The Shape of Things to Come

Convergence, if it hasn’t already sunk in, is well and truly upon us. The old distinctions between telecommunications and networking — voice and long haul on the one hand, corporate data networks on the other — means less and less every day, and soon will mean next to nothing.

That was one message that came through loud and clear recently when CIO Canada convened a virtual panel of telecom industry gurus. We asked Bob Singleton, Raju Rishi and Bill Conner to gaze into their crystal balls and tell us what CIOs will or should be losing sleep over in the months and years ahead. Mr. Singleton is Cisco Systems’s Toronto-based director of systems engineering; Mr. Rishi is director for international IP solutions at Lucent’s Bell Laboratories in New Jersey; and Mr. Conner is Nortel Networks’s president of enterprise solutions in Dallas.

Mostly we heard about the blurring of boundaries. It’s clear that convergence — of data and voice networks, of WAN and LAN — means the old divisions within the enterprise must come down too, if they haven’t already. Indeed, for some CIOs, dismantling duplicate, overlapping telecom and IT organizations should be a top priority. But as our panellists make clear, doing that only lays the groundwork.

BOB SINGLETON, CISCO SYSTEMS — Creating the Vision of the VPN

We originally asked our panellists to predict the technologies and trends that would impact the enterprise two, three and five years out. The first thing Bob Singleton tells us when we hook up with him by phone from his car somewhere in deepest Toronto: forget the five-year forecast.

“It’s pretty hard to look more than two to three years out these days,” Singleton says. Things are moving too fast in the networking world, and there’s too much uncertainty. “But I can tell you what we see in the next 12 to 36 months.”

Singleton and Cisco figure there are three emerging network technologies that directly or indirectly will fundamentally change the way many enterprises operate over the next few years: virtual private networks (VPNs), voice over IP, and broadband access. It goes without saying that Cisco has a finger in all three pies.

VPNs: VPNs, private networks created using public wires — including the Internet — will be vital because they reduce or eliminate the need to build expensive wide area network infrastructure, and they help open up important new channels to suppliers, customers, telecommuters and travelling employees.

“I think VPNs will be important to the future success of a lot of companies,” Singleton says. “They’re obviously part of the transition from the old- to the new-world economy. VPNs are very much new-world technology.”

VPNs use encryption and other security mechanisms to ensure that only authorized users can access the network and that the data cannot be intercepted. It’s as if corporate data travels in a protected tunnel or conduit through the public infrastructure.

What’s holding VPNs back? “People are still fussing about security and privacy,” is the way Singleton puts it.

He makes it clear this is more a matter of perception than the actual state of the technology. The technology, he insists, is mature and well tested. All the pieces are in place: client, server, network. “It’s just a question of corporations tying them all together to create the vision of the VPN. And that’s what’s starting to happen today.”

The real pay-off will come for companies that Internet-enable key applications such as inventory control and supply chain management and then use VPN technology to forge new electronic links with suppliers and customers, he says. It will allow them to reduce administrative costs, turn-around times and overall time to market for the company’s own services and products.

IP Voice: “IP voice will have a huge impact on the enterprise,” Singleton says. Some companies are already beginning to replace traditional PBXs and analogue phone networks with phone systems that move voice in IP packets over existing data networks.

IP phone systems can virtually eliminate administrative and maintenance costs related to moves, adds and changes — it’s possible to unplug an IP phone in one office and plug it in in another without making any alterations to software or hardware. They also reduce costs by eliminating redundant network infrastructures.

A cost of ownership study conducted for Cisco by DMR Consulting Group and Telecom Applications Research Alliance, both of Halifax, found IP voice systems were 20 per cent to 30 per cent cheaper to operate than PSTN-based PBXs or phone company Centrex services.

Perhaps more important, Singleton says, IP voice enables new applications such as video conferencing, unified messaging that consolidates e-mail, fax, voice and even video messages in one mailbox, and Web “push” applications that deliver timely information to screen-equipped IP phones.

Hold-backs include concerns about voice posing a network traffic problem — unjustified, Singleton says — and the perception that using IP voice means making compromises on connection and voice quality due to delays in delivering voice packets. Quality of Service (QoS) technology solves those network latency problems, Singleton says.

Companies that are moving into a new building, have a PBX lease expiring or have run out of capacity on their current phone system should all be looking at IP voice systems over the next two years, he argues.

Broadband: VPNs and IP voice are just two of the most critical applications driving demand for network bandwidth — both in the enterprise network and beyond. There are others. “E-training is starting to become an interesting play,” Singleton notes. Interest in video and multimedia conferencing is slowly growing. And more and more companies are using the Internet or intranets to transmit live video for corporate communications.

“One thing we’ve learned,” Singleton says, “is that networks never get smaller, because applications continue to drive more bandwidth requirements. To address that, CIOs need to be looking for higher-speed access.” That may mean optical technology between LAN or WAN and carriers, DSL and cable or wireless alternatives for telecommuters and branch offices.

Indeed, capacity planning — ensuring availability and reliability of network resources and services — has become a more and more important part of a CIO’s agenda. “If you want to talk about what should be keeping the CIO awake at night,” Singleton says, “that’s a big one for sure.”

RAJU RISHI, LUCENT’S BELL LABS — The Future is Wireless

Working at Bell Labs, the R&D mecca where transistors and the laser were invented, puts Raju Rishi in perfect position to view the road ahead. Rishi sees four often interrelated trends that will have a major impact on enterprises over the next few years: wirelessness, the “virtualization” of the office, improved user interfaces and the rise of “the collaborative environment.”

Wirelessness: A number of technical breakthroughs have already made wireless devices commonplace. But that process will accelerate in the future, Rishi says. Miniaturization — with a kind of Moore’s Law in effect that reduces the size of communicating devices by a quarter every five years — is one factor. The opening of new radio spectrum is another. Add increased battery wattage and reduced power consumption in portable devices and you have a recipe for virtually universal wirelessness.

“Anything wired now,” says Rishi, “will become free of wires in the future.”

That will extend from today’s commercial realm of cell and PCS phones and communicating PDAs into the enterprise as wireless LANs, PBXs and multiservice IP networks become more feasible. They’ll feature phones that share all attributes and functions of a wired phone but can be carried anywhere in a building or campus, and laptop computers that are always connected to the network wherever they are.

It’s true that wireless bandwidth will always lag behind wired networks but demand for increased bandwidth is driven by just 20 per cent of the network-using population, Rishi notes. Eleven megabits per second (Mbps) shared on a wireless network is possible today; 25Mbps wireless nets are six to 12 months away. “For most end users,” he points out, “that is more than sufficient.”

The Virtual Office: The virtualization of the office — people working from homes, widely distributed satellite offices, or from wherever they happen to find themselves — is something else CIOs should be planning for now, Rishi says. Virtualization will bring crucial benefits such as cost savings on real estate, increased productivity and improved employee retention.

Once again, it’s a series of technological developments that will make it more feasible in the future. One, clearly, is higher bandwidth to the home, thanks to DSL, cable and wireless access services. The widespread use of the same easy-to-use Web browser interface in a variety of corporate applications will also make it easier for employees to work unsupervised remotely.

Increased intelligence on the network will make it possible for the network to track employees wherever they go and deliver calls, mail and information without the kinds of delays and clumsy programming required today. Application service providers (ASPs) may even offer virtual PBXs in the future. “We haven’t seen that yet,” Rishu says. “But don’t be surprised to see it in the next six to twelve months.”

Finally, the rise of the storage area network — mirror databases strategically located around the country to optimize access response times and network costs — will help make it easier to break the ties that bind companies to one location.

Intelligent Interfaces: Rishi sees speech recognition — a technology with which Bell Labs has long been associated — and artificial intelligence (AI) as key enablers for new user interfaces that will revolutionize telecom and computing in the future.

The use of artificial intelligence is fuelling “tremendous advances” in speech recognition that will soon make it much easier to “train” speech systems capable of virtually 100-per-cent accuracy, Rishi says. Computer users will exploit the technology first to issue voice-based “hot-key” commands that perform multi-step tasks, he predicts, but improved dictation systems will also emerge.

There may be cultural barriers to the widespread use of speech recognition in offices, he concedes — the distraction of people babbling at their computers in an open concept office — but wireless headsets could solve that problem. Having a roomful of people murmuring into headsets seems to work fine in call centres, Rishi notes.

Other natural applications for speech technology:

• enabling voice-response systems to navigate an internal phone or information system, replacing touchtone interfaces that often frustrate customers and other callers;

• “media conversion” — translating voice mail to text for easier delivery.

Looking even further out, AI will be used in expert agents that use data mining techniques to predict the likelihood of finding a person at a given network address, for example, or data visualization tools to analyse and make recommendations about provisioning network or call centre resources.

The Collaborative Environment: In the virtual enterprise of the future, how will people work together? In electronic collaborative environments, Rishi says.

These collaborative environments will be created using products such as Lucent’s prototypical Definity Anywhere. They make it possible for members of a virtual corporation distributed across the country to meet, talk, share data, work interactively with the same computing applications, and even see each other in living colour — although Rishi concedes “there isn’t as much demand so far for video.”

The data portion of the network traffic generated by such collaborative environments is relatively small, but the multimedia components will drive demand for more and more bandwidth.

Indeed, one of the biggest challenges for CIOs in the next few years, says Rishi, will be establishing network traffic priority schemes and utilization policies to keep traffic manageable and control costs. This becomes increasingly important as separate voice and data networks collapse into a single multiservice network.

F. WILLIAM CONNER, NORTEL — Revolution in Network Architecture

We asked Bill Conner to talk about new technologies that would impact the enterprise, and he provided input on two technologies that Nortel is extensively involved in: Open IP and the Optical Internet.

Open IP: Open IP is Nortel’s set of software enabling tools for embedding routing intelligence in all manner of devices, from the tiniest wireless hand-held “IP appliances” to the latest terabit data switches. “I think we’ll see an absolute revolution in network architecture,” Conner says. “You can already see the momentum starting to build. Stay tuned and watch this space.”

It’s true that Nortel has already enlisted something like 200 partners in its Open IP initiative, including heavy-hitters such as Hewlett-Packard, Intel, Motorola and Microsoft. But it may be early to talk about a revolution. Some pundits, in fact, say Open IP is an attempted end-run around arch-rival, routing market leader Cisco Systems.

Open IP (OIP) is a set of low-level programming tools and code modules that Nortel will make available to any developer. They will allow manufacturers to embed routing intelligence in hardware, processors and custom chips. “We asked ourselves,” Conner says of the conceptual starting point for Open IP, “how do we take routing code and unbundle it, make it modular, make it usable in a distributed architecture?” Open IP is the answer.

One impetus for the project, he says, was that today’s hardware routers, even those that drive the supposedly open Internet, continue to use proprietary protocols (mainly Cisco’s). The result is that, contrary to the usual industry trend, router prices have actually risen over the past several years.

Any OIP-enabled device — wherever it is in the network, including at the end points — could route data, including voice and video in multiservice networks. This translates to higher data throughput, better bandwidth utilization and greater reliability, Conner says. It could also mean lower administrative and maintenance costs for IT departments no longer so reliant on mega-hardware routers.

The most important benefit, Conner says, is that Open IP will enable new kinds of highly tailored services, many as yet unimagined. A field service technician, for example, could use a PCS phone to dial into a company intranet. Intelligence in the phone would facilitate routing Web-based information to the PCS screen, providing just-in-time procedural instructions.

“The service technician doesn’t even have to know all the service steps required,” Conner says. “The system will push information to him — this is step one, this is step two — including the customer’s service history. And it’s all screen- and voice-enabled. It’s all showing up on a PCS screen.”

Open IP is open in the same sense as Sun Microsystems’s Java programming language. Nortel will be the sole owner of OIP Environment modules and frameworks and will sell them to developers, but developers will get source code and are free to modify it, although modifications will not necessarily be integrated into the core product.

Will Open IP spark a revolution in networking? As Conner, says, “stay tuned and watch this space.” Analysts we consulted are cautious but believe Open IP could, indeed, have significant long-term implications. That said, there is very little CIOs can do about Open IP now, except be aware that manufacturers will be offering products in the future that could provide alternatives to today’s proprietary routing architectures.

The Optical Internet: The notion of the Optical Internet — extending optical fibre to every layer of the network, including eventually to the individual corporate desktop or home, and using one packet-switching system from end to end — may not be exclusive to Nortel, but the company does have a dominant share of the emerging market for optical switching equipment, Conner points out.

The potential benefits of end-to-end fibre networks include huge increases in bandwidth availability — Nortel has 10 gigabit lasers today and is experimenting with 40- and 80-

gig lasers in its labs — and much simplified network architectures.

The Optical Internet starts with extending fibre from the carrier to the enterprise, but it will move to the desktop in the longer term to support high-speed applications such as voice, video and multimedia.

How far down the road is the Optical Internet? Closer than you think, Conner says. CIOs should be seriously planning for it today.

Tony Martell is a freelance writer specializing in information technology and IT management. He is based in London, Ont.