Home > Consultations > Consultation Documents > Telecoms Review > Telecoms Review  > Annex J

J.1 Many of the possible changes to the telecoms sector discussed in Section 5 of the Phase 1 consultation document are the result of technological change. This annex describes what Ofcom believes are likely to be the most significant technological changes in the period up to the end of the decade.

J.2 Technology trends can be considered at a number of different levels. There is an overarching technological trend towards increasing performance and reducing costs. This overarching trend implies a number of subordinate trends, each of which has particular implications for telecoms markets. Individual technologies and developments, for example the evolution from 2G to 3G mobile, represent a further level of detail.

J.3 This annex is focused on the overall direction of technological change and its key implications for markets. It is not focused at the level of individual technologies. Though individual technologies may result in fundamental, lasting changes, those new technologies are nonetheless likely to be manifestations of the more general trends discussed below.

Ovearching direction of technological change: increasing performance and reducing costs

J.4 In almost all areas of IT and telecoms technology, there is a significant and persistent trend towards increasing performance and capability. Moore's Law ^(-128-) covering the increasing capability of silicon is well known, and increasing computer processing power and memory are manifestations of this. However, performance increases can also be seen to a greater or lesser extent in fibre-optic bandwidth, spectrum exploitation, compression technologies, display technology, power efficiency and battery performance. Many of these performance increases are due to increasing miniaturisation (more on a chip) and greater integration (fewer chips to achieve a given function). Alongside the trend towards increased performance, unit costs of these technologies have been dropping.

J.5 This overarching trend is important because increasing performance, in particular that driven by miniaturisation, makes possible types of terminal and service that could not previously have been achieved. It is therefore a major driver to innovation.

J.6 This trend is also important because of its impact on costs. Greater miniaturisation and integration is the main reason why apparatus and customer devices are becoming cheaper. Much network infrastructure also shows very strong scale economies, so up to a point increasing traffic volumes result in decreasing unit costs, which in turn can stimulate greater usage.

J.7 This overarching trend has a number of subordinate trends that have particular implications for telecoms markets. These are discussed in the remainder of this section, and are:

• analogue to digital;
• circuit-switched to packet-switched networks;
• fixed to mobile or mobility;
• tethered to tetherless;
• dial-up to 'always on';
• narrowband to broadband;
• telecom-specific to IT-generic;
• operator-centric standards to vendor-centric standards;
• asymmetric to symmetric; and
• centralised to distributed.

Analogue to digital

J.8 Networks and services are progressively becoming digital rather than analogue; i.e. signals are transmitted in the form of encoded bits of information which are then decoded by equipment connected to the network. For example, the core of the telephone network has been fully digital since 1998. The two analogue mobile networks have now closed. However, fixed telephony is still, on the whole, analogue at the point that it is delivered to the customer. In future, fixed telephony is likely increasingly to be digital at the point that it is delivered to the customer. The growth of broadband may prove to be the vehicle for this.

J.9 This trend is important because digital delivery allows a richer set of services to be offered to consumers. For example, SMS, MMS and e-mail are all available from a GSM phone as well as voice. As the fixed telephone network becomes digital all the way to the customer, a similarly rich set of services will be available over the fixed network. Digital presentation also allows a more integrated way of delivering services and often uses the available bandwidth more efficiently.

Circuit-switched to packet-switched networks

J.10 Telephony networks have for many years been circuit-switched. In a circuit-switched transmission, an end-to-end pathway (or 'circuit') is created through the network for the duration of a call. The growth of IP-based data services now makes packet transport increasingly the dominant transport mode. In a packet-switched transmission, the data to be sent is divided into many 'packets', each of which is routed individually over the network, and then reassembled by the reception equipment. This is shown in figure 50. Whereas now the internet is often accessed over a voice network, in future voice calls are likely to be delivered over packet networks.

Figure 50: Circuit-switched and packet-switched networks

J.11 This trend is important because packet networks are more efficient and less costly than circuit-switched networks. They may also allow new forms of supplier in the market; for example, voice providers who deliver services over a third party's IP network. Furthermore, packet-based networks offer an opportunity to carry a greater range of services over a single network. They can deliver both voice and data, and are used for both fixed and mobile services. Integrating Voice over IP and video services into the traditional PSTN does, however, raise some technical issues of end-to-end quality.

Fixed to mobile or mobility

J.12 The number of mobile connections has now exceeded the number of fixed telephone lines. This huge penetration of mobile phones has changed the entire telecoms landscape by providing an alternative commercially viable, near-ubiquitous infrastructure which provides telephony services. Additionally, consumers are demanding 'mobility' services to support their increasingly nomadic work and lifestyles. They require services that can be used 'on the pause' or by access from any available fixed line ('mobility').

J.13 This trend is fundamental to the shape of the whole telecoms market and the type of services that consumers are demanding. The residential mobile market has all the hallmarks of a truly consumerdriven market and has broken away from 'utility attitudes' towards the fixed telephone business.

Tethered to tetherless

J.14 Within homes and businesses, fixed services are increasingly being used over 'tetherless' (short range radio-based) systems, such as DECT, Bluetooth and WiFi. Tetherless television is a likely future development. Consumers value the freedom, flexibility and uncluttered nature of tetherless delivery.

J.15 This is important because it is likely to have a major impact on types of services which will be demanded and the devices used to access them. It may also have an impact on spectrum policy, for example because many people see Ultra Wide Band as the natural technology for tetherless television.

Dial-up to 'always on'

J.16 The market is moving from narrowband dial-up access to broadband 'always on' services. Much of the growth in today's broadband connections is for moderate amounts of bandwidth, such as 128kbit/s or 150kbit/s. A major attraction of broadband to these consumers could be the 'always on' characteristic as well as the greater bandwidth.

J.17 This is important because 'always on' stimulates the growth and development of more real-time services, such as instant messaging, where permanent connectivity enhances value. Studies also show that ' always on' customers spend more time online, and their usage is likely to continue to increase. As a result, the current 'unlimited use' broadband packages aimed at migrating customers from dial-up may not necessarily be sustainable in the long term. ISPs in future may need to find ways to contain or charge for the ever increasing volumes of data that result from 'always on' connectivity.

Narrowband to broadband

J.18 The present trend is for migration from narrowband data services to a range of broadband services. Narrowband services vary in speed according to the nature of the modem and connection used, but are (by definition) under 128kbit/s. Most current broadband deployment for residential and SME customers is in the range 128kbit/s to 2mbit/s. But in future broadband connection speed requirements might increase. There could be a virtuous circle of take-up as more broadband customers consume more broadband services and their expectations of throughput and responsiveness increase. At the same time, service providers may become more confident in providing more bandwidth-intensive content andapplications. In this way, the emergence of new types of visual services, such as time-shifted television and video on demand, could drive the demand for bandwidths closer to those currently provided by digital TV.

J.19 Were this demand for increased connection speed to come about, it would be important because greater bandwidth demands will affect the investment needed in networks and their relative economics. This could at some stage make connection to the home commercially viable via fibre or high bandwidth radio. If this occurs then the physical constraint that the access network puts on maximum bandwidth will disappear, and pressures will focus elsewhere in the network, for example on the scalability of content servers and the architecture of servers and caches. This may also impact on the viability of ISPs' flat rate tariffs and cause a shift towards volume or valuerelated tariffs.

Telecom-specific to IT-generic

J.20 The unit cost of telecoms devices is critically dependent on volumes. Equipment used in mass market private telecoms and IT systems has a cost advantage over the low volume, more specialised equipment traditionally procured by telecoms operators. This has led to the adaptation of technologies designed originally for LAN applications to use in wide-area networks. Prime examples are Ethernet services and WiFi. Similar effects could emerge in the production of software.

J.21 This convergence of IT and telecoms technologies is important because it will have a downward effect on network costs, and provide opportunities for operators capable of exploiting these technologies. It could lead to the decline of many traditional telecoms systems and architectures, such as ATM over SDH. Ethernet presentation to customers allows more 'plug'n'play', while WiFi will allow roaming between private and public 'hotspot' services. WiFi is also currently playing a role in widening the rural availability of broadband. The use of generic IT software in delivering telecom services could lower costs and accelerate time to market.

Operator-centric standards to vendor-centric standards

J.22 Traditionally, telecoms standards-making was dominated by the incumbent telephone operators working in formal standards bodies, such as the International Telecoms Union (ITU) and the European Technical Standards Institute (ETSI). In the present market, there are far more operators and service providers, and fewer global equipment manufacturers. These vendors have a clear interest in reducing product variants to create global scale, as the high cost of development and short product lives militate against being able to survive on a narrow national market. Therefore vendors have a greater interest, and therefore participate more, in harmonization and standards-making. In addition, standards are now far more likely to emerge from commercially driven 'consensus groups', including such bodies as the Internet Engineering Task Force (IETF), which may often be fast-tracked into formal standards bodies at a later stage.

J.23 This is important because it can affect what choices operators have in creating their own bespoke products and services. It tends to focus innovation on less standardised value-added services where the operator or service provider can develop or commission their own software development. Such services, especially at the early stages of market development, are unlikely to support interoperability with competitors, though this may emerge at a later stage when the benefits of first market entry have been exhausted.

Asymmetric to symmetric

J.24 Traditionally, telecoms networks have supported peerto- peer communications, such as telephony, telex and email. Broadband was thought to be different and would be dominated by client-server architectures. As such networks were expected mainly to be used for delivery of centrally sourced services and content. These architectures tend to be asymmetric, with more content being delivered from the centre to the edge. In practice, however, broadband services are currently being used heavily for peer-to-peer file sharing and gaming. This implies a more symmetrical demand for bandwidth than is provided for with current ADSL and cable modem technologies.

J.25 If this trend continues, it could have a fundamental effect on the nature of network architecture and future investments. Peer-to-peer services will put the intelligence at the edge of the network using an increasing range of devices, such as set-top boxes, games boxes, PDAs as well as the traditional PC. The growth of peer-to-peer services could accelerate the penetration of broadband, as network externalities are greater than with client-server models.

Centralised to distributed

J.26 Because of the growth of symmetric applications, intelligence is growing at the edge of networks, as greater functionality can reside in increasingly capable peer-edge devices. Such edge devices may be either customer terminals or new forms of service provider node.

J.27 This is important because the location of intelligence in the network critically affects the way in which innovation occurs, how it is developed and who influences it. If intelligence grows at the edge, the demands on the network can be simpler. Instead of transport, control and management all being centralised and embedded in the network, these will separate into different 'layers'. The control or customer management functions could be provided either by the network operator, or by independent players on top of the network provider's facilities. This is likely to change present views on what is the natural boundary between wholesale and retail markets. It may also give greater power to consumers in terms of choice of services.

---

Footnotes:

128:- In 1965 Gordon Moore, co-founder of Intel, observed that the number of transistors per square inch on integrated circuits had doubled every year since the integrated circuit was invented. Moore predicted that this trend would continue for the foreseeable future. In subsequent years, the pace slowed down a bit, but data density has doubled approximately every 18 months, and this is the current definition of Moore's Law, which Moore himself has blessed. Similar trends apply to processing speeds, storage densities and transmission bandwidths.

Back to top