Chapter 3: Wired Networks
3.1: Fibre Optic Networks
The original ARPANET, the earliest form of the Internet, came online in 1969 using leased 56 Kbps telephone lines as backbones between four universities [72: 50-54]. A second network, NSFNET, was launched in the late 1970s also using 56 Kbps leased lines. NSFNET was seen as the successor to the ARPANET, and they eventually merged into one. These networks grew very rapidly and their backbones quickly became overwhelmed. The first practical fibre optic technology had just been developed in 1970, and it continued to be developed and improved through that decade. By the 1980s, a number of companies including Sprint and MCI began to install the first commercial fibre optic networks across the United States. NSFNET version 2 used 448 Kbps fibre optic channels leased from MCI as its backbone, but this was also quickly overloaded and was upgraded to 1.5 Mbps by 1990 [72: 54-56]. In 1990, the fibre backbone was upgraded to 45Mbps when ANS (Advanced Networks and Services) took over the network to form ANSNET. By 1995, ANSNET was sold to America Online and several commercial fibre networks were growing in number and speed, which today form the backbone of the modern Internet.
Fiber optic technology has been growing very quickly since its first installations in the 1980s, and it now reaches a top speed of about 10 Gbps with a maximum theoretical limit of 50 Tbps [72: 93]. Fibre optic cabling uses the Optical Carrier (OC-x) naming convention, where OC-192 identifies fibre that runs at 9.953 Gbps, the fastest available commercially [56]. That's fast enough to transfer a full-length DVD movie in under four seconds! Because fibre optic cabling is always shared between several users, however, anything near this level of performance is never achievable in practice. OC-192 uses several multiplexed SONET channels to operate at 192 times the speed of a single 51.8 Mbps OC-1 SONET channel. These super-capacity fibre channels are used as backbones between major cities and to bridge major networks, while lower-capacity fibre channels branch off from these major backbones to serve other areas.
Fibre optic cabling is composed of a thin glass core, with its thickness ![Figure 4: Fibre optic cable [30]](figures/4.jpg)
measured in microns. A protective cladding coats this glass core to prevent breakage or excessive bending, and a plastic jacket forms the outside layer of the cable. The nodes on either end of a length of fibre will have an emitter, a detector, or both. An emitter is a light-emitting diode (LED) or a laser which transmits an optical pulse along the fibre. A detector will detect an incoming pulse on the fibre and either pass it along as an optical pulse to an emitter, or will convert it to an electrical pulse. The presence of light on the fibre indicates a binary 1, while the absence of light indicates a binary 0, and these pulses of light travel along the fibre at nearly the speed of light. Networking devices like repeaters, optical amplifiers, switches and multiplexers connect these lengths of fibre together to form the backbone and lower-level pipelines. Repeaters or amplifiers are needed only about every 50 to 140 km on long lines [72: 99], while switches and multiplexers are used at junctions and end points in the network.
Deployments of fibre are usually limited to urban areas or the routes between urban areas. They often follow major highways or rail lines and can![Figure 5: Fibre Switch/Multiplexer SAP [35]](figures/5.jpg)
be installed on poles, in trenches, under water, or in underground wiring conduit. The cost to run 1 kilometre of fibre is estimated to be between $10,000 and $20,000 installed [52], which explains why fibre is rarely found in rural areas. Unless the cost to deploy fibre comes down significantly or the government heavily subsidizes such projects, it's not likely to be profitable in areas with a low population density. When a fibre pipeline is run into a community, it is terminated by a Service Access Point (SAP), which is usually a fibre switch. Companies can then lease the use of some of the bandwidth provided by that SAP for their own private use, or to resell the service as an Internet Service Provider. The existence of a broadband SAP in a community is a prerequisite for DSL, Cable, or any other kind of broadband Internet service to be offered in that area.
A new technology that is already quite popular in parts of Asia and the United States is FTTP, or Fibre to the Premises. Fibre nodes with a capacity of up to 500 homes are installed in neighbourhoods, and fibre is run from the node to each home, each with a capacity of 100 Mbps [78]. This single medium can carry telephone, Internet and television services into the premises, and it is seen as the future of Internetworking. Due to the high cost of deployment, however, it is not likely to be feasible for rural areas for many, many years.
© Jake Cormier, 2006 [jake (at) stormcloudstudios.com]
Completed as a partial requirement for the degree of Bachelor of Science (specialized)
Department of Computer Science :: Algoma University College :: Sault Ste. Marie, Ontario :: Spring 2006