xDSL: Fact or Fiction?

Web Languages
Typography
  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times

Want higher bandwidth at a lower cost with no change to your existing copper wiring? Digital Subscriber Line (DSL) technology may deliver and become the hottest thing in telecommunications since the 1984 AT&T breakup. This article explains what DSL technology is and how it can affect your future communications capacity.

Since I am directly involved in the telecommunications business (I work for a long- distance carrier), I am often asked questions about anything that is even remotely related to telecommunications. People often seem surprised when I don’t have answers on the tip of my tongue. When this happens, I am reminded of when I first moved to California and my family in the Midwest would ask me if I’d met any movie stars yet.

There is so much happening in telecommunications today that, even if I wanted to, I don’t think I could keep up with it all. Sometimes, though, I am close enough to get a feel for what’s hot and what’s not. Digital Subscriber Line (DSL) technology, for instance, could be the hottest thing to hit the communications business since the breakup of AT&T in
1984.

The demand for high bandwidth access to the Internet and other advanced data services has been steadily increasing for some time. Several solutions have been offered for the demand for more bandwidth: ISDN,

56-kilobits-per-second (Kbps) modems, Cable modems, ATM (see “Is Asynchronous Transfer Mode the Next Super-protocol?” in this issue), and Satellite Internet access, to name a few. DSL technologies could be the most promising solution yet. With speeds starting at 64 Kbps and going all the way to 25 megabits per second (Mbps), all over a standard telephone line, it definitely has potential. “So, why hasn’t the DSL revolution started yet?” one might ask. Well, maybe it has.

xDSL: What It Is and How It Works

xDSL is the name generally used to refer to a group of DSL technologies. There are (at last count) six different types of DSLs in various stages of development and implementation. For most of this article, I’ll focus on Asymmetric DSL (ADSL) since it is

the most likely candidate to be deployed to the masses. I will give a brief description of some of the other “flavors” later, but until then, I will be referring specifically to ADSL.

Before I begin, I will take a moment to explain the important difference between upstream and downstream. Upstream refers to a transmission going from the user to the network (i.e., uploads). Downstream refers to a transmission from the network to the user
(i.e., downloads). The network could be any network, but it is generally a local telephone company or an Internet service provider (ISP). Now that the terminology is clear, I will explain how ADSL works.

Normally, the telephone line that comes into your home is made up of two copper wires. These wires are commonly referred to as a twisted pair—not because they suffer from any psychosis but because the wires are twisted so that the transmit and receive channels are reversed as they travel between the network and the user (i.e., the network transmit channel will be “received” by the user’s receive channel). This configuration is standard for virtually all residential phone lines in the United States. Attaching an ADSL modem on each end of a twisted-pair telephone line creates an ADSL circuit. This ADSL circuit is split into three channels by a device known as a splitter (see Figure 1). The first channel is the Plain Old Telephone Service (POTS) channel that you use for your voice service. The second is a high-bandwidth downstream channel, and the third is a full duplex channel used primarily for upstream traffic. The POTS channel is separated from the other channels so that you can use your phone or fax machine while you surf the ’net with the other channels. Separating the POTS channel also ensures that, if your ADSL service fails, the POTS service will remain uninterrupted. Using current technologies, ADSL is capable of delivering up to 8 Mbps downstream and up to 1 Mbps upstream—hence, the name Asymmetric DSL because the data rates are uneven, or asymmetrical. By comparison, a T- 1 line is symmetrical and generally delivers 1.544 Mbps both upstream and downstream.

The Different DSL Technologies

Each type of DSL has its own bandwidth limits and operating characteristics (as shown in Figure 2). I have already described a little about ADSL, so now I’ll introduce the rest of the xDSL crew:

• High-bit-rate DSL (HDSL) is the oldest of the DSL technologies. HDSL is used mostly by telephone companies and requires two twisted pairs (i.e., two phone lines). The phone companies use HDSL mostly to set up T-1/E-1 links for public and private networks. HDSL delivers 1.5 Mbps upstream and downstream.

• Rate-adaptive DSL (RADSL) is a hybrid type of ADSL that makes is possible for the modems on either end of a connection to use a “handshake” similar to that of existing modems to automatically adjust their transmission speed. RADSL provides about the same bandwidth as ADSL (8 Mbps downstream and 1 Mbps upstream).

• Single-line DSL (SDSL) is a modified HDSL technology. It is designed to provide 1.5 Mbps in both directions over a single line. Currently, this technology is limited to distances less than 8,000 feet from the Point of Presence (POP). The POP is generally the nearest telephone company Central Office (CO).

• Very-high-bit-rate DSL (VDSL) is similar to SDSL but offers up to 25 Mbps downstream and 3 Mbps upstream. VDSL is the newest of the xDSL technologies, and there is a lot of talk going on about how it could be implemented. But, thus far, it’s only talk. The problem that needs to be addressed before there is widespread use of VDSL is that it is limited to less than 1,000 feet from the POP.

• DSL.Lite (also known as G.Lite) is a lower-speed ADSL technology that is being designed by a group of companies (including Microsoft, Intel, Compaq, and all the regional Bell companies) who are hoping to deliver ADSL-like service to consumers without some of the deployment issues associated with full-speed ADSL. This group, known as the Universal ADSL Working Group (UAWG), has set a Christmas 1998 deadline for itself to ship PCs equipped with DSL.Lite modems. It has also proposed a

preliminary set of operating standards for DSL.Lite that was released in August. DSL.Lite is capable of delivering up to 1 Mbps downstream and roughly 384 Kbps upstream.

All of the xDSL technologies have distinct advantages over other existing high- bandwidth solutions. The most obvious of these is speed. But there are also some not-so- obvious advantages. xDSL is a dedicated circuit, so you are always “connected.” There’s no need to dial up. There are no per-minute charges associated with xDSL like those normally associated with ISDN. Lastly, xDSL doesn’t “share” bandwidth with other subscribers as with a cable modem.

A word of warning: Some vendors are attempting to use the popularity of xDSL products to market their existing ISDN technology. IDSL is simply a dedicated ISDN circuit, so don’t be fooled by the name. It’s the same old ISDN in a new package.

Limitations of xDSL

One of the drawbacks of xDSL technologies is that they are distance sensitive. In essence, the longer the copper wire they are transmitted on, the less capacity that can be delivered to the user. For instance, VDSL can deliver up to 25 Mbps downstream and 3 Mbps upstream. But it will only operate at that capacity up to about 1,000 feet from the POP. This is the most extreme case, but it illustrates the point. ADSL works best at distances from 1,000 to 12,000 feet (about 2.5 miles). At that distance, it will deliver up to 8 Mbps downstream and 1 Mbps upstream. Between 12,000 and 18,000 feet (2.5 to 3.5 miles), ADSL will deliver 1.5 to 2 Mbps downstream and 640 Kbps upstream. As of this writing, ADSL will not work beyond 18,000 feet, although I have heard rumors about someone having extended the distance out to 22,000 feet. Some of the brightest minds in the world are busily trying to overcome this limitation, so I expect that a solution will be found soon. In fact, there is a solution available today, but it’s too expensive to be practical. By using repeaters and/or amplifiers that restore the signal quality, the distance can be extended, but installing repeaters is a very expensive process.

As I mentioned before, one issue that the UAWG group is trying to address is ADSL/DSL.Lite deployment (check http://www. uawg.org for details). Currently, in order to install ADSL at a customer location, a technician must first install a splitter to protect the data and voice services from interfering with each other. One of the advantages of the DSL.Lite product is that it doesn’t require a splitter. One of the drawbacks with both ADSL and DSL Lite is that every other product that is plugged into the phone line may need filters installed to avoid modem interference. The more likely scenario is that these products will be replaced with new appliances that have the filters. At first glance, this replacement process doesn’t seem to be much of a problem. But if your installation requires filters, think of all the things that can be attached to your phone line: alarm systems, fax machines, phones, answering machines, etc. It could get pretty expensive.

Compatibility is not an issue for users. Lately, there has been a lot of press about xDSL incompatibility. For users, the incompatibility between the different types of xDSL is not normally a problem. It is only a problem if you purchase the xDSL modem hardware and then move to another area where the telephone company (or other vendor) is using a different type of xDSL. Many of the companies implementing xDSL technologies will allow you to lease or rent the necessary equipment and include that cost in your monthly bill. So, unless you plan on purchasing your own equipment, which will cost you anywhere from $300 to $500, incompatibility problems needn’t concern you. The incompatibility issues are for the carriers and hardware vendors to worry about.

Internet Access: Can ADSL Really Deliver?

If you are going to use xDSL for your connection to the Internet, then you should consider the topology of the Internet backbone. The bandwidth available to you when you are connected to the Internet follows a weakest-link rule (i.e., your bandwidth is only as high as the lowest bandwidth connection between you and whatever point you are attempting to access). Most of the smaller Internet service providers are using a T-1, or possibly several T-1s, to connect to the Internet backbone.

Now, think of your connection to your ISP as a pipe (as you’ve probably seen in advertisements); as you increase the bandwidth, the pipe gets bigger. The ADSL connection can be compared to having two pipes. The ADSL downstream pipe will be roughly 4 and 1/2 times bigger than the downstream pipe connecting your ISP to the Internet backbone (if your ISP has only a single T-1 connection), and the ADSL upstream pipe will be about 3/4 the size of your ISP’s. Since the majority of Internet traffic is downstream, the problem is that you have a bigger pipe than your ISP does. So, even though you have the capacity to receive 8 Mbps, your ISP is only capable of delivering a maximum of 1.544 Mbps (over a single T-1), and that capacity is shared by all the other users on the ISP’s network.

So, in actuality, you are still going to be limited to whatever capacity your ISP is capable of delivering. This issue extends to anywhere you attempt to access. If you are trying to download a file from a server in India that is connected to the Internet backbone over a 56-Kbps line, then you are obviously going to be limited to 56 Kbps. Even given this type of limitation, an ADSL connection is going to give far better performance than a 56-Kbps modem or even an ISDN line.

If you are considering ADSL as an intranet backbone, I say “If it’s available, use it!” The issues regarding the weakest link still apply but to a much lesser degree. In setting up an intranet, the entire network (in theory) is under your control, so you can get more bandwidth where necessary and/or replace existing T-1/E-1 links with ADSL links that will probably be less expensive anyway.

What Will It Cost Me?

Even a cursory examination of ADSL pricing will leave you confused. It certainly left me confused until I thought about some of the telecommunication economics involved with a low-cost, high-bandwidth product. Right now, T-1 access to the Internet with
1.544-Mbps access costs somewhere between $1,200 and $2,300 a month. That price doesn’t include the equipment or setup costs. So, here comes ADSL, which is being offered in some areas for as low as $59.95 a month. It doesn’t take an economics degree to see that someone is not going to be especially happy about this. As a result, ADSL pricing is all over the board. I have seen advertised prices anywhere from $59.95 a month to $990.00 a month. In addition to the monthly charges, there will most likely be a setup or installation charge. This charge also varies from provider to provider. Depending on what is included in the setup/installation charge, it could range from $125 to $900.

Because there are normally two parties involved with ADSL, the Local Exchange Company (LEC) and the ISP, the advertised prices are not always what they seem. There are several questions to ask when you are pricing ADSL service. First and foremost, ask if ADSL is even available in your area. This will help you avoid wasting your time with the other questions. Second, find out if the setup price includes hardware, installation, and/or inside wire. Third, if you are getting ADSL from an ISP, then ask if the monthly price includes LEC charges. The reverse is also true. That is, if you get ADSL from the LEC, then find out if Internet access is included in the monthly price.

To Regulate or Not to Regulate

Normally, I am not one who pays a great deal of attention to what goes on in Washington, D.C., but in the case of communications, I am very interested. I know that laws and “legalese” are pretty dry stuff, so I’ll try to keep it short. Most of you are probably familiar with the Telecommunications Act of 1996. If you are not, then you may want to take a look at it at the Federal Communications Commission (FCC) Web site at http://www.fcc.gov.

Before I continue, you must understand the difference between Incumbent Local Exchange Carriers (ILECs) and Competitive Local Exchange Carriers (CLECs). The ILECs are the companies that were federally regulated monopolies prior to the Telecommunications Act of 1996. These companies include GTE and all the Bell companies (Pacific Bell, Nevada Bell, Bell Atlantic, etc.). CLECs are the companies that

have been competing with the ILECs since the Telecommunications Act ended the monopoly on local telephone service. The ILECs and the CLECs are in the midst of a battle based on Section 706 of the Act. Section 706 states that the state Public Utilities Commissions and the FCC are to be partners in encouraging communications carriers to deploy advanced/high-bandwidth communications capabilities. It seems like a simple proposition: Encourage both ILECs and CLECs and create a healthy competitive environment for both of them. But meeting these goals is trickier than it seems.

The ILECs, such as U.S. West and Bell Atlantic, are seeking “regulatory relief” to encourage their deployment of the types of services referred to in Section 706. This request is just a fancy way of saying that they: a) want to be allowed to provide these services in areas where they are not currently allowed to, and b) do not want to be forced to make these services available to resellers at a wholesale discount. Again, this sounds reasonable, but it doesn’t provide for a competitive environment, which is required by the Telecommunications Act.

The CLECs are not happy with the current situation, either. They are arguing (to the FCC) that in order to compete effectively with the ILECs, they need better access to the ILEC-controlled local loops and better collocation rights. This request also makes sense on its face. The CLECs are saying that they want access to the twisted-pair wire that is currently controlled by the ILECs. But if the ILECs are forced to provide this access, then they lose their marketing advantage ($100 billion of twisted-pair wire already in place). So it goes around and around. On August 6, the FCC ruled in favor of forcing the ILECs to give better access to the CLECs and also preventing the ILECs from delivering high- bandwidth data services without making these services available to resellers at a discount. The ILECs have appealed the ruling, so the battle is far from over. The ILECs will still be allowed to provide these services, but they will have to do so through a subsidiary company.

I have some strong opinions on these issues, but I will spare you from them. Let me just say that the ILECs have done nothing to convince me that they want anything other than to preserve, as much as possible, the monopoly that existed for them until the Telecommunications Act was passed in 1996. (I told you that I had strong opinions.) Anyone who is interested in seeing high-speed data services deployed rapidly and effectively should be paying close attention to the legal issues concerning how and when the resolution of this situation will take place.

The Final Word

The xDSL technology, like most modern marvels, will probably get simpler to install and set up, and it will probably cost less as it becomes more widely accepted and deployed. On the residential front, I personally feel that it is still a little too pricey if you are just going to surf the ’net. If you have a home-based business and want to host your own Web site, then the price is more than reasonable. If you are an IT professional looking for a cost-effective Internet/intranet access solution, then ADSL could be the way to go, if it’s available (see the sidebar, “ADSL and the AS/400”). Right now, availability is the key. For those of you fortunate (or perhaps not so fortunate) enough to live in densely populated areas, ADSL will be available sooner than later. For us who live in outlying areas, it may not be available for a year or two. You can find out more about when and where ADSL is being deployed at the ADSL Forum Web site at http://www.adsl.com.

ADSL and the AS/400

At first glance, the promise of ADSL may not seem too exciting to the AS/400 crowd. But if you examine where IBM is trying to position the AS/400, you’ll find that there is much to be gained from ADSL. IBM’s marketing strategy for the AS/400 seems to be twofold.

First, the AS/400 will be an e-commerce server. Second, it will provide the back-office support for the Java-based NCs like the IBM Network Station 1000. With that in mind, it is easy to see how implementing ADSL or any low-cost, high-bandwidth network solution could be significant to AS/400 shops.

Using ADSL is very similar to connecting the AS/400 to a PC network. In fact, the same equipment is needed. The setup will vary depending on the specifics of your implementation. The ADSL circuit comes into a network concentrator, and the AS/400 needs to have a Network Interface Card (NIC) of either 10 Mbps or 100 Mbps.

Once the AS/400 is connected to the network and ADSL circuit is hung off the concentrator, you’re done. Whatever is on the other end of the ADSL circuit, be it the Internet or your own intranet, is on your network.

POTS Upstream Channel Downstream Channel

3.4 30 138 1104 (kHz)

Figure 1: ADSL bandwidth is divided into three channels

xDSL type Downstream Distance

Data Rate from POP

ADSL 1.5 to 8.0 Mbps 0 to 12,000 ft.

Up to 768 Kbps 12,000 to 18,000 ft. HDSL 1.5 Mbps 0 to 12,000 ft. (requires two twisted pairs) SDSL 1.5 Mbps 0 to 8,000 ft.

RADSL 1.5 to 8.0 Mbps 0 to 12,000 ft.

Up to 768 Kbps 12,000 to 18,000 ft. VDSL Up to 25 Mbps 0 to 1,000 ft. DSL.Lite Up to 1.0 Mbps 0 to 18,000 ft.

Figure 2: Here’s how the different flavors of xDSL stack up against each other

BLOG COMMENTS POWERED BY DISQUS