Now that you are aware of some of the terminology associated with PSTN, let's look at its operation and how that affects wide area networking.
You, as a subscriber to telephone services, are responsible for supplying your own phone equipment and an appropriate place that a phone company can connect to. In "telephonese", your equipment, be it a single line or whole phone system, is referred to as CPE (Customer Provided Equipment). The point where the phone company hooks up to your system is known as the demarcation point or demarc as explained in the terminology section.
The phone company, in attaching its wiring to yours, completes what is known as a local loop (also previously defined). This loop is simply a single circuit hooked back to the phone company's central office (or CO). The CO contains switches that can connect your line to anyone else's line.
COs themselves are connected to other COs via lines called trunks. Trunks can carry large amounts of information, whether voice or data, between COs. An interesting tidbit is that trunks can occasionally get overloaded which means callers cannot get an available line on them. The result is a "busy" signal, but one that pulses at a faster rate than a normal busy signal. This fast-pulsing sound is referred to as a "trunk busy".
Trunks may link callers from within a single CO or from multiple COs. The COs themselves are responsible for creating the ringing on your telephone. When your call traverses from one CO to another, the ringing you hear is not synchronized with the ringing that the person you are calling hears. As a result, someone may pick up the phone to answer even before you hear it ringing in your handset. The trunk lines connecting COs may have differing technology when it comes to putting several conversations at once on them. Some just contain circuits that allow you to hear what someone else is saying while you are talking. Others use sophisticated means of taking signals from several callers, chopping them up into bits and pieces, then interspersing the pieces of several calls on a single line. On the other end of the line, these bits and pieces are reassembled into coherent conversations for several people. More on that aspect in a later chapter.
Figure 4-1: A temporary path can be created from your phone through a CO, its trunk system, and another CO to reach your intended receiver.
COs exist within certain regions called LATAs defined previously. LATAs are used in determining who gets to charge a caller. More than one CO may be located in a LATA. LATAs were created as the Federal Communications Commission broke up American Telephone and Telegraph in 1984. The long-distance companies like AT&T, US Sprint and MCI determine rates for calls from one LATA to another. If a call is placed inside a LATA to another point inside the same LATA, then local companies determine the rate and classification of the call. The local company may be one of several Bell Operating Companies (BOCs) that also resulted from the break up of AT&T in 1984. Incidentally, regulations also provide that each LATA has a CO that has a POP or Point Of Presence for each long distance carrier. That way, consumers may choose any long distance carrier that they would like. Long distance calls are simple routed to the CO that has the POP for the customer's long distance carrier.
In a nutshell, the PSTN is composed of several layers. Subscriber equipment composes the first layer. It connects to a local loop attached to a CO which is the second layer. Next the trunk lines connecting the switching stations comprise a third layer. Finally, the fourth layer is the long distance companies who govern what is known as inter-LATA calls.
2. Wide Area Networking With Telephone Lines
From a wide area networking point of view, the above information is important for a number of reasons. For one, someone intent on using standard phone lines for digital communications must be aware of the limitations of such lines. The filtering processes that take place limit the effective bandwidth of voice lines considerably. That is why callers you speak with often sound "tinny". This filtering reduces the bandwidth (and consequently the speed) of data transmissions as well. When purchasing a modem, you may select one that operates at 9600 baud (that's bits per second). This speed is pushing the upper limits of what the limited bandwidth of standard phone services will allow. However, the actual throughput of data is often increased by using what is known as data compression. That means that more data is squeezed in over the same limited bandwidth.
AT&T's "True Voice" technology is a sound enhancement feature that boosts the bass frequencies of sounds that are typically zapped by the limited bandwidth and increases the volume of a call by 4 decibels. This does not improve the quality of lines for data transfer but does create the impression that a caller's voice sounds closer and more natural.
Secondly, voice-grade lines are not known for their quality. Telephone lines are susceptible to various types of electromagnetic interference or other natural (or man-made) disturbances. However, all-digital lines especially designed for data transfer are also susceptible to some problems. The good news is that the telephone companies do make an honest effort in most cases to keep their lines clean. A factor here is that when a number is dialed, a temporary pathway is set up through the phone company entities involved. By their very natures, temporary connections tend to be of a lower quality and more susceptible to problems than permanent connections. Also, since a phone call is really the creation of a point-to-point link for the caller by placing in several smaller links along the way, any single problem link can create poor quality for the whole. Since the same arrangement of links may not exist each time you call a particular number, some connections may appear to be very good while other seem very bad even though you have dialed the same number.
Thirdly, there is the cost factor. Standard voice-grade telephone lines are obviously going to be cheaper than dedicated digital-grade lines. So choosing standard lines may seem the best choice. However digital lines offer increased bandwidth by removing filtering devices. This means increase speed of data transfers. A common digital line, called a T-1 link, allows data throughput at around 1.5 million bits per second. This stands in stark contrast to the 64 thousand bits per second capability on voice-grade lines using maximum data compression.
Figure 4-2: Modems will allow wide area communication, but throughput will be inferior to digital leased lines like T-1.
As the demands of the digital world have increased along with the proliferation of digital technology, phone service providers have created a host of digital solutions. Faster lines are available to accommodate higher data throughput. Dedicated digital lines, if used for voice transmissions, could generally accommodate a large number of them. But in the case of digital lines, the entire bandwidth is reserved for data.
Standard Digital Lines
| Line Type | Data Rate | Voice Channels |
| Voice-Grade | 64 Kbps | 1 |
| T-1 | 1.54 Mbps | 24 |
| T-1C | 3.15 Mbps | 48 |
| T-2 | 6.31 Mbps | 96 |
| T-3 | 44.73 Mbps | 672 |
| T-4 | 274.18 Mbps | 4032 |
As you can see the phone company can offer you a number of choices, but there are factors to consider. First, a permanently wired line is generally of better and more manageable quality then the temporary lines of standard usage. Additionally, the most optimal choice according to cost is achieved when purchasing digital lines that stand to be used frequently and at all hours. Obviously, spending a great deal of money on lines that are only used from 8 a.m. to 5 p.m. is not utilizing your resources as efficiently as possible. Your expensive leased digital link lays idle for 15 hours. Some batch processing may be scheduled over night.
As mentioned earlier, advances in the digital realm are driving the phone service vendors to improve their data handling capacity and offerings. As a result a new standard has been developed called the Integrated Systems Digital Network or ISDN.
3. Integrated Services Digital Network
This newer technology is called ISDN for short. ISDN features tiers of services offered to companies or individuals that include both digital lines and voice-grade lines. The consumer may select from any one of these tiers depending on what the need is. All the data whether network communications or voice data is digital.
All data is multiplexed where several types of data may be carried a single physical wire. ISDN addresses the need to transmit and receive all sorts of data allowing the usage of voice and video mail, computer data, remote terminal input, interactive video, standard voice telephone, etc. Basically the possibilities are almost limitless. Any signal that can be placed in a digital format will be fair game for ISDN. In the future, the telephone and modem will merge yielding a single data handling unit.
ISDN consists of digital lines that are broken up into "channels". These channels support different types of data and different throughputs. The services provided by ISDN are also called "interfaces". If ISDN services are installed at a company, that organization has purchased a specific interface.
ISDN Services
ISDN Basic Rate Interface (BRI)-
Called 2B+D, "S" or "T" Interface
Supports 2 64 Kbps channels (2B stands for 2 bearer channels).
Bearer channels may carry a voice conversation or be used as a
high speed data link (64 Kbps compared with 9600 bps today).
Bearer channels may be subdivided into several lower speed data channels.
Supports 1 16 Kbps channel (D stands for 1 data channel).
Data channels carry the control information necessary to connect and tear down the voice connections on the bearer channels. This channel may carry a 9600 baud signal in addition to the necessary control information for the bearer channels. Typically it is subdivided into three sub-channels called s, t and p.
s channel - handles the signalling portion of a call controlling the setup and tearing down of a call.
t channel - used for handling special data (called telemetry). The data may be something like temperatures from remote thermometers or thermostats.
p channel - the 9600 baud digital channel.
Two unshielded twisted pairs are used for the S interface, which cannot be more than 1 kilometer from a switching station. The "T" interface provides the same services as the S interface only that it uses one unshielded twisted pair instead. A company or individual using the T interface could be located as far away as 10 kilometers from a CO.
Primary Rate Interface (PRI)-
Called 23B+D or 30B+D depending on bandwidths needed. These are the ISDN equivalents to modern T-1 lines. They may also handle 23 and 30 voice channels respectively. The 23B+D delivers throughputs of 1.544 Mbps while 30B+D delivers 2.040 Mbps. These arrangements feature separate 16 Kbps D channels for handling control information. PRI uses two twisted pairs to provide such services.
Additional channels available-
Channel A - Standard 4 KHz voice conversation line.
Channel C - 8 or 16 Kbps line for handling control information (called out-of-band signalling) - very similar to D channel.
Channel E - 64 Kbps channel devoted to ISDN signaling and data handling.
Channel H - Purely digital data channel available at speeds of 384, 1536 or 1920 Kbps.
ISDN is a radical departure from what we have traditionally considered as phone services. An important thing to consider is that the ISDN standard is being promoted globally by the Consultative Committee on International Telephony and Telegraphy (or CCITT). As the world standard-setting body, the CCITT pretty well drives development on a global scale. The advantage will be standard ISDN services available anywhere.
ISDN telephones are now pretty expensive rivaling the costs of personal computers. Their front looks peculiar because it features a small LCD screen. The phone also sports an RS-232-C connector for attaching data devices. Typically they are feature-laden. One such feature is ID tracking. Every ISDN device sends out control information along the D channel to create voice connections. That info includes the source caller's telephone number. In addition, the special D channel can transmit other data as well - like credit card information. This leads to all sorts of phone order possibilities. The home shopping networks, complete with their onscreen graphical ordering interfaces, will get extra help from an accommodating ISDN device.
Some of the possibilities with future enhanced ISDN phones include:
· Ability to restrict callers from calling you.
· Presentation of caller's phone number and other information for call screening.
· Restriction of ISDN number from being sent to person being called.
· Call waiting where incoming caller information is displayed for user and person may choose to reject, accept or forward new call.
· Callers may be transferring data while carrying on a conversation.
· Telephones may support E-Mail for unattended sets.
· Caller may specify simultaneous ringing of telephones in multiple locations.
· Caller may send data to multiple locations simultaneously.
ISDN services may utilize circuit switching, which is used today for voice lines, or it may use packet switching. Packet switching is ideal for digital data, because this sort of information is grouped into frames or packets, which are simply a collection of bytes of data. In packet switching, information inside the packets is read as to where the packet is going. Then each packet is individually routed to its destination. Since there may be more than one pathway to the same destination, packets may be routed down more than one path to the endpoint. This occurs because a determination is made for each packet as to which is the fastest pathway. Since the dynamics of network pathways are constantly changing, this results in one path being the optimal path in one instant and not the optimal path the next. Consequently the destination gets a flurry of packets from many pathways. This would pretty well confuse the destination except for one thing. The packets being received all contain sequence information as to what order they were sent in. As a result the receiver can reconstruct the original message by placing the packets in the correct sequence. ISDN will implement packet switching for long distance transfer of data. Plans call for charges to be based on the number of packets one sends.
There are a number of reasons why ISDN will yet take some time to implement:
· Standard not fully implemented around the world.
· ISDN equipment must be installed at all COs.
· Analog (or non-digital) switches far exceed the number of digital switches.
· Rural sites may not change over for many years.
· Consumers will have to purchase special telephone sets that support ISDN.
· Marketing and public relations campaigns must convince consumers of the necessity of ISDN products.
· Early ISDN equipment is still very price-formidable.
In summary, ISDN offers a lot to the digital-oriented consumer, but its implementation will take some time before coming mainstream. In the mean time, determining how to get the best performance from services offered by the phone entities can require research. The next page contains a brief summary of the options.
Digital Data Throughput Options
Option |
Performance |
Cost |
| Voice grade line with 1200/2400 baud modem | Very slow at 1.2 and 2.4 Kbps respectively | Very low. |
| Voice grade line with 9600 baud modem | Faster, but comparatively slow to other technologies with data compression yielding more performance | Still very low |
| Dedicated line services | Range from 1.2 to 64 Kbps | Moderate to high |
| ISDN Services | Offers 64 Kbps on digital bearer channel | Moderate to high depending on no. of lines requested |
| T-1 line | Yields high performance when compared with voice grade lines - 1.5 Mbps | High |
| T-1C line | Yields higher performance yet with speeds reaching up to 3.15 Mbps | High |
| T-2 line | Higher performance still with throughputs of 6.3 Mbps | Very high |
| T-3 line | Very high throughput at 44.7 Mbps - requires fiber optic cable | Very high |
| T-4 line | Extremely high throughput as wide area links go - over 274 Mbps - uses coax or fiber | Extremely high |
The pricing of the above levels of service are so variable from region to region that a practical comparison would be difficult. Be aware that there are three tiers of tariffs that affect the lines - local, state and interstate. In most instances, interstate tariffs levied by the FCC are the most costly. The least expensive dedicated digital lines are those that utilize only one CO.
In general, the Public Telephone System (PSTN) provides a pre-existing network ideal for linking wide area nodes together. The cost of such linkage has to be weighed against throughput speeds. Quite often, the phone system provides a more economical choice against other wide area networking options to be discussed in this chapter. The obvious migration in the future for the telephone services is digital. The proliferation if ISDN service is bound to occur, faster in some places, slower in others. Tennessee, the home of Atrium Learning Center, has proven to be one of the most aggressive states in the US with full ISDN services available statewide during 1994. Other states may not have the funding or backers of such a progressive implementation. Whatever the case, the increased usage of ISDN services will yield better wide area networking access for everyone.
Microwave Technology
Microwaves lie on the electromagnetic spectrum between radio waves and light waves. As a means of carrying data, microwaves offer advantages because their high frequency is less affected by atmospheric conditions. Technically speaking, microwaves lie between 890 MHz (that's 890 million waves per second) to 5 GHz (or 5 billion waves per second). This type of transmission is used for both point-to-point connections over land or water (called terrestrial) or satellite communications, and is known for being "line of sight". This simply means that the microwave transmitter and receiver have to be aimed at each other.
Figure 4-3: Microwave technology is used to transmit data over satellite links.
Microwave technology that is earth-based allows us to get around restrictions that might be placed on physical cabling. For instance, if you were needing to send data over a cavernous gorge with a raging river, you might find laying cable a bit perilous. By setting up special microwave equipment, you could just beam the signal from one point to another, potentially saving both money and life. A common use for microwave technology is beaming network data between buildings where installing cable between them would be tough or more expensive.
Microwaves are also used for satellite technology. Since microwaves aren't strongly affected by atmospheric conditions, they constitute an excellent choice for broadcasting to satellites and back. Satellites themselves are extremely expensive, so large companies often pool their resources together and jointly purchase them. Then satellite resources are leased out.
From a security point of view, microwaves may not be that safe. Anyone may intercept, jam or in some cases, supersede a microwave signal. Millions of Home Box Office(TM) viewers were surprised to see a message appear on the screen one evening from a techno-wizard who beamed his on signal to the satellite carrying this popular pay channel. The little prank showed a little weakness that earth-to-satellite transmissions have - that is susceptibility.
From a reliability point of view, microwave technology fares well. Most earth-based systems beam a signal up to about 30 miles before the signal has to be repeated via a microwave repeater station. This station merely receives the signal and regenerates it. Then the signal is re-broadcast to the next receiving station. Generally the signals are not affected much by atmospheric conditions though rain and fog will occasionally affect them.
Satellite microwave technology uses frequencies less susceptible to atmospheric conditions. Communications satellites use what is known as a geosynchronous orbit. That means they orbit the earth at the same rate that the earth is spinning on its axis. What effectively happens is that the satellite stays in one place above the earth. This is an interesting feat because gravity is constantly pulling on the satellite, so it has to orbit at a certain speed to maintain freedom from the gravity that would pull it to its fiery destruction. If the satellite orbits too fast, it will not be geosynchronous and it will move further from the earth. If it orbits too slowly, it will not be geosynchronous and it will fall to the earth. Scientists have learned the optimal geosynchronous orbit lies 22,300 miles from the earth's surface. That means a microwave signal must travel this distance to reach the satellite and a signal must travel the same distance coming back to your remote site on the earth.
As you can imagine, this great distance interjects a time delay into the whole transmission and reception scheme. You have probably experienced this first hand if you have ever called overseas via a satellite link. You might hear your own voice echoing back to you when you say something. This is a problem associated with satellite technology that affects not just voice but digital data as well. Fortunately this problem is alleviated through the use of "echo suppression circuitry".
An advantage of microwave technology is the ability to receive transmissions from portable receivers. In terrestrial systems, this means the transmitter and receiver should be aimed, but once accomplished, can yield a good signal. In satellite technology, the portable unit must be aimed at the satellite. For security purposes, the beams of some satellites are "narrow beams" limiting access to certain geographical parts of the world. Satellites may also use "wide beam" transmission. This microwave carried message can be received on land, sea, or in the air.
Use of this technology requires FCC licensing in the United States and foreign licensing in other countries. This process will add some time to using a microwave system, so one should plan on it as a part of any implementation program. In addition, all microwave equipment must be approved for the safety of the users and to avoid violation of frequency guidelines. Oh, and one other note: Don't place body parts on or near a microwave transmission apparatus. Cooked limbs are useless limbs.
Laser Technology
Light has a much higher frequency than microwave does. Higher frequencies mean that more data may be carried by the wave, and light can carry a great deal of data. In laser technology a very intense beam of concentrated light is used to carry a signal. Typically, this beam is not visible to humans because the frequencies used are just below that of what we can see. This range of light is known as infrared.
Infrared light can be received by a special device known as a photodiode. A photodiode will allow a current to pass through it just as long as the device is exposed to light. When the light stops, so does the current. In the case of laser transmission, the light pulses in response to the data that is being carried by it. Technically speaking, the laser light is "modulated" by the data signal. This pulsing creates an interruption of the current moving through the photodiode. Since the light pulses in response to the data traveling with it, this pattern is replicated through the photodiode. Consequently, the exact pattern of digital data can be reconstructed.
Laser systems benefit from their superior speed of data transfer comparable to that of fiber optics. However, there are problems as well. For one thing, laser is susceptible to atmospheric conditions. It tends to diffuse rapidly when exposed to fog, rain, snow, etc. It also attenuates (or decreases in strength) over short distances. For these reasons, there are practical limits as to how far you can go with laser. Another problem that laser suffers is the fact that it is extremely directional - that is, the transmitter and receiver must be perfectly in line. Only recently one networking professional complained to this author of problems incurred with their laser system. The company was using laser between two buildings. The laser target on the receivers was five inches in diameter. However, during the day, temperature changes and the like would cause the buildings to move enough to throw the lasers off target. The wide area network link would fail and users would become aggravated. Fortunately these stories are far and few between, but be aware of laser's limitations.
Controversy in Laserland
The concept of the LASER (Light Amplification from the Stimulated Emission of Radiation) was first put forth by a Columbia University graduate student working in physics in 1957. Gordon Gould posited that light could be generated in a highly concentrated form by producing it at a particular frequency with all the tiny waves in phase (lined up together). Realizing that this technology could be useful, Gould filed a patent for lasers in 1959. The only problem was that one of his profs at Columbia had stolen the idea and, along with a gentleman from Bell Labs, had filed for the patents in 1958. Gould was crushed and took the matter into litigation. Justice won out, and 20 years after conceiving of the laser, Gould was awarded the patents. He became an instant multi-multi-millionaire as a result. One wonders if there was anything left over after the legal fees were paid.
Switching Technology
Moving data across wide areas requires that we be able to form a link from the sending point to the receiving point. As mentioned previously in this chapter, those end-to-end connections are often created from a series of links in between. The entire connection is susceptible to one failed link in a chain of links that allow our message to pass through. To create these links, switching has to take place.
Switching in this instance, means selecting a pathway that gets information to, or at least closer to, its destination, and then directing the data down that path. This switching may mean creating circuits from one point to the other much as the phone company does. Switching may mean selecting the best paths between machines that store messages. Switching may mean breaking up data messages into small units (packets) and individually directing each packet down the best path. Whatever the case, each method offers particular advantages and disadvantages.
1. Message Switching
Whenever we use the postal service, we are using a variation of message switching. In this technology, an entire message (analogous to a letter) is routed to a destination (recipient). The message will follow several paths and decisions as to the proper pathway must also be made. The first decision is yours. You must take the letter to the appropriate mailbox. Placing the letter in the "metered" mailbox for instance is a good way to slow it down if you have placed a stamp on it, so the mailbox holds the letter until the postman is ready to receive it. The postman transports the letter from the box to the post office. Hopefully he will choose a route to the post office that requires minimal delay. Now the post office holds the letter until the postal system is ready for it.
Once the letter is processed locally, a decision has to be made as the most expeditious manner of getting to a post office near its destination. The letter may go by air or truck to a receiving area that holds the letter until a postal service representative picks it up and takes it to the remote post office. Once there, it is held until it is processed and the proper route for delivering it is selected. Next the postman must get the letter and deliver it to the intended person or company where it is utilized.
Figure 4-4: A message switching network is also called a "store and forward" network. This system requires substantial memory resources and lengthy overhead.
What you have seen in this example is an example of a store and forward network. In this scheme, messages are received in their entirety and stored while a route to or closer to their destination is being determined. Once the route has been established, then the message is sent to either its final destination or another intermediate storage and forwarding point.
One of the prime users of this type of technology is electronic mail systems. In E-mail networks, all devices storing E-mail (typically called E-mail or mail servers) may not be connected all of the time. They may only connect with one another intermittently. When you think about it, that's not a bad idea. Many E-mail messages are not urgent, and paying for a dedicated line between servers could be expensive. E-mail servers can periodically dial each other up and upload (transmit) whatever messages need to be sent. In fact, the dialing process could take place at night to reduce phone line rates. So it makes sense just to store messages and upload several hours worth of messages at once. More sophisticated systems could actually look at the messages and not call any server for which it didn't have a message. In fact, each E-mail server may not be able to call every server, but only a few. Then those few servers might be able to call other servers, setting up a network in which a message could eventually get to almost anywhere.
No exclusive pathway is required to get the message from Point A to Point B, only a series of links that are created as needed. The message itself contains information as to where it is going. The nodes along the way temporarily store the message and select the next best route to send it on, maintain a listing of those pathways. The message is stored as many times as is necessary to forward it to the right place.
The machines (or nodes) that are used in message switching are not really special machines other than the fact that they generally require a lot of disk space to accommodate long messages. Large hard disks can get expensive, so there is an economic factor to this technology. The machine also has the capability of sending multiple copies of a message out so it could be duplicated for broadcast to other nodes.
Since messages are stored, we can choose how long we want to hold them. In fact we may give them a low priority that indicates that the message is held until all other messages of a higher priority have been sent. This prevents a clamoring for the nodes by many messages at once. In fact, storing data helps keep traffic minimized to some degree which always a positive thing on a network.
A bad point from a wide area network point of view is that storing and forward on a message switched system is slow. Obviously to store a whole message, select a route, and then retransmit the message takes time. This mechanism is totally unsuitable for communicating in real time (interactively, with no gaps in communication). Playing a computer video game that includes a lot of interaction with an opponent on the network would not work very well with message switching, for instance.
In summary, message switching offers good use of network resources providing several features stemming from the storage of messages along the way. However, the expense and delay of the store and forward schema are not practical for some organizations.
2. Packet Switching
Packet switching involves the breaking up of messages into smaller components called packets. Packets often range in size from about 600 bytes to over 4000 bytes depending on the system involved. Each packet contains source and destination information, and is treated as an individual message. These mini-messages are received and routed through optimal pathways by various nodes on a wide area network. There are two major types of packets to be switched, so let's look at their characteristics.
The first type of packet is called a datagram. The name datagram is reminiscent of the term "telegram", and this similarity is no coincidence. When one sends a telegram, they are leaving a lot up to chance. Think about it. If you call the telegram in, you have no absolute guarantee that the message will be sent to the proper destination. As it is being sent, you have no guarantee that the information may not be damaged in some way. Once received on the other end, there is no absolute guarantee that the delivery will take place. Now, probabilities are high that your telegram will be delivered completely intact, but there is always the slightest margin of error that can occur. Can you imagine what would happen if even one letter was transposed on a telegram bound for anxious parents during wartime. "Your son is not dead," may become "Your son is now dead." What a difference!
Datagrams are simply broadcasts to a remote node. There is never a guarantee they'll get there or that the message will remain intact, a fact that could be worsened by a packet switched network. You see, when directing datagrams, there may be more than one route to take along the way. Each individual packet is directed down what seems to be the optimal path at transmit time. Obviously pathways can become better or worse depending on their congestion levels or whether or not they are operating at all. So a datagram from a message may end up taking a different pathway from another datagram from the same message. That doesn't really matter except for the fact that you must keep the packets in order or the message will be garbled.
Figure 4-5: Packet switching networks treat each packet as an individual message to be routed. Requires less overhead then message switching. Messages are broken into packets and reassembled via the PADs.
To alleviate this ordering problem, packet switched networks incorporate a special device known as a Packet Assembler/Disassembler device (or PAD). The job of a PAD device is to make sure the packets are placed in the right order as they are received. But how do you know what the right order is? This problem is solved by placing a sequence number in each packet designating which packet of a message is which. The PAD simply looks at that number in the packet and is able to subsequently reassemble the message that was originally sent. The pad is also responsible for taking messages coming into the network, breaking them up into packets and then assigning sequence numbers to each packet.
Datagrams do not utilize any sort of relationship between the sender and receiver, such as agreeing on packet size. Datagrams also do not typically use acknowledgments, which is a packet sent from receiver to sender acknowledging the receipt of a particular datagram. However, the other major kind of communication mechanism in a packet switched network does utilize the two elements just mentioned.
As a datagram is making its way toward its destination, decisions are being made on the fly concerning the pathway each packet will take. To improve reliability, a decision concerning the best pathway to a destination could be made prior to any data being sent. In this manner, a single, static path could be set up between two communicating parties, one in which the two would use exclusively for communicating with one another. This pathway is known as a virtual circuit.
The idea behind virtual circuits is to remove some of the chance factors involved with datagrams. When creating a virtual circuit, the sender and receiver agree on which path will be used and on packet size. Then during the process of communicating, acknowledgments are sent from receiver to sender in order to verify receipt of the packets. Typically, information is traded between the two communicating entities concerning errors and speed of the transfer. These two factors are known as error control and flow control. A relationship can be set up to last long term spanning several communication sessions or just on a session-by-session basis.
The whole purpose behind virtual circuits is reliability. Though virtual circuits create overhead for communication, they are necessary to ensure that data travels safely from senders to receivers. This is especially important for critical applications. Novell's NetWare uses virtual circuits in allowing administrators to control the file server remotely and in handling communications associated with printing.
In comparing datagram and virtual circuit packet switching with other switching technologies, you must consider several factors. First of all, packet switching is faster because messages are not stored in their entirety for later retrieval. Each packet is small enough to be stored in a routing machine's memory until it can be routed an instant later. Secondly, packet switching allows the avoidance of pathway failure due to excessive traffic loads or mechanical problems. This is accomplished by routing packets along pathways that are the most free and clear. Thirdly, using packet switching allows us to use pathways that may not ordinarily get much traffic. Instead of concentrating on a few paths that are always busy, packet switching spreads the load of communication across several paths.
Packet switching does, however, involve some technology. Intelligent decisions have to be made concerning pathways, and that requires sophisticated machinery. Obviously while path decisions are being made, packets are being held, and that adds overhead as far as time is concerned. In addition, the very nature of temporary pathways is to be less reliable than transmitting data along a fixed physical link, so packet switching can be less reliable than another type of switching we are yet to explore.
3. Circuit Switching
Figure 4-6: In circuit switched networks, a single pathway is set up at the outset of communication and used throughout. This method is very reliable.
Circuit switching involves the creation of a physical path for data flow between a sender and receiver. This method is used to create the links between you and another caller using the phone system. The whole connection of sender to receiver is called a "circuit". Circuit switching offers advantages associated with a physical pathway - like reliability of transfer, because no other devices are contending for the path.
The problems associated with circuit switching is that overhead is required to create the physical pathway. It takes time to put all the links in place to complete the circuit. Once established, the circuit offers dedicated bandwidth to the sender and receiver. This condition is great while each of the pair are actively sending and receiving, but when the channel becomes idle, all that bandwidth is wasted.
In short, circuit switching offers the advantages of having dedicated communication channel between senders and receivers, but suffers from overhead to create the channel as well as maintain it even after transmission is halted.
All three of these switching technologies are implemented in modern WAN systems. System developers must make decisions as to which method best suits their needs.
Chapter 4 Study Tips
1. Know the operation and terminology of the Public Switched Telephone Network (PSTN).
2. Know how wide area networking is accomplished via telephone lines.
3. Know the different types of digital lines and their throughput levels as well as their relative costs to one another.
4. Know what services are offered by ISDN and why ISDN is an improvement over our conventional telephone systems.
5. Know why ISDN will take time to fully implement.
6. Know how microwaves are used in wide area networking.
7. Know how LASERs are used.
8. Know the operation of the three major types of switching.
9. Be able to compare the switching technologies with one another according to time overhead, cost, resource utilization and reliability.
Table of Contents Chapter 3 Chapter 5 Netguru.net Glossary
Copyright© 1993 & 1997 by Atrium Technical Inc. All rights reserved. Illegal to copy without written permission of ATI.