September 2000

BY DAVID MARKOWITZ

Since the world found out about the Web in the early '90s, businesses and consumers have been screaming for faster and more reliable access. A whole new industry has grown up around trying to provide broadband solutions that deliver what customers want without ripping out infrastructures that have been evolving to the tune of billions of dollars since Ma Bell was a girl. Today, we have many flavors of broadband access, and each of them is right for somebody. Unfortunately, what's good for the customer might not be so good for the service provider, and vice versa. There's a lot of talk about the converged networks of the future -- a utopian place, where provisioning on demand exists, and where all services are available to all people, all the time, at prices both they and their providers can afford -- but so far, it has been just talk. What we need is a new approach to convergence that reconciles economic reality with the ever-expanding service and application demands of businesses and consumers.

There are multiple types of broadband access, all of which have evolved at the junction of provider necessity and customer demand. Cable and the many flavors of DSL currently dominate, with wireless rapidly coming up to speed. Each of these transport types, even those under the dubious heading of xDSL, has an appropriate place in the landscape of the last mile, but none of them fully solves the access problem.

Today DSL and cable are the solutions of choice for broadband access in the last mile, but DSL has made more inroads as a data-only service than as a technology used for converged voice, video, and data services. Cable is converging video and data, but the voice piece of the converged services model is still unrealized. CLECs and service providers need a cost effective method to deliver data, video, and voice over the broadband link, and one size doesn't fit all.

In the beginning, broadband meant T1 service, but even today, T1 service costs too much for the average consumer. And even though the price is still falling, T1 and fractional T1 are based on circuit-switched networks that were designed for voice rather than data traffic. DSL and cable technologies emerged because the majority of the potential customer base could not afford the cost of T1 or fractional T1 services. Broadband DSL and cable open up broadband to the masses.

The many flavors of DSL meet different market needs. Variations include:

• Symmetric DSL (SDSL), which has been touted as an ideal protocol for the small to medium enterprise (SME). The symmetrical data flow is best suited to the needs of a business whose broadband usage is bi-directional. SDSL also holds the promise of providing toll quality voice services at significantly lower costs than traditional POTS or Centrex services, ultimately lowering SME operational costs. But SDSL is limited in the distance it can travel (12,000 feet), and that limit is further challenged by the unfortunate state of the copper infrastructure in older areas. A further barrier is that digital loop carriers (DLCs), which have become the de facto standard in new developments, are not unbundled and are restricted to DS-0 channeling protocols.
• IDSL (ISDN-enabled DSL) is a symmetrical DSL service that can be deployed for longer distances (up to 18,000 feet) and is not restricted by the limitations of a DLC, but it's expensive, and at 128 Kbps, it's not really broadband at all. When no other option exists, IDSL is a viable option for data, but it will never have the horsepower to carry voice or video.
• ADSL (Asymmetrical DSL) using the G.Lite standard (G.Lite ADSL) can travel farther than SDSL (up to 18,000 feet) while offering true broadband speeds of 1.5 Mbps. But ADSL offers faster download speeds than upload speeds. Asymmetrical options are valuable for residential users who pull more data down from the Internet or an Intranet than they send up, but it will never be able to carry multiple toll quality voice (VoDSL) communications and therefore loses much of its advantage for the SME.
• Full Rate ADSL has the same VoDSL restrictions as G.Lite ADSL, but it can provide up to eight Mbps of bandwidth in locations that can support its power requirements and fall within its distance limitations.
• To add to the mix, new versions of high-density symmetrical DSL will be coming to market over the next year or so.

CABLE HAS THE BANDWIDTH BUT NOT THE REACH
And then there is cable. Cable is an easier sell to consumers who want broadband access. They already have the cable running past their houses, and many of them already have experience with cable TV. Cable broadband prospects are primarily interested in digital video services and faster Internet access. They probably don't care that cable won't soon replace their telephone connections, because they're not particularly concerned about the cost of their heavily subsidized POTS service.

But cable is generally not available outside of residential areas, and it's unlikely that a cable service provider could justify the cost of laying cable to the light industrial areas where the SMEs tend to locate. Providers that plan to make the effort to lay new infrastructure to such locations are laying fiber. Further, SMEs demand a level of service that cable can't guarantee -- cable uses shared transport, and the more people there are on a particular section of the network, the slower it goes.

MULTI-ACCESS NEEDS, SINGLE-ACCESS BUDGET
Each flavor of broadband access meets a need, but service providers and CLECs must limit their choices in order to control costs. The more broadband options they offer, the more complex their networks become, and the more expertise, equipment, and personnel costs they incur. Limiting options limits operational costs, and today, CLECs and providers must narrow their focus in order to remain profitable. Settling on one broadband option is best for providers, but it means that consumers don't have much choice in the matter. Residential customers might be able to choose one flavor of DSL or cable, for example, but their choices are likely to be limited by what's economically feasible in their area. Fewer choices generally mean higher prices, too.

As a result, we have a fragmenting provider base whose members focus on residential, on SMEs, or on multi-tenant units. Some providers focus on voice or data. Other providers focus on data services, content distribution, or applications hosting, and so on. Some of these providers hope to grow and expand their offerings into more areas, but there are a lot of little guys who'll never do more than one thing at a time, because high volume and narrow focus are frequently incompatible.

For service providers and CLECs, offering converged voice and data services without losing their shirts requires converged platforms that are simply not available today. And at the other end of the market broadband hungry customers are required to know a whole lot more than they should have to in order to make the best choices, and they can't get everything they want in any case.

MANAGING THE OPTIONS
With broadband access options expanding rather than shrinking, the strategy of limiting customer choices won't likely pay off. The communications business is far too robust and competitive to take a step backward, and ripping out infrastructure that, for all its faults, is doing the job in many ways, isn't really an option. Instead, carriers and providers need to find a way to blend and comprehensively manage their various broadband systems in a way that expands customer choices and supports emerging applications without costing too much.

The answer lies in greater intelligence at the edge of the network and in managing that intelligence. A truly converged network is an environment where any voice, data, or video application can be deployed without increasing costs. It's an environment where provisioning is done without truck rolls and where there is centralized, cost-effective management and troubleshooting for all network elements. It's a place where the end user gets one monthly statement for voice, video, and data and where business or residential customers can have broadband their way without putting service providers in the red.

The idea of fully converged networks is pretty simple, but implementation has been much more complicated. Larger equipment providers promise convergence, but they can't deliver because their own internal politics prevent their individual business units from working as a team: One group sees another as a threat, and they are forever comparing their bottom lines and sales figures.

Small equipment manufacturers promise a working product for part of the solution, but then they leave it to providers and CLECs to make a complete system work. A fully converged network is beyond their scope of expertise, and they have more pressing concerns such as keeping up with the Joneses, fighting larger competitors, and trying to stay ahead of the newest set of regulations, deregulations, chipsets, and standards bodies. Not to mention that guy in the "garage" down the street.

Everybody wants to sell a product, but nobody is looking at the whole picture. One vendor says its DSLAM is better; another says its access concentrator saves money. One says IP is the future and won't support ATM, while ILECs (Incumbent Local Exchange Carriers) fret about their ATM-heavy networks. What's really needed is a solution that encompasses all transport mediums and broadband protocols, and that means moving intelligence closer to the point of service delivery.

The only way to offer two adjacent customers a real choice in broadband solutions is to make the local distribution point intelligent and manageable enough to support multiple options and to allow providers to effectively manage and provision those options. Rather than spewing out more point products that address aspects of convergence, carriers and service providers need a transport-agnostic platform that accommodates their existing infrastructures and yet offers fully managed voice, data, and video services. Until we have such a solution, broadband in the last mile will never meet its full potential or the needs of its myriad applications. Choice is as American as apple pie. The question is, who will provide it?

David Markowitz is director of marketing for Zhone Technologies, Inc. Zhone is developing a complete line of telecommunications infrastructure products for the local access network. The company's products bring end-to-end integration and management to the local loop, making it possible for carriers and service providers to offer a full suite of managed services while optimizing bandwidth, reducing costs, and speeding up provisioning.

[ Return To The September 2000 Table Of Contents ]

BY DAVID GIBOR

The key to the data communications revolution is making high-capacity connectivity affordable. Laying fiber comes with enormous costs -- more than $150,000 per mile -- and often requires a long process of getting permission from the various landowners and government agencies responsible for the fiber path. The costs and time to lay fiber are significant factors that block the bandwidth revolution today.

As alternatives to fiber, there are various technologies that help expedite data communications over copper lines (xDSL) or broadband cable. Unfortunately, the communications capacity that these technologies can deliver is limited by the physical characteristics of the existing infrastructure, and therefore, limited to a few Mbps increase in the best case. An alternative that can deliver more bandwidth to customers is wireless deployment. For example, ultra high-capacity broadband wireless systems operating at 622 Mbps are on the horizon.

The advantage of high-capacity wireless deployment is that the systems can provide fiber capacity (>155 Mbps) and quality (BER<10�13) without depending on existing infrastructure. The installation of a wireless system takes a few hours, and it can be moved as demands shift, for example, when a company relocates to a new campus. Immediate deployment, high capacity, and flexibility are the main points that make wireless solutions a prime choice for high-capacity network deployment.

GOING ONCE, GOING TWICE...
Recent auctions in the 39 GHz band are drawing more players into the broadband wireless market. As this global trend continues, demand for broadband wireless technology will increase. CLECs who have purchased spectrum in this and other bands are ready to use their valuable investment, and wireless broadband is becoming a technology of choice.

CLEC networks are also becoming more and more complex. As opposed to earlier days when voice was the prime revenue generator, data networks are taking the lead today. High-speed SONET, IP, or ATM infrastructure is being rolled out, and many of these networks are wireless. New emerging high-capacity wireless systems offer flexibility by supporting multiple transport protocols and interfaces in a single modular system architecture.

Global service providers need a high degree of flexibility to meet different standards, such as those set by the FCC and ETSI, and to operate across a wide range of frequencies. Modular broadband wireless equipment provides this flexibility, allowing for quick deployment of new networks globally. Recognizing potential in international markets, service providers have purchased frequency licenses and are deploying networks in Europe, South America, and the Far East.

The marketplace is forcing service providers to offer more network configurations, greater capacity, and faster speeds -- all at lower cost. To meet these demands, service providers need a seamless technology that provides high-capacity connectivity from the backbone across the metropolitan infrastructure to the last-mile access door. Equally important, the technology must deliver flexibility throughout the providers' networks.

Today, that can be achieved only with fiber or high-capacity broadband wireless. The time and cost savings broadband wireless offers are making it a significant driving force in this fiercely competitive market, helping to resolve the biggest challenges of the communications revolution -- access and affordability. 

David Gibor is director of product management for Giganet, Ltd. Giganet's FibeAir broadband wireless systems operate at 18, 23, 26, 28 (LMDS), and 38 GHz. They are deployed in over 20 countries on five continents.

[ Return To The September 2000 Table Of Contents ]

BY TOM CURRAN

In the past five years since VocalTec set in motion the VoIP revolution, companies and individuals have been scrambling to define, develop, and discover the best solution to most completely capture the true potential of this new reality. Without going into the various gatekeepers, IP-PBXs, switches, routers, and the tunneling or signaling that transpires between the PSTN world and the IP world, let's instead focus on the "last foot," or the actual connection for placing or receiving calls via IP.

These calls can be placed from a business, office, or home via wireless local loop, satellite, cable modem, radio frequency, or even infrared. However, how do these technology transport vehicles actually connect to the phone? A year ago, there were only one or two viable options for connecting the device to the network. Today, the answer is not quite as simple as it was. Serious development efforts have resulted in a wealth of devices flooding the market.

The names of these devices vary. It all depends on who is doing the describing and on their market focus. Internet appliances, terminal adapters, IP phones, network telephones -- these are just a few of the names being applied to this class of device.

Essentially, what began as a PC bundled with a handset and 12-key pad, has evolved into a sophisticated device offering that serves a variety of needs extremely well. There are expensive, sophisticated devices that offer Palm Pilot-like displays for browsing the Web and scrolling through your address book, and simple boxes that connect analog phones to the Ethernet network. Prices range from around $2,000 down to less than $100 depending on the features you're looking for.

Among the problems that edge devices (client-based IP phones) have to overcome is a world where networks are still mixed and standards are still evolving. So when bringing a device to market, you must consider the various realities of that market. If you own the network, then you can take advantage of proprietary phones across your network routers using your own IP-PBX for a great system. But when supplying solutions for developers, telcos, service providers, OEMs, or others, you must consider all of the protocols available today, including H.323, MGCP, Megaco, SIP, and even proprietary ones if that's what the solution calls for.

The need to offer multiple protocols on the same device is obvious: When a call is placed or received you must insure that the same protocol is used to decode the voice packet using the same codecs. This must happen in real time and be invisible to the user.

Now, once you have the packet voice taken care of, you need to consider the environment of the phone or the desktop workspace. Is there one Ethernet port connection, in which case the device must have a built-in hub or switch to connect to the PC or another device such as a printer? Does the customer require PSTN connectivity for lifeline needs? With regard to the internal switch or hub, should you have control over the port priority? Or do you prefer to control packet priority with special flags to bypass other packets?

When analyzing the environment, it is extremely important to consider the visual and tactile experience, i.e., does the customer wish to continue using the same phone they are currently using or do they wish to learn a new fully featured device that can browse the Web and have all features on-board? What type of displays are necessary if any at all?

Today all of these are possible and available. Perhaps the biggest concern regarding VoIP is the actual business model and the relative cost tradeoff between investing in new systems and the depreciation write-off of existing systems. The cost to the end user must be scalable and make business sense relative to their need to switch to a different system.

Fortunately, there are many players in this space today offering a wide range of solutions. The "Last Foot of the Last Mile" used to be a lonely place to be without many choices, but today and for the foreseeable future, the power of choice has shifted squarely to the consumer.

Tom Curran is vice president of marketing and sales for ShelCad Communications, Ltd. ShelCad specializes in the development of innovative CTI, Internet telephony, and VoIP products. Their vanguard products link PCs with any analog phone to support IP calls and added-value features. The company's technological developments include Voice over IP, network computer telephony integration, PCMCIA, Ethernet, and ASIC products. 

[ Return To The September 2000 Table Of Contents ]