Sei sulla pagina 1di 4

CHAPTER 9

T 17A LOCAL AREA NETWORK


A Local Area Network (LAN) is a system for interconnecting data communicating components within a relatively confined space. A system is a group of inter related parts with the focus on the interrelationship and LANs are concerned with methods of communication among their components, which are mutually compatible. LANs are most commonly contained within one building but may spread to contiguous building. LANs are privately owned and so are not subject to regulation by networks. This fact, coupled with the simple, symmetrical topologies used in LANs, facilitates most acceptable speeds of 1 to 10 Mbps. Advantages: The major advantage of LANs is their ability to share equipment, data, and software. A variety of workstations can share common resources. While some visualizes LAN as an information transport system which provides medium to high speed connections between users in a local building others work at LAN as a group of computers connected together through adapters and cables. Applications of LAN are made available to everyone on the network. A LAN has architectural flexibility that is provided through minor changes to the system structure. The main advantages of LAN are: Information resources can be shared among processors. Each workstation having individual processor, the computing power is independent of the As the workstations are dedicated, response time is very less. The system is highly reliable. System architecture can be easily expanded. Data can be worked on concurrently by multiple processor. Software and hardware are generally transparent to users.

A LAN is composed of four basic elements. A transmission medium A mechanism for control of transmission over the medium An interface to the network for the host computers and Protocols to govern the relations between process operating.

A twisted pair network cannot handle high speed transmissions. Baseband signals, where modulation of carrier is not used are generally less expensive to transmit but frequency division multiplexing cannot be used. So, only one attached device can transmit at a time. Broadband system can use FDM but requires sophisticated design and installation and is more expensive. Broadband uses cable spectrum immediately avoid some higher frequency carrier. Protocol: A protocol is rule of conduct which must be understood by all devices connected to the network. As per IEEE 802 reference, two standard protocols-carrier sensing multiple access (CSMA) and token passing are used. CSMA checks whether the line is busy and sends messages when the line is free. If a collision is detected, it waits for a random period, before transmitting again. As there is no guarantee for transmission deadlines, the protocol is difficult to use for real time application. But it is simple to implement. Token passing on the other hand, is

normally used in Ring network and electronic tokens are employed to keep the devices in an orderly transmission sequence. Transmission is possible only when the token is passed through the device. Tokens are not usable for real time applications. Both CSMA and token passing are baseband protocols. In broadband system, a channel controller regulates a different sort of traffic flow. For transmitting, a channel is chosen and the controller is informed about which device is to be called. If it is free, the controller picks up a clear channel for the transmission. Originally LAN supported three main classes of services local access capability, server sharing and the ability to develop an integrated system. With local access capability, computing resources reside in relatively few locations. The requirement is to provide users at terminals with an access. This calls for server sharing which should support high speed high volume applications. Since shared resources are used an error control protocol should be provided by the network. Presently, an increasing number of applications require simultaneous real time capabilities. With these sophisticated requirements and integrations, LANs have nowadays evolved to multiuser multitasking system. However, product must consider three critical user concerns ease of use, accounting for managerial applications and full networking compatibility. Bus Interface Unit: In LAN, a node is a computer resource connected by a link to one or more neighboring nodes. A topology connected set of nodes is a network. A Bus Interface Unit (BIU) is a LAN node to which other devices e.g., server, work stations are attached via standard communications interface. It can be treated as a package of flexible and sometimes economical hardware and software. Network control functions implemented in BIU hardware, software and firmware are usually those of the physical and data link layers. BIU translates signals for the node interface into form and format necessary for transmission over the carrier. Besides this, it also drives the message over the medium, assures error detection, and provides power and ground isolation. The BIU integrate (a) An RF modem for media access, (b) A UP based network controller, (c) A timer, (d) a status display, (d) A RAM buffer, (f) A UP based control of the user interface, (g) Serial I/O interface unit and (h) A physical standard e.g., RS 232, RS 449. Actually a BIU assumes the role of a front end computer. For security reason, BIU should be protected by a password. LAN Controller: All components of BIU except the modern are interconnected by a standard commercially available component bus compatible with the selected. The Media Access Control (MAC) adapter and the Mega bus interface appear as intelligent I/O controllers on the bus. The Boot PROM supports the down line loading of software to the RAM and dumping of RAM. It basically tests and talks to the host. RAM must be able to accommodate operating system, software processes needed for operation of the controller and sufficient buffer space. The Mega bus Interface is practically, a special requirement. A versatile MAC adapter can be designed so that either a token bus or token ring protocol can be supported by the same hardware LAN controller also contains a for executing logical link control software. LAN control software should be implemented in a high level language with structured programming. Critical Aspects: Provision of proper interconnection of workstation should satisfy organizational, procedural and system issues besides technical setting. With end users devoting attention to the implementation problems regarding the operation levels the design development and installation can have more attention from

computer specialists. The main points which should be considered when planning are:a) b) c) d) e) f) g) h) Ability to build up a flexible network No transmission line expansion cost in near future Reliability of the network Compatibility with upward software versions Security control Possibility of integrating diverse baseband LANs by using broadband network. Extensibility with access to other networks Speed of action.

Ethernet: It is one of the popular privately owned data communications system. Its geographical area is usually limited to a section of a building or the entire building. Though it is a baseband LAN, it can be implemented by broadband CATV cable. Ethernet on CATV offers the ability to use other channels for video voice and other communications with the same network wiring. There is a way to set up a small Ethernet LAN without connecting transreceivers to coaxial cable. It can be done with Interlink cluster module from Intel and plain twisted pair wiring. This module allows upto 9 Ethernet workstations without using coaxial cable or transreceiver, Normally Ethernet employ either of the two protocols-xerox Network Systems (XNS) or TCP/IP. Among the advantages of Ethernet are its easy installation and expansion facility by use of only a small number of shared active components. But it cannot make effective use of high speed transmission and at the same time, maintain effective usage of such media. Also there is no link level encrytion possible in Ethernet causing security problem. Coaxial cable is an unbalanced transmission device. So, electromagnetic compatibility between the network and adjacent equipment is more difficult but it has more bandwidth and dependability. In Ethernet system, a stations interface connects bit serially to a transceiver which is tied to the passive cable. The transceiver consists of line driver, line receiver and collision detector. Workstations may be connected to Ethernet without transceivers using Intel link cluster module also. Initially the bus configuration i.e., the wiring procedure to interconnect the PCBs of the workstation is chosen. Then the topology can be extended through repeater stations. An Ethernet controller board in watch workstation of the LAN is then needed to connect parallel multilens of the workstation to a single serial connection on transmission line. A software is required for controlling communication over Ethernet. Ethernet permits users in a network to share printer resources and so, a dedicated printer is not necessary for each PC. Ethernet software is installed on the PC designated as a print server, which has one or more printers attached for shared use. A Etherprint command sets up a link into a shared printer. Data is sent transparently across the network to the server. Comparison Between Ethernet & Arcnet Parameter Ethernet ARCnet RawBit Speed 10 Mbps 2.5Mb/s Ethernets faster Bit Rate stand in ARCnet, though slower on the face good stead in burst utilisation of the of it, gives an excellent throughput, cable. especially when the load on the However, this parameter should not network is fairly even. Also, as the be taken as it is, since the effective number of nodes on the network through put is also dependent on the increases, the throughout does not topology used, and the utilization of as drastically as it does in the case

Topology

the network. Linear Bus The linear Bus is the simplest topology. The cable consumption is also the least and conceptually it is easy to connect nodes to the to the central trunk line, through drop cables (thick Ethernet) or Tconnectors (Cheapernet) but the network is prone to failures, and diagnostics is quite hard. Ethernet The chance of one cable fault bringing down the entire network is quite high. Require accessories like repeaters, as the number of nodes increases. In addition there are restrictions on the minimum distance between nodes maximum number of repeaters between nodes, maximum number of nodes per segment, and maximum number of segments in the network. CSMA/CD This Wait & Transmit protocol has the problem of slowing down in the effective throughout when the network traffic gets constantly heavier, especially due to the addition of extra nodes.

of Ethernet. Star-wired Bus Though a cable-intensive, and therefore costly system, the ARCnet physical topology is the most reliable. All the nodes are protected and isolated from each other, and so there is virtually no chance for a node failure b-ringing the network down. Diagnostics is relatively simple. However, one has to ARCnet Conceptualize the entire network, before commencing cable laying. Requires Active & Passive Hubs. Active Hubs also act as signal boosters. Relatively less number of connection rules, except that after a passive hub, there can only be a node.

Parameter Accessories

Protocol

Max. Limits

Terminator

Token Passing A variation of the Token Ring Protocol adopted by IBM, this deterministic method, ensure that each station has almost equal chance for transmission. therefore, the degradation on increase of active nodes, is less steep. Thin Ethernet: Max segments length 255 nodes in one network 600 ft. Max. no. of nodes 30 per Maximum length of 2000 ft segment. Min. distance between between 2 ft. Active Hub & nodes Thick Ethernet: Max segments- node/hub Maximum length of 100 5 per network Max segment length- ft. between Passive Hub & 1600 ft. Max. no of nodes 100 per node/Active Hub. segment. Min. distance between trans receivers -8 ft. Max transreceiver cable length 165 ft. 50 ohms 93 Ohms.

Potrebbero piacerti anche