November 1997

BY STEFAN KNIGHT

There have been a host of articles written about xDSL, many touching on the application of this technology to business and residential access to the Internet. The remote access industry has been burned in the past for touting the great era of fullservice networking to the home. Some of these promises included movies on demand and video phone conferencing to the far corners of the globe. While these are two extreme ends of the spectrum — from the outlandish to the perhaps plausible — the real question remains, will John Q. Public really pay for it?

There is a new twist in this debate, a twist which hinges on CTI, another application that — similar to Video-OnDemand (VOD) and multimedia — invokes a spectrum of reactions as to its viability. This article will discuss the very real use of xDSL to enable CTI applications in the small office home office (SOHO) market and how carriers can deploy and charge for these services.

DON’T BELIEVE THE HYPE
In today’s climate, it seems that every new and improved telecommunications technology is a roller coaster, starting with hype and ending with a sigh of resignation when the “big one” gets away. For years now, many special interest groups have been trying to find a way to penetrate deeper and exploit more fully the residential telecommunications market by promoting broadband services into the home. Each new ray of hope, FTTH, FTTC, ATM, ISDN, cable modems, original ADSL services, satellite, etc., have great technological promise but stall prior to reaching their promised ubiquity. There are many reasons for this, including a lack of a business case, implementation problems, slow-moving standards, and a lack of compelling applications at compelling price points.

The latter is the greatest black hole for the would-be entrepreneur set to embark into the local loop to seek fame and fortune. Residential consumers are prepared to pay for water, electricity, gas, and phone services, but no matter how cool the Internet is or how great video conferencing would be, they are not seen as necessary utilities by the consumer. It is ultimately a question of money, what kinds of services and what tariffs? Tariffs are the key to the success of new technologies and services to the home. We all watch our long-distance bills and turn blue when they approach $100 a month, so who is going to pay more than $19 a month for any service no matter how technologically “cool” or “hip” it may be?

BIG BUSINESS PROVING GROUND
Business, that’s who. The real application that can make use of broadband services to the home, pay for it, and contribute to the bottom line is: turfing office workers out into the virtual workplace of their homes. “Hoteling” is a fashionable concept today and one which is being implemented by real companies. In this model, workers have laptops and docking stations in their home office. They telecommute from home the majority of the time. When they do need to go into the office they reserve a cubicle or office and have their phone routed to this virtual desk. The data side of the connection is handled quite efficiently by having a docking station with a LAN connection for the work-at-home user’s laptop computer. The facts and figures speak for themselves: It costs money to maintain office cubicles, lights, heating, parking, and various support services for office workers. The direct savings to businesses come from reduced operating overhead, productivity gains, and tax incentives from local and federal agencies. The indirect savings are tangible to the community at large and take the form of less road repairs and fewer traffic snarls. We can show a return on investment for work at home applications but we still have not addressed whether working from home as a service is mature enough to offer.

THE HOME OFFICE
Working at home today is not like working in the office. To emulate the office, you need multiple phone lines, fax lines, and data networking lines. Setting up a home office can be trying on the nerves. Ordering multiple POTS lines has never been an easy chore. Ordering ISDN may make you wish you were back on the road to work in a traffic jam. Once the lines are in, the main problem of working from home begins to surface.

How are calls routed from work to the home number? What happens when little Bobby picks up line 2 and says something inappropriate to your best customer? Many corporations still do not believe that a true office environment can be made in the home, even though much of the office work done does not require workers to be geographically collocated. Many workers the world over are glued to terminals and phones eight hours a day in office cubicles. In effect, the only real difference between working in an office cubicle or from home is the equipment and infrastructure. Of course, there are the coffee klatches and free donuts one may miss out on when telecommuting to work several days a week. What would it be worth to have a true office environment at home? What would it be like, and is it feasible with today’s technology and equipment?

The answer to the first question is up for grabs, but the second is within reach. In the ideal work-at-home environment, the worker would make real productivity gains in standard office tasks due to the lack of interruption and improved concentration that isolation affords. Ideally, these gains would not be offset by a lack of business services such as fax, multi-line phone, and simultaneous data access, in real time. The typical corporate office has copy centers, dedicated fax lines, highspeed data access, and a PBX system for phone lines, all on demand as needed. Therein lies the typical work-athome dilemma.

The work-at-home office today typically has a PC, a fax line, and a phone line. These facilities pale in comparison to the typical office fax server, PBX, LAN, and copy center. Some brave souls have installed ISDN for higher speed data access and some companies subsidize T1 access to work-at-home employees — the lucky ones. But as anyone that has ever set up a home office knows, ordering and installing these services is neither easy nor scalable.

ADSL is supposed to take hold of and reshape the work-at-home, remote corporate LAN access market and free us all from the drudgeries of the morning commute. Does ADSL address these issues or are we merely adding to our home office setup nightmare by adding a new type of technology with new installation issues? ADSL does solve these issues when combined with the right transport technology to offer multiple services and guarantee the quality of those services.

ATM OVER ADSL
ATM in ADSL networks is a reality. ATM technology was designed from the ground up to run over various physical mediums such as copper, fiber, wireless, and of course ADSL on top of copper. ATM has the distinct advantage that it was also designed to carry voice, video, and data traffic. These services can be guaranteed at a determined quality using ATM’s rich management features. ATM is a point-to-point technology similar to the existing PSTN. It is no wonder that telcos the world over have embraced ATM and have played a large part in standardizing and promoting it. ADSL works harmoniously with ATM as yet another ATM physical layer, allowing ATM networks to extend across links of single pair copper up to eighteen thousand feet into the home. ATM is synonymous with B-ISDN (Broadband ISDN) and is a natural fit for provisioning broadband services over ADSL links. Each day, we read of a new ADSL trial in the works and proclamations by carriers as to when, how, and how much ADSL they will be rolling out.

An overwhelming majority of carriers have chosen ATM as the lowest cost, most feature rich, scalable backbone network to interconnect ADSL access devices to the core network. ATM has even become fashionable for the multiplexing of frame relay circuits over long hauls and it seems only a matter of time until it completely dominates the WAN. For once, a clearly strategic decision will have tangible benefits to the end user as well as the telco. ATM is a boon to the marketing folks as it will enable a host of voice and video applications as well as data to ride the same infrastructure. Without ATM, ADSL is merely a faster data pipe which does not address the massive transition underway by “so-called” service providers to become “real” service providers as opposed to bandwidth providers. The value add for the next century will be applications and transaction processing versus infrastructure. So how does this all relate to CTI?

ONE MAN’S CTI IS ANOTHER MAN’S...
CTI is an elusive acronym. CTI means many things to many people. To some it means a LAN interface to the PBX which allows the user to interact with the 100+ features of the PBX through a GUI. To others, it means the computer is the phone. And to others it may mean the LAN is the PBX. I prefer a definition of CTI that combines models one and two above.

As we are seeing more and more, the computer is the network. CTI can be the migration from the PBX to a virtual PBX over LAN and WAN, front-ended by telephonyenabled PCs. Now, this presents a number of thankfully diminishing problems as PCs and LANs were traditionally designed to move connectionless, latency insensitive data and not highquality, latency sensitive voice. Enter ATM into the equation.

ATM provides true circuit emulation of voice and guaranteed quality of service contracts for voice channels interspersed with different types of connection oriented and connectionless data services. ATM is the smart multiplexer for voice, video, and data services. Thus it stands to reason that CTI and ATM would walk hand in hand down the aisle to a perfect marriage inside the PC. Well, “stands to reason” does not make it so.

I can hear the IP telephony doubts rankling in the back of your head as you read this. IP telephony is fine for low quality, delay insensitive conversations across the Internet. Quality is definitely suspect as the Internet grows and has bandwidth demands placed on it disproportionately to its rate of bandwidth growth. The next objection is generally, RSVP will solve this by guaranteeing bandwidth. At any one time there is preciously little bandwidth to go around and it remains to be seen if ISPs will utilize RSVP to guarantee higher service provisions for IP telephony over and above other types of data when there is no premium being charged for the service. Like with all good commercial services, IP telephony must present a business case and tariffing model.

All that being said, what business of any size would trust its telephony operation to the Internet? Any takers? I didn’t think so. Quality and reliability are two issues that can not be skirted when trying to roll out a business-wide telephony solution. Here is where we come to the compelling relationship between CTI, ATM, and ADSL. Separately, each is an interesting topic for conversation, together they provide the framework for a “real” business application that changes the work-athome paradigm forever.

CTI OVER ATM OVER ADSL
There are companies today, undoubtedly profiled in this very magazine, that do CTI over ATM with all the quality of pin-drop fidelity. Telephony circuit emulation to the desktop or dropped off an ADSL modem to an ATMenabled phone device is a reality today and at surprisingly low costs when compared against the world of the proprietary, costly PBX. There are real dollars and cents savings advantages to migrating telephony to LANs from PBXs, not to mention tremendous gains in the usability of telephony features when presented through desktop GUIs. The concept boils down to virtualizing the PBX and combining multiple services over a single infrastructure which meets today’s requirements and scales to tomorrow’s. Once these services become part of a comprehensive network (Figure 1) they can be transferred from place to place across the network provided the necessary security mechanisms have been put in place. Here is where the work at home office begins to become a true business application.

HOW DOES IT WORK?
My corporate office deploys a call server which takes in POTS lines on the low end all the way up to fractional T1 on the high end. This server uses stan-dardsbased mechanisms and APIs to process calls and store voice mail. The server is attached to the corporate ATM backbone and thus can emulate voice circuits to ATM-enabled desktops or ATM-enabled telephony distribution points (similar to hubs) for telephones fed by an ATM link. The corporate LAN user interacts with the virtual PBX call management software residing on the server across the network through GUIs. The corporate network has a connection to the WAN which is ATM-based at speeds ranging from T1 and above.

Work-at-home users connect to the corporate LAN resources over the ADSL access network. The work-athome user has an ATM interface in their PC which the telephone connects to. Alternatively the telephone connects directly to the ATMenabled ADSL modem and the last foot to the PC is Ethernet. Both models work, but it is more effective to carry the ATM traffic across the last foot to the PC, especially since PC ATM cards have fallen dramatically in price and the range of services offered over ATM is so much greater. ADSL modems today predominantly connect to the ADSL network via ATM though they may present either Ethernet or ATM or a both ports to the PC. These modems are typically external to the PC. The next generation of modems is under development which will make internal cards a reality. These cards will be predominantly ATM-based and offer native ATM services within the PC. It is also likely that the next generation of ADSL modems will offer direct telephony connections and circuit emulation services over ATM.

This scenario changes the way the work-at-home user views their phone and data resources. Immediately, the user can stop using their POTS line for business calls. The user can have a provisioned on-the-fly, by-thecall server management system from a few to a vast number of telephone extensions. This process is easily administered through a GUI.

Additionally, some of these telephony circuits can be used for fax services. On top of all of that there is enough data bandwidth to access databases in real time, video conference with folks in the office, and send copy jobs to the virtual copy center. Now the work at home environment begins to closely parallel the office environment and new productivity gains aren’t wasted on the inefficiencies of home office equipment and setup.

LECs AND THE LAST FOOT
One can imagine a terrific boon for the Local Exchange Carrier (LEC) as well in this brave new world. The LEC can now offer the consumer multiple phone extensions and data access across a single pair of copper into the home. Each of these extensions and services can be tariffed and contribute new revenue streams. The beauty of this scenario is that no new copper needs to be pulled, no old copper needs to be terminated, and aside from the possibility of an initial truck roll to troubleshoot in-home wiring issues, no new labor is required.

This boon can also be the LEC’s bust if they do not act decisively and aggressively to promote ATM in the last foot between the ADSL modem and the PC. It is a foregone conclusion that the next wave of internal modems will be predominantly ATM-based but there is still a turf war brewing for the last foot between external modems and the PC. Technically, there is a very good reason to have an end-to-end ATM architecture. Ethernet in the last foot relegates the services provided to those which tolerate the latency introduced by variable length datagrams being routed or bridged at the modem between ATM and Ethernet. The implications of this simply mean that real-time, constant bit rate applications lack any real quality of service. The reason that this can be a bust is that the market which is unfolding is one of new competition for customers. For the first time ever, competitive local exchange carriers (CLECs) can rent a copper pair to the residence. Armed with the ability to offer so many services over this single pair to the home, the LEC that does not choose an extensible ADSL rollout strategy that focuses on services may very well be doomed when pitted against a CLEC that does. Only time will tell if this prophecy is borne out.

CONCLUSION
ADSL is a straightforward technology which can offer broadband data access to remote users over existing copper lines. These attributes make ADSL the compelling page turner that it is. A vast majority of telcos and the equipment vendors that supply them are betting on ADSL to be the big one that doesn’t get away. Equally, the press and analyst community see ISDN’s writing on the wall and will sit back and wait before pronouncing an early victory for this technology. ADSL when combined with ATM’s attributes makes for a compelling rollout model that offers great promise and seemingly loads the deck in favor of ADSL’s ultimate success. What is clear is that services, tariffs, and applications will either make ADSL ubiquitous or relegate it to niche deployments as has happen with ISDN.

CTI will be one of the services which transforms this technology from a bit pump into a powerful business tool that reshapes the way the world works. The PBX of the future will be bought as a combination of hardware and software at the local computer store — much like LANs are today. We will see the evolution away from proprietary telephony exchanges to open, standardsbased systems that ride the same wires as our data traffic. CTI has the promise to be the next true revolution in personal computing and networking. This will also see the transformation of the very nature of these revolutions from primarily quantum leaps in hardware to quantum leaps in services to the end user, a dramatic paradigm shift.

Stefan Knight is director of product marketing at ATML, a provider of complete, silicon-with-software solutions to companies wishing to enter the market quickly with low-cost xDSL (Digital Subscriber Line) products for carrier networks. ATML’s ATOM ASIC (appli-cationspecific integrated circuit) is the first integrated ATM ASIC and software solution for those providing xDSL systems. For more information, visit the company’s Web site at www.atminc.com.