters: cost and security. Costs accrue to WAN links (and bandwidth) which are generally quite a bit greater than LAN connection costs. Considerable planning is often required to obtain exactly the right WAN facilities which will meet company needs within budget-imposed limitations.
Security is another consideration that requires close attention by network infrastructure designers. Exposing internal networks, systems, data, and applications to outsiders via a WAN is risky business. Hackers intrusions, computer virus and other malicious software, electronic espionage, and other threats are increasingly prevalent. The responses of some 273 organizations collected by the Computer Security Institute (San Francisco, California) and the Federal Bureau of Investigation's Computer Intrusion Squad as part of the 2000 Computer Crime and Security Survey revealed several alarming facts:
Ninety percent of respondents (primarily large corporations and government agencies) detected computer security breaches within the last 12 months.
Seventy percent reported a variety of serious computer security breaches other than the most common ones of computer viruses, laptop theft, or employee "net abuse"—for example, theft of proprietary information, financial fraud, system penetration from outsiders, denial of service attacks, and sabotage of data or networks.
Seventy-four percent acknowledged financial losses due to computer breaches.
Forty-two percent were willing and/or able to quantify their financial losses. The losses from these 273 respondents totaled $265,589,940 (the average annual total over the last three years was $120,240,180).
According to the Computer Security Institute, survey results illustrate that computer crime threats to large corporations and government agencies come from both inside and outside their electronic perimeters, confirming the trend in previous years. Seventy-one percent of respondents detected unauthorized access by insiders. But for the third year in a row, more respondents (59%) cited their Internet connection as a frequent point of attack than cited their internal systems as a frequent point of attack (38%).
The point is that networks frequently provide the point of entry for security threats. Thus, provisions must be made within the network for preventing unauthorized access both to the network and to the systems connected to it. Common risk reduction methods include the deployment of network firewalls, the use of data encryption, the creation of virtual private networks, and the implementation of access control and user authentication technologies.
These security measures must be balanced against other considerations, such as cost of deployment and the user-friendliness of the resulting security capability. That is to say, none of the methods are perfect shields against dedicated attacks. Thus, many organizations add intrusion detection capabilities to their prophylactic measures for intrusion prevention. Taken together, these measures constitute an important part of the infrastructure requirements of the network layer, as depicted in Figure 1–4.
){kind=link}
Figure 1–4 Network layer functions.