Confessions of an Educational Heretic
Yüklə 12,06 Mb.
|
- Bu səhifədəki naviqasiya:
- (1) Physical Layer.
- (2) Data Link layer.
- (3) Network layer.
- (4) Transport layer.
- (5) Session Layer.
- (6) Presentation Layer. M
- Application Layer.
Distinction. The main distinction differentiating point-to-point with broadcast subnets is that point-to-point subnets utilize data transmission in a bit-by-bit method along store-and-forward pathways that bring and originate communication to its destination, while a broadcast network utilizes data transmission in a bit-by-bit method to all addressed IMPs without completion of the entire communication at any one time. That is, all IMP’s receive the first bit of information, then the second, etc., until all the information is broadcast and received. IMPs in a broadcast network may take their turn broadcasting during a preassigned time slot or whenever a situation merits a broadcast. While the first is called a static broadcast, the latter is referred to as dynamic. The dynamic subnet broadcast has the advantage of not wasting channel space during broadcast inactivity. CHAPTER VI Network Topologies Network topologies include star, ring, loop, bus, tree, mesh and fully interconnected. LANs may consist of either a combination of two or more topologies in as simple or complex a manner as a designer may incorporate or a situation may require. Each topology is briefly described below with K.C.E. Gee's summary of advantages and disadvantages from his book, Introduction to Local Computer Networks. Star. Star topologies as the name implies consist of a centralized computer from which all information is processed and fed out to peripheral devices. The plain ordinary telephone system (POTS) is an example of star topology. POTS maintains a central switching center such as a local exchange, from which all users can be reached or to which all users connect. From this centralized location other telephones can be reached through interconnection to their exchange from which the end user is connected. Extending the telephone switching concept for data requires that the network incorporate control techniques for setting up a connection, efficient transfer of data, completion of error-free transmission, breaking the connection and making it available for further communication. Thus, packet switching techniques are utilized to accomplish this task according to a standardized protocol. The term packet refers to small chunks of data which contain the sender, user, data content and error checking information. The star network topology is well suited for use within the physical plant in which a local area network exists and for which the main requirements are on-site communications. A convenient way of visualizing star topology is placing the central computer in the center such as the axle of a wheel with spokes emanating outward to devices which are attached. Advantages
Disadvantages
Ring. Ring topology is defined as the connection of each node to two and only two other nodes. In ring topology no single node has overall control over the others dictating when messages can be sent or received. In a ring, it is obvious that end-devices cannot be directed together, thus the ring requires repeaters or transceivers connected together for the end-devices to communicate with each other. While this may be a negative, it does eliminate the need for the central node or computer that is required in the star topology. This reduces complexity and thus expense, substituting instead relatively inexpensive repeaters and transceivers located at each node. Messages are thus originated by a node and regenerated by repeaters as they circulate around the ring in one direction thus simplifying broadcasting. What appears to be the greatest strength of a ring the inexpensive physical structure becomes a weakness as the topology depends upon each node to consistently function all the time in order for the network to maintain integrity. Reliability thus becomes an issue. The addition of extra nodes thus becomes by virtue of definition requiring an inoperative network for the duration of the upgrade. That is, the addition of a node requires the breaking of a ring in order to make room for the addition of the new node. Network downtime is a consequence. The addition of many new nodes can also present a problem in terms of complexity. Ring topology is best implemented at the outset of LAN design and perhaps connectors with jumpers in place at the site of possible future expansion would minimize disruption when additional nodes are required.
Disadvantages
Loop. Loop and ring topology is similar. In a loop however, control is expressed by a single node which determines what a circuit will be used for and what messages are sent along the circuit thus minimizing or eliminating the need for repeaters. Loops are used in conjunction with other topologies where the control node exists and where low transmission speeds are the norm. Loops originated from pre-local area network technology where access was required to mainframes and minicomputers. Advantages
Disadvantages
Bus. A bus consists of a discrete circuit containing two ends which along the way has devices attached. This allows for all devices along the bus to access a single transmission. Gee likens the bus to that of “normal broadcast radio transmission” where a listener may tune into a particular station by dialing in the frequency. Thus radio transmissions are broadband. A bus topology is considered broadband similar to radio transmissions, however, network radio transmissions are typically made along a cable rather than propagating through the ether (wireless). The ether is the medium through which radio waves (RF or radio frequency) propagate. A bus may also be baseband. Baseband refers to the transmission of pulses which represent the binary “1” and “0” which comprise all data. A baseband bus also has the added distinction of having connections separate from the data transmission mechanism. That is, a cable sends messages between nodes while separate access and interface peripherals are present. Advantages
Disadvantages
Tree. Tree topology incorporates bus topology with branching. Thus, it is a series of buses connected together. Since the branches may be many and of varying lengths, introducing components and adding connectors and branches creates changes on the bus that require attention. Tree shaped broadband buses such as Ethernet utilize a backbone bus (indicating that a node is in use) and a data bus. Finally, tree topology has the same advantages and disadvantages as bus topology. CHAPTER VII
There are seven layers within the structure of the modern network architecture which is referred to as the OSI protocol stack. These are labeled as follows:
The communications subnet boundary is comprised of the lowest three layers: physical, data link and network. (1) Physical Layer. This layer connects devices together physically within the network. Since data is transmitted as a stream of “1s” and “0s”, the physical layer must determine what form the “1s” and “0s” take, the length of each data bit, the voltage on the line which represents them, how the connections between devices are made and whether data is transmitted in one direction or both simultaneously, etc. The standard for the physical layer requires that the method of connection remain the standard everywhere, i.e. types of cable used, number of pins used in the connectors at the ends of the cable, the electrical characteristics of the cable itself, i.e. characteristic impedance, etc. The physical layer defines the medium which the host is using to communicate. It may include, but is not limited to, twisted wire pair, coaxial cable, fiber optic cable (FDDI), serial connections (RS232), radio frequency (RF), etc. New media are expected and as standards develop they will be incorporated within the physical layer. Standards are maintained by national and international organizations which include the International Organization for Standardization (ISO) and Electronic Industries Association (EIA). Examples of physical layer implementation include, but are not limited to, Ethernet, 10Base2, 10BaseT, 100baseT, Token Ring, Arcnet, FDDI and wireless (which includes microwaves, frequency modulation or FM, etc.). The Geneva, Switzerland based International Organization for Standardization (ISO) is a worldwide federation of standards organization. Established in 1947, ISO's aim is,
[http://www.iso.ch/infoe/intro.html] The Electronic Industries Association (EIA), established in 1927, is the primary trade organization representing high technology within the United States. Pete McCloskey, president of EIA, describes the organization's role as to, ...enhance the competitiveness of the American producer including such valuable services as, technical standards development, market analysis, government relations, trade shows and seminar programs... [http://www.eia.org/Pfm_msg.htm] RS232 is a standard interface approved by the EIA where the letters RS stand for recommended standard. Most, if not all computers, come equipped with two serial or RS232 ports as do all modems and many display monitors and mouse devices. Ethernet refers to a technology which allows information to be transmitted between computers at speeds of 10 and 100 million bits per second (MBPS). The 10-Mbps Ethernet media varieties include the original thick coaxial system, as well as thin coaxial, twisted-pair, and fiber optic systems. The most recent Ethernet standard defines the new 100-Mbps Fast Ethernet system which operates over twisted-pair and fiber optic media. [http://www.ots.utexas.edu/ethernet/100quickref/ch1qr_2.html#HEADING1] (2) Data Link layer. Since the physical layer is the actual connections within the network where raw data traverses the network's pathways without facilitation, there needs to exist a means by which the data can arrive intact as a faithful reproduction of the original. This is the task of the data link layer. This layer breaks data up into frames sending them sequentially to a destination, numbering them so that if a frame arrives with errors it may be resent and placed in the correct order upon acceptance. The data link layer processes ACKS, or acknowledgments, verifying the error free reception of the frames. This is accomplished by adding bit patterns at both the beginning and the end of the frame. Thus, while the physical layer is prone to error, being nothing more than a hardwired infrastructure of hardware, the data link layer is responsible for orchestrating the orderly and error-free transmission of the data over the physical layer. It presents the raw data error free to the next layer, the network layer. (3) Network layer. The network layer is the upper boundary of the communications subnet. This layer is responsible for converting messages from the source host, converting them into packets and assuring that they are directed toward their destination. The network layer is responsible for maintaining the flow of traffic, that is ascertaining circuit activity and taking steps so that a proliferation of packets do not interfere with each other and unduly affect throughput. This layer provides to the upper layers a means of transmitting “datagrams” over the network to a specified host. This datagram service provides no confirmation of safe delivery of the information. The transmissions are “connectionless” meaning that there is no “continuing conversation” set up between the two hosts. One datagram may have nothing to do with the next one, and indeed, may arrive out of order if they are related. This layer is also responsible for assigning addresses to the hosts and routing packets between interconnected networks. [http://www.ee.siue.edu/~rwalden/networking/network.html] The Network layer performs its functions through the implementation of packet switching within the network. Packet switching is the actual error-free routing of individual packets within the network. Known as X.25, this protocol encompasses the first three layers of the 7-layer ISO network architecture and within the network layer is called X.25 Packet Layer Protocol (PLP). A modified version of X.25 protocol is used in amateur radio networks (packet radio). Called AX.25, the protocol substitutes amateur radio call letters as routing indicators for established X.25 network addressing. The user end of any network is referred to as Data Terminating Equipment (DTE) while the carrier equipment is Data-Circuit terminating Equipment (DCE). The X.25 protocol thus allows DTE equipment over DCE to communicate with other DTE equipment. Each DTE user must possess a unique address on the network. The connections on the network can take place in one of two fashions: switched virtual circuits (SVC) or permanent virtual circuits (PVC). SVC’s are connections which take place on demand using call request packets and the address of the remote destination. The connections are initiated, data transfer complete and the connection disconnected. PVC’s are connections which are permanently in force waiting to be used. The connections are established by administration by a Packet Switched Network Administration. Thus, there is no need for a callup. The topological network experiment (below) and the BBA LAN will use PVC.
The Transport layer provides the means to establish, maintain, and release transport connections on behalf of session entities. It provides reliable end-to-end data transport. Error checking and other reliability features are handled by the protocols in the Transport layer... [http://ganges.cs.tcd.ie/4ba2/transport/5.intro1.html]
TCP stands for Transmission Control Protocol. TCP allows client PC's and UNIX servers to communicate with each other. UNIX is an operating system which is, …responsible for managing computer resources and allocating and scheduling these resources to serve the needs of individual users…With its programming tools, application programs, and networking facilities, UNIX provides a powerful, versatile computing environment. Use of UNIX is widespread, and with the advent of UNIX workstations, it continues to grow in popularity. [http://www-portfolio.stanford.edu/103841] If the network is using Novell, then TCP can be replaced by IPX/SPX. IPX/SPX is a Novell compatible protocol set. Similarly when the network is using a Windows for Workgroups or Windows NT environment, TCP/IP would be replaced by Netbios/NetBEUI. Using Netbios/NetBEUI, each station in the network is given a name and the protocol is used for the transmission of messages using addressed datagrams. A Netbios/NetBEUI name is the name of a software process or service containing 16 characters that is unique on the network. (5) Session Layer. The session layer is what the user sees when connected to the network. The process by which the user establishes a connection on another machine takes place in the session layer. Thus, a user establishing a connection with another machine on the network is taking part in a session. The session layer involves password authentication, addressing of session users and conversion of the session address to a transport layer address so that not only are users able to communicate but so may their transport station’s programs. Tanenbaum calls the “operation of setting up a session between two processes…binding.” (6) Presentation Layer. Many different terminals are in use around the world which require conversion solutions for successful network integration. An example might be a terminal at one end which uses a different character set than the terminal at the other end. While this is becoming less and less of a problem as time goes on, there still remains the need to convert settings to accommodate them. It is in the presentation layer where this conversion takes place. The presentation layer performs functions that are requested sufficiently often to warrant finding a general solution for them, rather than letting each user solve the problems. Other possibilities for presentation layer functions include text compression and encryption. It is much easier for the presentation layer to provide a secure means of data transmission, for example, than to leave its implementation up to the individual user. The International Standards Organization (ISO )model offers common abstract data forms which are associated with such distributed applications as encryption and compression. These syntax forms are known as abstract syntax notations or ASN.1. Within ASN.1 there are the basic encoding rules (BER). In actuality, data encryption can take place anywhere within the network model with the most suitable layers being the physical, transport and transportation layers. Each has consequences and there has been controversy at which layer the encryption should take place. Consider the following: When encryption is done on the physical layer, an encryption unit is inserted between each computer and the physical medium. Every bit leaving the computer is encrypted and every bit entering a computer is decrypted. This scheme is called link encryption. It is simple , but relatively inflexible. When encryption is done in the transport layer, the entire session is encrypted. A more sophisticated approach is to put it in the presentation layer, so that only those data structures or fields requiring encryption must suffer the overhead of it. [http://ganges.cs.tcd.ie/4ba2/presentation/encryption.html] Thus, it appears likely that the presentation layer will continue to be the level at which encryption and compression will take place with consequent additions to the basic encoding rules. The need for such additions will only become more and more pronounced as more and more data and transactions take place on the Internet requiring secure environments and efficiency in transmission.
CHAPTER VIII Topology Experiment In an attempt to further understand the intricacies of networking and in preparing for a network that was to be implemented at BBA, I attempted to create a small local area network in my home. It was a good way to put into practice some of the theory of construction, operation and maintenance. Operating under the assumption that the more complex LAN at BBA would be built upon the model of a small, basic LAN, I connected an older Insight 486 DX 33 MHz system running an AMD 5x86 133 MHz microprocessor (network-named TurboL) with a new Cybermax 686 PR233+ system (network-named Cybermax). The goal was to establish a LAN such that the Insight and Cybermax machines could share data, programs, resources, peripherals, printers, disk drives, SCSI devices such as the Iomega JAZ drive and graphics scanner, modem and most importantly, Internet access. The two machine LAN uses Ethernet technology (as described above) which utilizes baseband bus topology. Either machine, individually or simultaneously must have the capability for controlling the modem, making a TCP/IP connection to an Internet Service Provider (ISP) and, independently of each other, being capable of browsing the Internet, sending and receiving e-mail, file transfer protocol, etc. Yüklə 12,06 Mb. Dostları ilə paylaş:
|