by Arthur Cooper
How does the Plain Old Telephone System (POTS) tie in with wide area networking or local area networking? The answer is not that simple anymore. This chapter discusses the POTS issue as it relates to LAN/WAN internetworking. First, following a brief history of POTS, we look at the function and workings of a single dial-up POTS line. Then, we explore the various uses of dedicated, singular POTS lines between network locations. Following that, we look at aggregate arrangements on POTS connections between nodes and/or network locations.
One of the most misunderstood technologies in use throughout the world is the POTS circuit. Almost anyone on earth who can speak has used a telephone in the last year. In the process of talking on the phone, the signal being transmitted from one end to the other undergoes many transformations. Sixty years ago, the signal being transmitted would remain in analog form from one end to the other. The only signal regeneration that occurred would have been, perhaps, an amplification or two along the signal's path and between switching points. The term Via Net Loss (VNL) was used in those days to calculate the total loss in decibels on a POTS path. Test tones and decibel (dB) or volume unit (VU) meters were the only pieces of test equipment needed to keep POTS lines running efficiently. These meters simply show the amount of gain or loss inherent on a POTS circuit. The decibel is a measurement that is related to the ratio of input to output. The volume unit was similar to the decibel, but it was used to measure complex signals such as the human voice.
In the 1940s and 1950s, POTS circuits would have been far too noisy to be put to any use other than what they were originally intended for: two humans conversing on a telephone. However, since the 1960s, there has been a marked improvement in the way POTS circuits are connected, switched, and transmitted. These changes were due to the advent and proliferation of Bell T-1 transmission systems. As the Bell System began to incorporate T-1 systems throughout the Public Switched Network, the quality and quantity of POTS circuits increased steadily.
During the late 1970s and early 1980s, the Bell System made great efforts to improve the quality of the Digital Access Cross-Connect System (DACCS), which all of the T-1 and T-3 digital carriers traversed. The DACCS system was notorious in the 1980s for minor bursts, frame slips, timing lags, and other similar transmission problems. However, POTS service benefited greatly from the many upgrades made to the DACCS and other digital services of the Bell System. In fact, as digital subscriber service improved, POTS lines were looked at as an alternative to expensive, dedicated, point-to-point data connections. Although the Bell System was aggressively marketing 56Kbps (DS0) and 1.544Mbps (T-1/DS1) circuits to customers, the POTS circuit was just "tagging along for the ride."
After all, on most of these early DS1 systems, the major use for the T-1 stream (1.544Mbps) was to terminate on a D3 or D4 channel bank. These channel banks would in turn deliver wonderful, clear, crisp POTS service from one end to the other of the DS1 stream. What a waste of bandwidth these systems incorporated in those days! An entire 64Kbps slot of a DS1 was being used for Grandma and Grandpa to speak to each other. It was at this point POTS lines were looked at as an alternative to expensive, dedicated, point-to-point data connections. Many of the networks in existence at that time were actually collections of dedicated circuits. There really was no structure or flexibility on these "networks." If the line to Peoria was down, Peoria was out of luck.
These dedicated circuits seemed to be marvels to information systems (IS) professionals. Generally Ma Bell would install the circuit and the related modem. What the IS professionals of that day did not understand was that the modem in use on any particular data circuit was being routed to an interface card. This interface card, in most cases, existed in a channel bank somewhere. If the IS professionals had any inclination or ability to realize this fact, they probably would have immediately reviewed the way they were doing their internetworking and connectivity. These point-to-point circuits were outrageously priced in the 1970s and early 1980s. With the costs of dedicated data circuits reaching levels of thousands and thousands of dollars, many IS professionals began to look at POTS as a cheap, reliable backup.
Dedicated 9600bps and slower circuits were, for the most part, extremely reliable. Once IS professionals realized these circuits were traversing the same path as their telephone conversations, the POTS world became extremely important to the data communications world. Companies such as RACAL-MILGO, PENRIL, MOTOROLA, and many others began to sell dial-up and dedicated modems at a rapid rate. As we moved into the 1980s, POTS became more than just a dial-up backup to dedicated circuits. Network computer systems and nodes began to incorporate POTS dial-up support within the major routines of their operating systems and functions. POTS circuits were finally accepted as legitimate members of the data communications family.
The most important thing to keep in mind when looking at the function of POTS lines is the fact that these lines, or circuits, were originally intended to be used by human beings. Also, the original concept of the Bell System was a series of POTS trunks interconnecting the major cities of the United States. Within these major cities, telephone operators would manually connect the calls coming in over the trunks and local lines. The term local loop has often been used to describe local POTS circuits, and it is a fairly accurate assessment of what is happening when one uses the circuit. With that in mind, let us now look at how these lines operate.
Each POTS circuit uses two wires to complete the path between an end user device (telephone, modem, and so on) and the local exchange switch. These wires are referred to as TIP and RING. Believe it or not, these terms go back to the days of cord patch boards manned by human operators--the TIP and the RING of the patchcord were wired out to each local subscriber; hence, these two terms.
Refer to Figure 12.1 for an illustration of how a POTS circuit functions.
FIGURE 12.1. Basic POTS circuit.
Notice that there is a switch on the user end of the POTS circuit. This switch refers to the electronic portion of your user device (telephone, modem, and so on) that closes the two-wire loop when you are demanding service from the local exchange carrier (or LEC). The other end of the circuit terminates somewhere in the LEC's local switching facilities. The point must be made here that there may or may not be actual metallic, copper connectivity between the end user and the LEC.
On the contrary, there may be a series of channel banks and/or multiplex units between these two endpoints of the POTS circuit. For the purposes of this discussion, let us assume there is two-wire copper connectivity of some sort between the user and the local switch providing dial tone. When the user picks up the phone and closes the two-wire loop, the local switching equipment senses a return of current on the loop. This is accomplished through completion of the DC circuit between the switch and the user device. In other words, when the user device goes "off-hook," a DC signal applied to one side of the circuit by the switch then returns on the other side of the circuit. The switch sees DC current flowing through the entire loop.
This is what causes the switch to recognize there is a request for service from the user. At that point, the switch connects a dial-tone generator to the loop. Remember, dial tone is in an analog format. When dial tone is sent over the loop and reaches the user device, either a human hears the tone, or a modem or some other device "senses" the tone and dialing commences. If the user device is incapable of producing Dual-Tone Multifrequency (DTMF) signals, sometimes referred to as touch tone, then dial pulse is used.
Dial pulse works quite simply. When the user device sends out dial pulses--that is, a person or a modem dials--the loop is simply opened for a short interval of time that directly corresponds with the number being dialed. In other words, if a 1 is dialed, the loop momentarily opens for one short interval. If a 5 is dialed, the loop momentarily opens and closes for five short intervals. When the digit has been sent, the loop is again closed. This series of openings and closings of the loop activates a relay in the switching equipment, and that relay provides the switch with the numbers actually being dialed by the user device. Also, the dial-tone generator is disconnected at the first sign of an incoming dial pulse.
Since the early 1970s, most telephone switches began the conversion to DTMF, touch-tone signaling. When a user device is using DTMF instead of dial pulse, the sequence is somewhat different. After the loop has been closed, the switch simultaneously connects a DTMF register/receiver and a dial-tone generator to the loop. After the user device "senses" the dial tone, or a human ear hears it, DTMF signals are sent to the switch by a user pushing the buttons on a phone or a modem transmitting the DTMF signals. Once again, the dial-tone generator is disconnected at the first sign of an incoming DTMF signal. All of the DTMF signal frequencies are within the audible spectrum, hence the "beep, beep, beep" sounds in your ear when using a touch-tone phone or listening to the dialing sounds a modem makes.
Now that we have seen how the switch senses a closing of the loop, provides dial tone, and then receives digits, it is easy to see how a POTS line functions. It is important to note, once again, that the example we have been using assumes a copper two-wire connection between the switching facility and the user device. In reality, however, most POTS circuits are not wired as straight copper to all user locations. For user locations located far from the switching facility, fiber and digital carrier systems are used between the switch and collocated wire plants in various parts of a city or town.
All of us have seen the green boxes at the side of the road with the Bell System logo on them. Most of these are covering up repeaters and subscriber loop equipment of some type. These repeaters and line interface units are what allow a local loop to be located far away from the actual switch providing dial tone on the circuit. Today, with the advent of microelectronics, the local telephone companies are using very small units connected to fiber runs underneath the streets. These units then break out into interface boxes with copper wire in them. Why doesn't the telephone company run straight fiber into the home? The answer is simple. Almost all user devices still use copper wire to connect to a POTS circuit. Also, until the cost of ripping up the streets of a city to run fiber connectivity into all buildings and houses comes into an acceptable range, the telephone company will no doubt continue to use the old tried-and-true copper wire that has been buried under the street for many, many years.
When interfacing POTS circuits, it is important to consider that a single POTS circuit consists of a TIP wire and a RING wire, as mentioned earlier in this chapter. The TIP and RING of the POTS circuit are generally wired to some sort of jack or interface. In the United States, the typical termination for a single line is wired according to USOC standard RJ-11. If there are two POTS circuits present in one spot, typically they are wired into one jack according to USOC standard RJ-14. If there are between three and twenty-five POTS lines in one location, typically they are all terminated on the left side of a 66-type demarcation block. This arrangement and the associated wiring is in accordance with USOC standard RJ-21X.
These standards may seem confusing, but they are actually simple. Think of the RJ-11 and RJ-14 as being standard, modular phone jacks. The RJ-21X is simply a 66-type termination block that utilizes a "quick-disconnect" on one side in order for the local exchange carrier's technicians to disconnect the network side from your equipment side should there be a difficulty or failure of some sort.
NOTE: The 66-type block is simply a square piece of plastic, usually white, with rows of metal pins sticking through it. The block allows you to connect wires to it by "punching" them down onto the pins. Each pin has a y-shaped opening on the end, and as you punch the wire into this y shape, the wire is "squished" between the sides of the y shape. This forms a strong mechanical and electrical connection to whatever is hooked up to the pin on the other side of the 66-type block. Typically, the other side of the pin is factory wired to a connector on the side of the block. In other instances, the rows of pins on the block are connected together horizontally. If this is the case, you will have to "punch down" something on either side of the pins to make a connection.
Having the disconnect also facilitates troubleshooting at your location should it be required. On the right side of this block, you would need to punch down any connections you need to make. The wiring you punch down, or connect, to this side could then be extended out to standard, modular phone jacks. Or in some cases, there may be a need to wire the right side through a series of patch panels or frame blocks. The truth of the matter is, after 1982 when the divestiture of the Bell System occurred, most local telephone companies do not really care what you do with your internal wiring.
For that reason, IS professionals have a wide range of options when it comes to deciding how their network will interface with the outside POTS world. It is with great apologies to all other configurations that I now suggest what I feel is an ideal connection situation. Let us assume that the LAN we are using for our discussion consists of approximately 50 workstations and one large server. The server has need to do dial-up connectivity with other LANs within the organization that are not located in the same city, so there are six POTS circuits connected to the server. Let us also assume that these POTS circuits are strictly dial-up, therefore they will be using 33.6Kbps dial-up modems of some type. (We discuss POTS modems later in this chapter.) Let us also assume that these dial-up lines are going to be used for local access to the LAN by the organization's employees.
In order to follow along with this discussion, please refer to Figure 12.2.
FIGURE 12.2. Basic connection diagram.
Notice that there is a block labeled RJ-21X. This is a 66-type block as mentioned earlier in this chapter. Also, notice that the premise, or local in-house, wiring is on the right side of this block as also mentioned earlier. This may vary according to the practices of the local exchange carrier. The next item on the circuit diagram (following it backward into the LAN) is a patch panel of some sort. Many people say this is an unnecessary step. I have heard people say for years that patch panels are wasted pieces of equipment. Further, people have told me that patch panels only add to the problems a circuit may develop. These statements are both false, in my opinion.
In the nearly 19 years I have been involved in networking, the ability to patch around failed equipment and circuits, coupled with the excellent monitoring and troubleshooting capabilities of patch panels, have convinced me of their necessity. In our dream configuration, patch panels will be a part of the circuit. These patch panels allow you to monitor and access your POTS circuits with ease and efficiency. These panels pay for themselves the first time you patch rapidly around a failed modem in a hunt sequence.
NOTE: A hunt-sequence refers to a group of POTS lines in which, typically, one telephone number is assigned to the entire group of lines. If the number is dialed, the local exchange carrier rings in on the first available line that is not in use. Likewise, if the lines are used for outward calls, the user's system uses the first available line to place the call on.
This ability to patch around a failed modem ensures that both inward and outward calls are completed on properly operating circuits only. The failed modem is removed from the configuration by the use of the patchcord in the patch panel.
Why not just swap out the bad modem with a good one? There may be a lot of reasons why this would have been the wrong approach. First, the modems may be rack mounted, and pulling modem cards from a "live" rack can damage them. Second, what if the modems are not collocated with the interfaces or the server? Rather than running back and forth and getting upset over the fact that you have wasted time and still not solved the problem, why not simply make the patch, perform the simple deactivation and activation mentioned, and return a high level of service to the system users? You can always go back at your leisure and repair or replace the modem. The good thing about having followed the patching procedure is the fact your users will have a lot less interruption of service when you do it this way.
Let us refer once again to Figure 12.2 and continue following the connectivity diagram. Now, from the patch panel backward, there is an interface of some type between the panel and the actual modems being used. This too can vary greatly, but the best configuration, in my opinion, is to have a modular connection cord between the patch panel and the modems themselves. Many companies, such as ADC, Inc., offer excellent modular patch panels to be used with POTS lines and modems. The great thing about using modular connections is the relative ease of use and cost associated with this type of connection. Ideally, the patch panel equipment and the modem equipment should be located together. In a single rack, on a single backboard mounted on a wall, or even inside a single cabinet or set of cabinets, is the best configuration to have when using POTS circuits.
Once again, looking at Figure 12.2 and following backward from the modems, you can see that the next connection is on the data terminal equipment (DTE) side of the modem. This connection goes back to the data communications equipment (DCE) portion of the server. This connection can vary greatly, but today, most people are using either Winchester-Type connections, or they are using standard cables with serial RS-232/DB-25 pin configurations.
NOTE: Winchester-Type connections do not necessarily have to be associated with V.35 interfaces; however, they usually are used for V.35 interfaces only. Also, a DB-25 connector, which is a standard 25-pin connection, may be used with V.35 interfaces.
Once again, the interfaces and cables are not important here. We are only trying to show the location of their placement within the circuit flow of a typical POTS connection to a network.
To sum this discussion up, let us review what we have learned in this section. Typically the local exchange carrier brings POTS lines in on an RJ-21X/66-block arrangement when there are three or more POTS lines. Typically, the user (you) connects on the right side of these blocks with some type of approved cabling. This cabling then completes the circuit to the output/input line connections of the modem(s). In between this connection, I am recommending patch panels be installed. The next step is the actual data cabling between the modem(s) and the network server or termination (router, bridge, and so on) equipment.
NOTE: The only exception to this would be for modems that are internal to the server or termination (router, bridge, and so on) equipment. In these cases, there would be no cabling between the DTE of the modem and the DCE of the server. With internal modems, the DTE/DCE connection is handled internally between the "slot" in the system and the printed circuit board of the modem.
This completes the interface between a POTS line and a network server. The only variation to this typical connection design would be if the network or LAN servers were connected to termination equipment such as routers or bridges. In that case, the modems for the POTS circuits would be between the routers or bridges and the patch panels mentioned earlier. The DCE connections would be on the routers or bridges and not on the server itself. Also, if the router or bridge is going to control the dialing of the POTS modem, it must be equipped with software capable of doing so. CISCO, Bay Networks, 3COMM, and most other intelligent bridges and routers are quite capable of handling modems and "modem-pools."
NOTE: A modem-pool refers to a collection of POTS lines and modems that are connected to a server, router, or bridge and that can be accessed by all of the users on the LAN via the server.
It would be difficult to cover every type of POTS modem that exists, so for the purposes of this discussion, I will introduce three basic classifications of POTS modems that are currently popular and in widespread use. The first type of POTS modem is the single-line, dial-up, external standalone type. The second type is the single-line, dial-up, internal standalone type. The third type is the multiline, dial-up, rack-mounted or shelf-mounted type.
NOTE: Chapter 13 of this book covers ISDN, so I mention here only the fact that ISDN modems are being considered POTS modems in many circles. There is a debate over whether a POTS line can only be the standard analog TIP and RING circuit. However, ISDN 128Kbps circuits are being made available for home and business use and are sometimes referred to as POTS ISDN lines.
Let us now look at these three basic classifications of POTS modems.
This type of modem is the most common one in use today. Many Internet Service Providers (ISPs) use external standalone POTS modems to comprise their local modem pools. A common brand of modem in use at ISP locations is the U.S. Robotics Dual-Standard/33.6 dial-up modem. ISP organizations simply use shelves or wall mounting to place all of these modems in close proximity to each other. If the RJ-21X connection for these modems is physically close to them, it makes no sense to wire patch panels in line with the modems as shown earlier in this chapter. In these cases, the right side of the RJ-21X block is wired to individual modular RJ-11 jacks. The modems are then connected to these RJ-11 jacks with common modular line cords.
In the case of a network using standalone modems, it is a good idea to try to locate the server or terminating equipment (router or bridge), the modems, and the RJ-11/14/21X interface points in close proximity to each other. Another factor to consider with these modems is the amount of heat they give off and/or can tolerate. If the closet or room where your interfaces are located is not cool or equipment-friendly, these standalone modems may overheat or malfunction. The newer modems are pretty reliable, but there is always the chance one may fail due to excessive heat.
This type of modem is basically identical to the first type with one exception, which concerns the placement of the modem. Since these modems are internal, they become an integral part of the circuit boards making up your server, router, or bridge. Most companies make internal POTS dial-up modems to fit their equipment. For example, CISCO Systems, Inc. makes modem cards that fit into their router products. These can use POTS or ISDN trunks for dial-up networking. If we are talking about a small network using a PC as the server, a standard internal modem operating at speeds between 300bps and 33.6Kbps can be used. The cost of these modems has come way down in the last two years. There are also companies that produce a single board holding more than one POTS modem and that can be plugged into a single slot on a PC's motherboard. This is an outstanding solution for a small LAN that uses POTS circuits as its main LAN/WAN interconnection method.
One point needs to be made again for patch panel configuration. With internal POTS modems, there are going to be no LEDs or lamps on the face of the modem that can be observed during the modem's operation. Many people call these idiot-lamps, but they can prove invaluable to someone who is troubleshooting a problem with a POTS modem or circuit. With an internal modem, there needs to be some sort of routine in the server, router, or bridge's internal operating system that extends this information out to the user or network administrator. If there are no software means of getting this information to the outside world, it can be very difficult to tell just what a modem is "doing" at different times during its normal operating cycle, hence, the necessity of good patch panel facilities between the modem(s) and the outside POTS circuit(s).
With only a speaker or a telephone butt set, a technician can patch into the monitoring portion of the patch panel and ascertain what the status of the modem is. I have seen many network administrators forced to reboot their servers when their POTS circuits and/or modems failed. Not only does this inconvenience users and cause data loss, but it also does not tell the administrator what the real cause of the problem was. In the case of a network using three or more internal POTS modems, a patch panel becomes an absolute necessity.
These types of modems are becoming extremely popular. There are many companies offering modem solutions in many different forms for LAN/WAN interconnection. A typical arrangement is a rack of modem cards in individual slots. From this rack are individual modular connections out to the POTS circuit interface equipment. The other side of this arrangement has an interface box or widget of some sort that interfaces the modem rack to the server, router, or bridge. The obvious benefit of these systems is that more than one connection can be made at a time, and therefore many POTS calls can be up at a time.
One of the best variations I have seen of this technology is used by CISCO Systems, Inc. They provide multiline POTS interface capabilities on their router systems to place on-demand calls between routers of different LAN systems. By using this type of interconnection, these so-called modem racks, or nests, can provide excellent, on-demand bandwidth between the routers of an organization's LANs.
There are also vendors that can provide racks capable of holding several of the internal modems mentioned in the last section. These racks of internal modems can be controlled by an interface widget that connects to serial ports on the server, router, or bridge, or directly to the LAN. In fact, some vendors are designing standalone racks of dial-up POTS modems that can interface directly with the internal structure of a LAN. For example, there could be a connection made by coaxial Ethernet to this nest of modems. A software package would allow all servers on the LAN to get to the modem nest and use one of the modems if one is indeed available.
Once again, there are simply too many variations of interfaces, racks, shelves, and electrical characteristics to attempt to list them here. However, IS professionals should be aware of the virtually limitless types of POTS modem racks, nests, or pools available for LAN/WAN interconnection.
If an organization needs more bandwidth than a single POTS line can provide at any given time, it makes sense to use more than one, right? Well, the simple truth of the matter is that there is currently a limit of 33.6Kbps on most singular POTS circuits. Since most servers on LANs are incapable themselves of using the bandwidth of a combination of two or more POTS lines at the same time, the highest amount of bandwidth most servers can gain from a POTS circuit at any given time is 33.6Kbps. In other words, if an organization's server in Maine is connected to another of its servers in Colorado through a POTS circuit, the two servers can only hope to have a bandwidth of 33.6Kbps available for their use, right? Wrong. There has been a recent rise in the number of vendors willing to provide the termination equipment capable of combining POTS circuits to provide higher aggregate bandwidths.
This started with the voice communications world back in the 1970s. In those days, Northern Telecom, Inc. came up with a system called Comm-II (or Comm-Two) equipment. This equipment was able to take 10 POTS circuits, and provide dial tone for maybe 12 to 18 circuits using only the path of the original 10. It is obvious how this equipment worked. It simply stole free time from circuits where no human voice was present and crammed data from other circuits with voice present into this stolen time.
Many would argue that T-1 was doing this job in the 1960s. Why did we need Comm-II equipment? You are comparing apples and oranges. A T-1 circuit is a digital stream that facilitates the transmission of 24 analog POTS circuits using two pairs of wires. The main reason the Bell System created T-1 was to maximize the use of its copper wire facilities and transmit more than one circuit over two pairs of wires. When Northern Telecom was creating their Comm-II equipment, T-1 was not readily available for use by private industry. Further, LANs were not nearly as prevalent in business as they are today. Rather, Northern Telecom was providing a business customer a means to connect more people than they were paying for.
Since the invention of this equipment, many other vendors have come up with ways to provide increased bandwidth using more than one POTS circuit simultaneously. However, today the most common use of combining POTS circuits is done within a router or an intelligent bridge. When a LAN/WAN interconnect design is put together with routers on the controlling end of the interconnection, it is possible to take POTS lines and combine them into aggregate connections. See Figure 12.3.
FIGURE 12.3. Aggregate POTS connectivity.
By using a router intelligent enough to put up two or more POTS calls to the other LAN's router, there can be an aggregate connection at speeds beyond the 33.6Kbps barrier of most POTS singular lines. This reduces the need for long connect times between LANs, and it speeds up the transfer of large files and/or e-mail between the LANs comprising an organization's WAN. Also, many of these new router systems are able to break out telephone and fax capabilities to be used when the server is not using the POTS circuits for data transmission.
It is interesting to note that at the time of this writing, many vendors are looking to increase POTS speeds to 56Kbps. One of the foremost vendors in this area is U.S. Robotics, Inc. (USR). It has devised a standard called X2 Technology. Their position is a very valid one: Because most of the Public Switched Telephone Network (PSTN) is now in a digital state and capable of passing higher bandwidths, USR feels the time has come to produce a system capable of passing 56Kbps over a standard POTS line. I am not talking about an ISDN line; most ISDN lines in use today are of the 128Kbps variety. No, rather, what I am talking about here is a chance to push the envelope of what was previously held as the limit for an analog POTS circuit's bandwidth.
If USR is effective in developing and marketing the X2 standard, we may see POTS circuits operating routinely at speeds up to 56Kbps. If we take this a step further and look at the intelligent capabilities being designed into routers in order to combine POTS circuits and increase bandwidth, many IS professionals will have to rethink the way they have been designing their LAN/WAN interconnections. A combination of X2-capable modems and intelligent routers will provide organizations with outstanding POTS aggregate bandwidth capabilities.
When using POTS lines in a data network, generally it is for the purpose of bringing connectivity in from a location far-removed from your local area network (LAN). In other words, the POTS line is being used to facilitate a connection to another LAN, which may be a part of your organization but is located across town, across the country, or across the world. The second reason POTS circuits are typically used is to facilitate the local or traveling corporate employee who needs access to the services of the corporate LAN when he or she is at home or away on company business.
There has been much discussion over the years as to just what constitutes a LAN or a WAN. In my opinion, any time you are networking individuals together who reside in the same office, room, or building within a corporation, you are involved in a LAN situation. When you start connecting these LANs to other LANs that may be across town, or in another country, then you are taking the step into Wide Area Networking (WAN). The connections made between LANs and WANs are the most misunderstood.
There are many schools of thought concerning what technology to use on LAN/WAN interconnections. For the purposes of this chapter, I shall make no attempt to pick a certain technology and declare it as the only one to use. Rather, I will highlight situations when POTS lines themselves would be the best candidates for LAN/WAN interconnections. There are basically three situations which I see as being correct for the use of POTS circuits as the means to connect LANs together, hence creating a WAN.
The first situation would be when only e-mail and a small amount of file sharing is needed between LANs in an organization. The second situation would be where the geographical separation of the LANs in an organization is far-flung. For example, a company with offices in Paris, France, and Los Angeles, California. The third and final situation would be where the organization has only one LAN and does business with other organizations only on an as-needed basis. As you can see, in all of these situations, it would make no sense to make LAN/WAN interconnections using expensive, leased connections between locations on the network. Further, in the last example, POTS lines are the only way the two LANs could make connections, unless they decide to use the Internet for connectivity and their Internet connection is by some other means. However, even in that scenario, there would most likely be a POTS connection to the Internet at one or both organization's networks. Let us explore these three situations individually.
In this situation, the best way to gain connectivity between two or more LANs in the organization would be to use a few POTS lines in a modem pool on each of the servers within the LAN structure. Whenever there is a situation requiring the LAN to connect with one of the other LANs in the organization, the server would be programmed and setup to automatically dial the other LAN's modem pool and make a connection. This seems to work really well when the organization's LANs need only small amounts of interaction between themselves. I have seen many organizations waste money and bandwidth by installing T-1s and even T-3s between themselves in an effort to provide a WAN interconnection between their LANs.
If they had only taken the time to do an effective network traffic analysis, they would have seen where a few POTS lines would have sufficed. It makes no sense to pay for bandwidth that is not being used. An on-demand modem pool can pass data between LANs extremely efficiently. Also, if the time factor is not as crucial, the servers can be programmed to make connectivity only with other LANs in the organization when the rates on the POTS lines are cheaper. If e-mail is the only thing being passed between LANs, it may also make sense to purchase POTS service from a local Internet service provider (ISP) and simply do a dial-up Internet connection at the organization's LAN locations.
The only benefit from using an ISP to pass e-mail is that most ISPs provide e-mail software of some sort for you to use. Either way, whether you use direct dial-up between LANs, or an ISP/Internet connection to pass e-mail, a modem pool using POTS service is the only way for these organizations to go.
This situation is very prevalent today with the advent of international business relations and the so-called "global village." There are many organizations with offices all over the world, and to try and maintain fixed, leased, or owned connections between these LANs would be expensive and unnecessary. For this reason, it only makes sense to use POTS connectivity between LANs. If an organization's LANs are sharing large amounts of file transfers, e-mail, or on-line data, it may become necessary to put in fixed, leased, or owned connections between the LANs. However, most situations involving LAN/WAN interconnect with organizations that are spread out can be facilitated through the use of POTS circuits.
Once again, as in the situation above, a modem pool connected to the server(s) on the LANs is the best method to use. Most LAN application software available today can be configured specifically to an organization's desires. If there is only a small amount of e-mail going back and forth, and it is possible to tie in with local ISPs at each LAN location, the modem pools would only be used to connect with the ISP. The mail would travel the Internet to reach its destination.
However, if security is a consideration, or if there is time-critical e-mail and file sharing that must occur, then an ISP would not be used, and the modems would be configured to direct dial each other and make a LAN-to-LAN connection. If the LANs begin to use longer and more costly connect times, the organization may decide to go with Frame Relay connectivity between LANs. This is covered in Chapter 17, "Frame Relay." Without going into a long discussion concerning Frame Relay, it too may use POTS circuits in modem-pools, or it may consist of routers and leased loops to the frame relay network provider.
The main thing to understand, once again, is the traffic of the organization's network. If the traffic is somewhat normal or moderate, POTS lines are the only way to go. The LAN administrators at all of the organization's locations would then need to purchase the cheapest long-distance service available and configure the network servers to dial only when needed and hopefully at the cheapest times. If a little common sense is used, POTS lines can provide the users at all of the organization's LAN locations with a decent level of service.
If an organization is small, and it has only one LAN in use, POTS is really the only sensible way to connect this LAN with the outside world. Typically, POTS lines would be used to gain access to an ISP in the organization's locale. If the LAN has need only for e-mail services and nothing else, a single POTS line to connect to a local ISP is the best way to go. It is amazing the amount of e-mail that can pass on a 33.6Kbps modem over a POTS line. It is just as amazing how many small organizations I have seen purchase too many POTS lines for their LAN. The typical ratio is 10 users per modem. That seems to work really well. With the cost of modems going down, and the cost of monthly Internet access through an ISP going down, using the Internet to pass corporate e-mail has become the best way to go. If security is a concern, there are many software-based encryption programs that can provide organizations with secure mail on the Internet.
Now that we have looked at the three situations involving POTS lines as the best candidates for LAN-to-WAN interconnection, I am sure there are a million exceptions to these three situations. However, if you look at the general points made in the earlier discussion concerning these three situations, I am sure you will agree POTS lines are a viable technology to be used when connecting LANs together. The only reason many network people still look at POTS interconnectivity as being a backup to a real LAN/WAN network connection is the fact they have fallen into the same trap as many IS professionals over the past 10 years.
Many network companies such as MCI and Wiltel offer cheap, fast interconnection between LANs. To many IS professionals, these types of connections are the only way to go. However, if they really knew how little of the bandwidth of those T-1s they purchased was actually being used, they might think twice about continuing to pay for them. Also, time-sharing has become a somewhat anachronistic way of doing business in the IS world. There are very few situations where people are using their PCs or terminals and interacting directly with a server over long distances.
Rather, this direct interaction most likely is occurring on a local level between the user terminal, or PC, and the LAN server the PC is connected to. If an organization has terminals and employees all over the world and these employees are interacting live, online, with a server that is located many, many miles away, it is time for that organization to give up the central control mindset that must rule in their network design ideology. Now I am not saying there are no situations in which this design might be correct. However, I am confident that when the analysis of an organization's business is thoroughly reviewed, it will make sense to provide modern, high-speed LAN functions at user locations. The next step is the connection of these LANs as the means to pass data between all the members of the organization. It is this connection, this LAN/WAN tie-in, that will be the most crucial part of any network. Most LANs in an organization do not need to interact with each other's servers on a constant basis. For that reason, POTS circuits may be the best choice in making this all-important LAN-to-WAN interconnection.
In any discussion of POTS, it is necessary to talk about the Serial Line Internet Protocol (SLIP) and the Point-to-Point Protocol (PPP). These two protocols are probably the ones most widely used on POTS lines connecting home/business users and the Internet. Also, as more and more enterprise networks are being built on TCP/IP-based platforms, SLIP and PPP protocols will provide the capabilities of POTS interconnection between LANs on these types of networks. Let us look briefly at both.
This is an older protocol that was developed to pass TCP/IP data transmission over serial lines. Because TCP/IP networks have become more important in recent years, it is important to know about this protocol. It was originally developed to be used over serial lines, and it provides a TCP/IP connection over POTS circuits. The SLIP protocol is not an Internet standard, and there are many different versions of SLIP floating around. However, it is capable of framing and transmitting IP datagrams on serial connections. Because it is not a standard protocol, SLIP has no maximum packet size specified. Therefore, any size can be used as long as both ends of the connection are using compatible packet sizes.
The PPP protocol is a standard protocol for use over serial line connections. The difference between PPP and SLIP is the fact that PPP is based on a standard developed by the ISO called High-Level Data Link Control (HDLC). HDLC is very common and has been incorporated into X.25, Frame Relay, and ISDN. Hence, PPP is more common and more widely accepted than SLIP. Once again, PPP provides a TCP/IP connection over POTS circuits just as SLIP does. The difference is the fact that PPP is a standard protocol and therefore more likely to be used.
There are many popular software packages that create a TCP/IP stack in the Windows or UNIX environment and that are capable of dialing into either ISP networks or into an organization's server. Once this dial-up connection is made through a POTS circuit, the system using the connection is capable of running most TCP/IP-based protocols and functions over the connection. As more and more enterprise networks go in the direction of the Internet and use TCP/IP as the backbone of their LAN/WAN interconnection services, there will still be a need to pass TCP/IP services over POTS circuits. SLIP and PPP are currently the only methods of doing so that are in common use today.
No matter how much one wants to get away from work when the day is over, there will always be a need for remote access into an organization's network by its members. In order to facilitate easy remote access to an organization's network, there is only one accepted method: POTS circuits. There are many different ways to provide remote access, however, typically it is done by simply dialing in on one of the POTS circuits assigned to your modem pool or nest and creating a data connection between your home or laptop PC and the network's server.
Today, most remote access users are employees on the go or those authorized to telecommute from their homes. These are the people who are most likely to need remote access to the organization's network and its resources. There are many ways to make these connections, but the simplest method is through the use of a dial-up modem on the remote user's PC or laptop. The other end of the picture is a modem, modem pool, or modem nest located as an integral part of the organization's network. The user simply dials into the network's equipment and the connection is made, after some sort of security check for the authenticity of the user.
NOTE: Telecommuting is a term given to those who work at home, yet are connected somehow into the organization's network and computer resources.
Organizations today are apt to use security checkpoints and other software tools to ensure that only authorized users are able to gain access to their networks. Once the user has connected to the organization's network, the server determines what type of connection in terms of use and appearance is afforded to the user. There are many client/server software application packages around today that allow a remote access connection to look and act nearly the same if not identically to the connection users receive when logged into the LAN on a locally connected workstation.
However, remote access is but one use of POTS other than network LAN/WAN interconnection. Another common use of POTS circuits is on Point-of-Sale (POS) transaction machines, such as credit card readers, cash registers, and Automated Teller Machine (ATM) equipment. In many ATM networks, banks have taken it upon themselves to use leased, fixed lines between terminals and banks. However, the smarter banks are using on-demand POTS circuits and modems to connect ATM machines, banks, and credit-approving networks. Whenever a credit card check is made on a small POTS-oriented credit card reader, an on-demand connection to a credit-approving network is made.
Still another use for POTS circuits that is becoming extremely common is the connection used by various state-run lotteries. In states where there is a lottery, there seems to be a store on every corner where lottery tickets can be purchased. However, what do you think is the glue holding these lottery ticket vendors to the main or central lottery network in the state's capital? A simple POTS circuit, of course. Whenever these vendors are involved in small numbers of ticket sales, an on-demand POTS circuit is used. Only when vendors consistently generate an extremely large number of ticket sales do states tend to install leased, fixed lines between themselves and the vendors. Even then, the leased line may actually be a POTS connection that is dialed and left up all day long.
As you can see, the many uses for connecting various networks together by way of POTS are astounding. It would be foolish to try cover all of them in this chapter. What we have tried to do is provide a general outline of what POTS lines are all about and how they can be used in the business of networking. It always amazes me when I see new uses for POTS lines. Over the years, I have dealt with many creative IS professionals who have come up with some pretty far-fetched uses for POTS circuits.
I even met a man once who used a POTS circuit to connect his LAN to the electric fence holding his horses. This was how he was able to know immediately if the electric fence on his property had been breached. The fence's alarm system had a POTS line terminated in its cabinet, and when a breach occurred, the system would dial the man's LAN and send a prearranged login sequence to the server. The server had been instructed to notify the man whenever this login name was used. Does it seem far fetched? Maybe. But to my friend, it was a means to continue working at his office and know that his horses and their fence were probably still intact.
Literally, we are moving to a time when the only analog portion of a telephone line will be at the user's handset. Until human beings are able to hear and speak in digital signals, that analog portion will remain intact. However foolish the preceding statement may seem, it illustrates dramatically the main reason why POTS lines have always been considered analog in nature and somehow less important than digital circuits and systems. While most of the public networks these POTS circuits traverse are digital, there is still a lot of copper wire buried under the streets of the world's cities. While some of this copper wire can be used for digital transmission, a lot of it is old and not in the best condition.
For these reasons, there will always be a certain stigma applied to POTS circuits by IS professionals. Even when I have not been involved in the conversation, I have overheard IS people talking about their nice, new fiber connections and their old, rotten copper connections. There is one fact that still remains true. The public switched network is the most rapid means of gaining connectivity into any area at any time. There will always be a need for this network, as people will always want to have telephones in their homes and businesses.
The telephone is still the most important device in any office. Do not let anyone tell you differently. As long as we continue to tie high-speed data technology to the technology of the public telephone network, there will be continued use of POTS circuits in LAN/WAN interconnections. With the advent of new technologies such as USR's X2 modem technology, the POTS circuit will continue to redefine its rightful place in the world of data networking. ISDN is the technology a lot of people are looking at now for dial-up network access. There is a lot of spirited talk of how ISDN will replace POTS. What I say to you is this: POTS will never go away. It will simply transform itself.
The term POTS will simply be applied to newer, different technologies that emerge in the public switched network. Perhaps all of us "old-timers," those who came to know and understand the fact that POTS meant wires and the ability to converse over them, will never be able to fully let go of POTS as we knew it. However, there will be a new generation of laptops, hand-held assistants, fax machines combined with cellular telephones, and other such innovations that will all simply reinvent the term POTS. The Plain Old Telephone System, POTS, may come to mean digital-switched circuits that run at speeds in excess of 1Mbps or better. I am confident the term POTS and the circuits denoted by the term POTS will be with us for many years to come.
© Copyright, Macmillan Computer Publishing. All rights reserved.
){kind=link}
){kind=link}
){kind=link}