TELECOMS - "Hello, I'd like more bandwidth".

One of the inevitable burdens that FDs have to shoulder in these high-tech times involves making judgements about the validity of spending plans for technologies the existence of which they were blissfully ignorant – until someone asked to spend money on them. Unless an FD has acquired the habit of scanning the computer press on a daily basis, the phrase “voice over IP” (VoIP) will convey nothing at all. In fact, this little phrase signals one of the most dramatic shifts in telephony for decades. FDs will come to relish the sound of VoIP, as it equates to substantial cash savings. The basic concept behind VoIP is relatively simple. Over the years, medium and large organisations have got into the habit of paying for leased lines to allow data to be moved efficiently between their various sites. For these lines they paid a flat fee and they then had the ability to move as much data to and fro down that “pipe” as its data carrying capacity, or bandwidth, would permit. The disadvantage in this set up was that the protocols (the basic set of rules governing the language of data transfer) required for moving data were not the same as those required for telephone conversations, so the data lines were limited to carrying data only. Hence, companies expected to pay separately for telephone conversations between their various offices, even if these sites were already connected by a leased line data network. What VoIP brings to the party is the ability for these same data lines to carry telephone conversations for free. Suddenly, a business case that was fully justified on a particular count (data), now brings additional advantages (voice), for a relatively modest additional spend. Once this is done, inter-office conversations “ride for free” for all time on the data lines. If those calls are relatively long distance over a wide area network (Wan – opposed to a local area network (Lan)), and if the volumes are substantial, the resulting cost savings can be very attractive. What has made all this possible is the increasingly widespread use of a microchip called the digital signal processor (DSP). This allows voice, video, fax and other analogue signals to be processed into a variety of digital formats. For those who like a little technical detail, the “IP” part of VoIP, stands for Internet protocol, and, as the name suggests, this is also the protocol that runs the Internet. The full name of the protocol used in VoIP services is TCP/IP, where the TCP part stands for transport control protocol. TCP/IP is just one of a number of data communications protocols, the other most successful examples being frame relay and ATM – asynchronous transfer mode. Suppliers such as Micom, a division of Northern Telecom, have been providing voice over frame relay for a decade or so, combining both voice and data. However, what makes all the difference with TCP/IP is that it is rapidly becoming the data communications protocol of choice for organisations around the globe. Micom, for example, points out that around 90% of organisations will be using TCP/IP by the end of next year. Adding VoIP services to an existing TCP/IP data network infrastructure is relatively straightforward. The presence of such a widespread, standardised network, together with the capabilities offered by the global Internet, is a powerful factor in creating the conditions for an explosion in the VoIP market over the next few years. Leslie Hansen, Nortel Micom’s US director of international marketing, reckons that since his company launched its VoIP services in 1996, it has shipped some 10,000 ports. “The numbers are still in the early adopter phase, but that phase is coming to an end,” he says. Other suppliers in the area are VocalTec/Dialogics and Vienna, owned by Newbridge. “In all, there are probably about 250,000 IP telephony users in organisations right now, but the numbers are rising all the time,” Hansen says. “We have seen people with savings of around $250,000 a quarter, with payback for implementing VoIP coming in six to nine months,” he adds. “Right now there is probably not a major telecoms carrier anywhere in the world that isn’t offering or thinking of adding VoIP capabilities to its network.” However, Hansen does not believe that pure cost savings from intra-organisational calls are the long-term future of VoIP. The limiting factor is that at some point the available bandwidth on the data network is bound to fill up. Then once the costs of upgrading the network are taken into account, the cost equation for implementing VoIP starts to look very different. The real long-term advantage, he claims, lies in the multi-media capabilities of mixed media IP Wan/Lan networks. “Multi-media is what will justify corporates investing in major network upgrades, rather than simply adding VoIP kit to their existing data networks,” he suggests. Avi Oron, vice president of R&D at TeleLynk points out that his company has yet another variation to offer on the theme of free or more cost-efficient use of corporate telephones using IP. TeleLynk’s system uses what Oron calls a Dial Server, or a “phone to IP resolver”, which translates telephone numbers into IP addresses, in combination with a piece of client software on each end-user’s PC. Users can then make calls to each other using the company’s existing Lan/Wan services, much as they would make ordinary telephone calls. Any user sitting at his or her PC can see immediately, from a tree-based list on the screen, if the user they want to talk to is engaged or available. Using a regular 64Kbps (64,000 bits per second) routing channel, the system can cope with up to 12 simultaneous calls, while a typical Lan can cope with 100 simultaneous calls. One of the unusual features of this system, according to Oron, is that TeleLynk has solved the problem of giving VoIP all the usual telephony trimmings, such as ringing, dial tone and engaged signals, which are often missing from IP telephony solutions. Possibly of more interest to FDs will be the fact that router-based systems such as TeleLynk makes it possible for users in, say, the south of England to route Midlands or north of England calls to external parties (say to clients or suppliers) via their branch office in that area. As Oron points out, the call travels for free on the company’s own data network, then is connected to the public sector network via a gateway at the relevant branch office. This means that the south of England speaker is in effect contacting someone at the other end of the country for the price of a local call. Such systems offer efficiencies in other areas, too. An executive working from home with the client software resident on his home PC can connect to the router in the office and then dial out anywhere in the world, if they have international calling privileges. The billing is automatically registered to the office phone and not to their private line, eliminating all the paper work associated with recouping the expense of business calls from home. Nor is VoIP the only transformation to hit the corporate telephone network. A rash of developments, many having to do with the new communications possibilities created by the World Wide Web, have sprung up. Take the traditional PABX exchange. Instead of that neatly designed, relatively expensive box sitting on the receptionist’s desk with its buttons and array of tiny, flashing red lights, the receptionist is likely, henceforth, to be sitting at an ordinary computer screen (assuming your company still has a receptionist!). Suppliers such as Mitel have found a way of collapsing all the functions of the PABX into an NT server, the same box that runs your file and print services. In many cases, of course, the receptionist too, has vanished into the machine, to be replaced by a structured string of choices. You can hardly dial a company these days without being greeted by some computer voice saying: “Press one for sales, two for support, three if you feel like strangling this messaging system …” The technological transformation has affected even what FDs would think of as traditional corporate telecommunications from national carriers such as BT. Here too, on the public telecoms networks, the introduction of new technologies has paved the way for a whole new range of telecommunications services. BT’s Robin Wiles points out that the move to virtual rather than actual circuits on BT’s own national ATM network has transformed the pricing structure for private data networks. The big revolution introduced by virtual circuits instead of the usual switched circuits, is that bandwidth can be configured according to demand, and scaled upwards or downwards. At a stroke this frees companies from having to live with the consequences of under or over estimating their bandwidth requirements. As Wiles explains, the new service, called CellStream, is not geared for time-sensitive traffic, such as voice (no one likes to partake in a chopped-up conversation, where some bits take longer, and occasionally considerably longer, than others to arrive). Like the Internet, CellStream is good at guaranteeing that data will get from point A to point B, but it doesn’t offer what are known as quality of service (QoS) guarantees, which means that a QoS-dependent service such as voice is not suited to CellStream in its present incarnation. Instead, CellStream’s main claim to fame is the way users can add incremental chunks of bandwidth, in 0.2 mega bits per second (Mbps) chunks. They can buy into the service at a starting point of 0.2Mbps, then go on up, as the need arises, to a maximum of 25Mbps. “This service is designed for customers with a regular need to send data between known sites,” Wiles explains. “If they need to add a new site to their Wan, the system is automatically reconfigured to accommodate the new connection point,” he says. BT guarantees to add extra bandwidth to a client’s virtual private circuits within a day nationally, and within three days internationally. For users who need rather less bandwidth than CellStream offers, BT now has a service called FrameStream, (its implementation of frame relay). This begins at 64Kbps and goes to a maximum of 2Mbps. Framestream can be linked to CellStream. Up to now BT has offered FrameStream solely as a data network, but as Wiles points out, at Network 98, BT demonstrated its voice over frame relay option, so this, too, can be expected as an option shortly. FrameStream is distance-independent, which means that the tariff is based on bandwidth usage, not on the geographic distance between points. FDs are likely to find their data network managers pressing for ways of connecting local area networks, which typically operate at a shared 10Mbps or shared 100Mpbs speed, to the Wan to give offices at different locations an inter-networking capability. Having the tariff calculated on a distance-independent basis can be a major consideration in such instances. Telephony can be a complicated issue to sift through for FDs when judging the merits of a business case. However, communications are now so central to corporate success that spending some time getting a feel for the issues and technologies involved will pay dividends for even the busiest FD. One thing FDs can be sure of, there is no shortage of suppliers falling over themselves to provide presentations on any and all of these new services. Don’t hang up.