Over the last two decades the status and function of telecommunications as an element in the business processes of enterprises of all types has altered almost beyond recognition. What was formerly considered simply as an aid to administration has become a critical element in the quest for corporate efficiency, competitive commercial advantage and superior market understanding. Indeed telecommunications and information technology (IT) are forecast to become the dominant components in the production, sales and delivery of products as diverse as entertainment, financial services, healthcare and security.
Narrowband voice still accounts for the large bulk of corporate communications traffic, and the private branch exchange (PBX), along with its smaller relative the key telephone system (KTS), are ubiquitous in business. Providing free internal calls, and simultaneously reducing the cost of external communications by sharing a limited number of outside lines on the basis that not everyone in a company will make or receive calls at the same time, office switchboards are now viewed as being essential office infrastructure by enterprises of any size. This is a perception which, according to a 1994 report from New York-headquartered market research concern Northern Business Information, could translate into a global market worth up to US$ 18.2 billion by 1998.
While voice predominates in corporate communications, the use of data is growing much faster. The first fully fledged datacommunication systems, such as those constructed to handle airline reservations, were only applicable to specific industry sectors. Data manipulation and transport has subsequently become vital to businesses and organizations in many different sectors.
Part of the reason for this is the widespread adoption of desktop computers as multi-purpose business tools, and the creation of local area networks (LANs) of these computers in an effort to increase office productivity, share IT resources and peripherals, and facilitate workgroup collaboration. According to a number of reports, including one published in 1995 by the Financial Times, global shipments of workstations and personal computers (PCs) will reach 100-150 million by the year 2000. Already, in many countries, locally-networked PCs dominate the market for new and replacement equipment.
Two LAN types are very well established. Those based on the Ethernet protocol operate at a basic rate of 10Mbps, (although Fast Ethernet shares 100Mbps between its users), and are said to account for over 50% of the installed LAN base worldwide. Token Ring LANs, running at up to 16Mbps, account for around 35% of the installed base. A third standard, the Fibre Distributed Data Interface (FDDI) is designed for campus applications and operates at a speed of 100Mbps. A copper-based version of FDDI has also been developed.
A common goal of business enterprises is to turn what are currently inward-looking telecommunication and IT systems into outward-looking, customer-oriented systems. The aim is to capture and control customers, and flatten the cost of doing so.
One of the approaches to this is to use new technology in the commercial world to rationalize customer-facing staff levels and remove intermediary organizations in the food chain. An example is provided by the retail trade which, according to the Germany futurologist Dr Rolf Homann, will shrink by 50% in the next ten years.
The consolidation of geographically separate PBXs and LANs into metropolitan and wide area networks (MANs and WANs) has a number of attractions in the context of lowering operational costs and increasing workgroup productivity. MAN and WAN construction is also a response to the increasing globalization of business and the need to shrink management hierarchies, and reflects the contemporary reliance on enterprise-wide access to information and data. All of this is reinforced by a move to client/server architectures in which branch office "clients" are connected to centrally- or regionally-based 'server' headquarters.
In general the internetworking of data-predominant LANs over MAN and WAN distances is more problematic than the internetworking of voice-predominant PBXs, although the latter has its share of difficulties when it comes to internationally networking an extended range of digital switchboard features.
One problem for LANs is that network protocols associated with distributed processing are less than ideal for wide area applications. Another is that LAN traffic is very bursty with the average utilization being typically between 10% and 30% of the peak bandwidth available. Users are then faced with a major MAN/WAN dimensioning problem: if MAN/WAN capacity is too low to accommodate peaks, traffic bottlenecks and poor response times follow; if configured for peak bandwidth at all times, operating the MAN/WAN becomes uneconomic.
Technology and service options available for deployment in corporate MANs and WANs include the following standards developed by the ITU:
These have significantly different characteristics and the particular solution chosen, or more likely combination of solutions chosen, will depend on the requirements of the particular enterprise in terms of MAN/WAN size, the location of access points, user numbers, traffic types and patterns, speed and throughput, and growth potential.
Despite a disappointing commercial debut in the mid-1980s, ISDN is now seen in some quarters as a useful means of low speed LAN interconnect. Tariffs are generally usage-based, and ISDN has often been used in a fall-back role with private circuits. The integration of ISDN with desktop video packages also extends its utility.
X.25, the ITU standard adopted in 1976, is a well-established packet technology. In operation, data is divided into packets and routed to its destination according to address information contained in the packets.
The technology has a number of attractions. Since packets from different users can be interleaved where appropriate over the same link, it is a cost-effective service for low volume, intermittent data.
And X.25 has comprehensive built-in error checking and management capabilities, and can be operated over poor quality lines. "You could run an X.25 network over barbed wire," jokes one vendor.
It also boasts a protocol conversion capability, allowing the interconnection of dissimilar types of terminals, and provides a degree of security through an access control mechanism.
On the downside, X.25 is relatively slow. While some equipment can run at 2Mbps, most of the extant networks peak at 64Kbps. As such, X.25 is not considered suitable for LAN interconnect where traffic bursts at high rates.
Frame relay is basically a very fast, stripped down version of X.25. Without the built-in error checking and management capabilities of its predecessor it can operate at speeds from 2 to 45Mbps. Actual frame relay data throughput can be ten times higher than X.25, even on a 64Kbps link.
Other positive attributes are efficient response times, and an ability to assist in the routing of data through complex networks. However, conventional frame relay is not suitable for voice and video applications, since error frames are "dumped", breaking up the signal. Also, with most current frame really technology, users have to specify their capacity requirements in advance.
This is not the case with SMDS, a connectionless service in which calls are set up and broken down in a similar fashion to calls in the telephone network. Users then only pay for the capacity they use. SMDS occupies a speed range somewhere between the 2Mbps of standard frame relay and the 155Mbps of ATM.
ATM initially operates at around 155Mbps - although some applications are presently run at 45Mbps - and is scaleable to 622Mbps and beyond. Although a fast packet technology like frame relay, it is able to accommodate voice, video and images as well as data due to its higher speed. Added to this it can support both frame relay and SMDS.
Nearly all ATM service development activity is currently targeted at enterprise networks, and observers do not anticipate ATM core public networks for some years. One reason is that the necessary large scale, ultra-fast switches are still at an early stage of development. "The development of effective traffic control mechanisms and their linkage to quality-of-service definitions is essential" adds UK consultancy firm Ovum.
As with ATM, both in principle and in practice, businesses can exploit the benefits of broadband frame relay and SMDS technologies in private networks. This is not a universal option, though. Broadband private networking is both capital-intensive and involves not inconsiderable technological and skills risks. Many corporations accordingly interconnect their LANs using public broadband networks and services.
This isn't the only option though. Virtual private networks (VPNs) and outsourcing are attracting considerable interest among corporate communications and Management Information Services (MIS) managers .
In VPNs, operators use special software to furnish capacity on the public network for the private use of each participating corporate customer. The potential advantages of this arrangement include shifting the technology and skills risks to the public telecommunications operator (PTO), easier access to international networking, and lower costs in circumstances where circuits sit idle for long periods of time. Voice traffic was an early candidate for the VPN treatment.
Australian carrier Telstra lays claim to having launched one of the world's first international VPNs. Udhay Mathialagan, the company's group manager for international products, says that VPNs are all about making the communications manager's life easier. He also observes that Telstra's international VPN can be one third cheaper than using dial up telephone lines.
But, naturally, VPNs don't meet all requirements. Real private networks, based on leased lines, can be cheaper when intensively used, and user control in this instance is higher. Network security may also be tighter.
The PTO management of private networks is a different proposition. In this, the operator offers end-to-end international service packages for corporate customers. Since different carriers have different areas of expertise, and no one PTO has sufficient resources to furnish all required service options on a global scale, those aspiring to success in this marketplace have been obliged to move forward through partnerships with other like-minded PTOs. These groupings are sometimes dubbed 'super-carrier' alliances.
Current super-carrier alliances include: the US$ 4.3 billion Concert partnership between BT and the second largest US long distance operator MCI; the Stentor consortium of Canadian carriers; the tandem alliance of the Unisource group of the Dutch, Swedish, Swiss and Spanish telcos with the WorldPartners consortium involving AT&T and a number of leading Asia-Pacific telecommunication carriers; and the proposed US$ 4.2 billion deal in which the Atlas team of France Telecom and Deutsche Telekom will make a 20% equity investment in the third largest US long haul carrier Sprint. This last alliance awaits approval in the US by the FCC (although the Justice Department has approved the deal) and in Brussels by the European Commission. Other super-carrier groupings contend that the French and German telecommunications national markets should be liberalized before the deal progresses; France Telecom, Deutsche Telekom and Sprint argue that they will only operate in fields already open to competition, prior to to European services being liberalized in 1998.
But one-stop shopping is not the whole of the super-carrier story. What many of the alliances are proposing is that they build, operate and manage total private networks on behalf of their customers. This arrangement is termed outsourcing, an activity which it is claimed can save between 5% and 20% of total telecommunications expenditure for the user.
Outsourcing alliances have their own set of challenges. It has been said that although the PTOs are arguably better positioned to cope with the technical, service and regulatory complexities involved in the construction of global networks, the current pace of technological change and re-regulation might mean that long term agreements negotiated between corporate users and outsourcing suppliers now may not be to the advantage of users in the future.
According to a 1994 report from UK telecommunications research concern CIT Research, of US$ 57.5 billion which will probably be devoted to telecommunications by the world's 1000 biggest corporations in the year 2003, expenditure on public telecommunications services could be worth around US$ 37.5 billion, with spending on self-provided or insourced services reaching another US$ 12 billion. This would leave outsource vendors with just US$ 8 billion.
An equally important concern is that operator alliances are faced with potentially serious conflicts of interests. Quite apart from bi-lateral agreements between operators belonging to different consortia, some consortia members also continue to act unilaterally in the outsourcing market-place, and some have interests in different outsourcing groupings.
In any event, the existence of such alliances raises some tricky questions concerning user choice. "What happens if you need services not handled by alliances? What if you need interconnection to countries not covered by alliances? What if you want a combination of carriers that are not pre-assembled in the alliance?" ponders Bill Coopman, Chairman of the Council of the International Telecommunications Users Group (Intug).
Value added network (VAN) services, which can be provided unilaterally by individual members of the super-carrier alliances, are less controversial. A partial list of services provided by the PTOs and other service providers which make up the international VAN community include X.25, frame relay, electronic data interchange (EDI), electronic mail and X.400 services, enhanced facsimile, computerized travel reservation services, on-line card authorization and videotex.
The characteristics of the VAN industry vary from country to country and region to region. For example, in its continuously updated study "VANS Markets Europe", Ovum finds that telcos constitute the largest group in the top 40 list of regional suppliers, with 14 companies accounting for 46% of revenue. Other big European VAN players are computing services companies and industry-specific service providers. The top five European VAN suppliers by revenue are France Telecom, Reuters, Spain's Telefónica, Deutsche Telekom and Unisource. Ovum thinks the regional market for these services could bepoised to achieve a 9% compound annual growth rate, which would hit a value of US$ 8.4 billion by the end of the decade.
Worldwide, commercial VAN service providers are setting up shop on the Internet in increasing numbers, with e-mail being a popular entry point for corporate users.
The Internet, which began life as a US military and academic/research network, is actually a network of networks. At the beginning of 1995 some 5 million host computers interlinked an estimated 30 million Internet users in over 90 countries. One projection is that by the year 2000 there could be as many as 125 million host computers in the Internet system. If that happens then will probably generate a greater volume of traffic than the global telephone network.
Commercialization of the Internet does, however, continue to be dogged by concerns about service quality, network management and security.
Because of its structure, the Internet is characterized by low throughput rates, and data can get lost. And by its very nature, a network of networks can sometimes lack a unified management structure. Equally worrying for commercial users is the apparent absence of security in the network, and the prospect that this will expose business computers to unauthorized intrusion, jeopardize the confidentiality of business data, and facilitate charge card mis-use and fraud. Over time, and given the enormous growth forecast for the system, the quality of the Internet will have be improved if business and commercial users are to be able to take true advantage of its possibility. It seems likely, given this, that more cohesive management mechanisms will be implemented, with increased security.
Commercial users are already beginning to install computer security systems and 'firewalls' which segregate LANs from the rest of the network. For the future, the use of encryption and the construction of secure commercial overlays on the Internet look likely. Communications and information security is not just a concern for Internet users and, in any case, a degree of protection already arises fortuitously from the sheer volumes of Internet users and traffic. As Alcatel NV director of marketing and business development Dominique de Boisseson points out, if messages are lost the likelihood of them ending up in the hands of anyone who can make illicit gain from receiving them is extremely low.
The Internet will undoubtedly become a major element in future enterprise networking: one estimate by US market research house Killen & Associates is that, in the year 2000, up to 8% of the estimated total US$ 8 trillion worth of goods and services purchased worldwide will be bought on-line. However, there's no suggestion that the Internet, or any other single network service or technology, will make all the other present-day options redundant. Corporate communications will continue to depend on a blend of services and technologies, and this blend will change as both businesses and network capabilities change.
Click Here to Return to Telecommunications Features and Press