Internet growth in Latin America:
(if you want to know where PDAs and Laptops are going wirelessly, consider the comments in these articles regarding the application of TDD for the pico cell. with Dell's presence in Latin America expanding, you'd think they should be able to capture a respectable percent of this growth)
telecommagazine.com
telecommagazine.com
The number of Internet accounts in the region was approximately 5 million at the end of 1998 and, according to Dataquest, will grow to 32 million by 2003. “We expect Internet growth in Latin America to be the highest in the world,” said David Schwartz, senior analyst for Dataquest.
Latin America Eyes Up Wireless Broadband Access
Industry & Market Update
In the competitive telecommunications access market in Latin America, wireless broadband access is a compelling solution. But choosing the appropriate technology platform is no easy task.
Kimberly Tassin
--------------------------------------------------------------------------------
The telecommunications access market in Latin America is a hotbed of activity. Over the last few years, privatisation has been occurring at a blistering pace and vast sums of capital have been invested as intense competition, globalisation and consolidation have reshaped the industry. Today, new international players are scrambling to build out infrastructure capable of providing more bandwidth to a population hungry for advanced services such as Internet access and corporate data networking. The number of Internet accounts in the region was approximately 5 million at the end of 1998 and, according to Dataquest, will grow to 32 million by 2003. “We expect Internet growth in Latin America to be the highest in the world,” said David Schwartz, senior analyst for Dataquest.
Competitive service providers are finding that wireless broadband access (WBA) technology is a viable solution for providing the needed bandwidth. Systems are affordable, available and service can be launched quickly. Latin America is ripe for the rollout of point-to-multipoint (PTM) networks. Even though the region lacks the backhaul infrastructure found in more developed countries, basic services are maturing and data is expected to account for 70 per cent to 90 per cent of network traffic within five years. Also, the population is largely urban, indicating the suitability of PTM networks, which are expected to serve small and medium-sized businesses or multiple dwelling units in dense metropolitan areas.
Until recently, WBA systems operating in Latin America were exclusively point-to-point (PTP). PTP establishes a wireless link between one hub and one remote site (customer). Point-to-multipoint systems allow service providers to connect a hub to multiple remote sites thereby creating greater network efficiency through sharing the hub resources and the costs among the several remotes. Point to-multipoint allows the service provider to not only build out his network incrementally, but also allows him to add remotes when demand requires. The shared-hub design offers tremendous potential for maximising revenue through efficient use of spectrum. Looking ahead, many believe that PTP will be used in areas where traffic is aggregated in backbone networks, whereas PTM will be used to serve the many buildings tributary to the backbone networks.
Latin America has progressed fairly quickly in licensing wireless broadband frequencies. Argentina, Brazil, Columbia and Mexico, the four largest economies, as well as Bolivia, Chile, Ecuador, Paraguay, Peru and Venezuela, have all licensed PTP or PTM frequencies. After several rounds of licensing in many countries, frequencies from 10 GHz to 38 GHz have been licensed, all with varying band allocations. During the licensing process, much debate occurred about which band plans to adopt. Some argued for adopting the US band plan, others argued for the Canadian band plan, and yet others pushed for a modified US band plan. The participants were unsuccessful in identifying a plan that would be accepted by all, so there are several plans.
Several key players have deployed or announced the intention to deploy point-to-multipoint service in Latin America. Techtel announced the expansion of its network in Argentina; Genesis, a subsidiary of Bell Canada International, has announced service in Venezuela; and others such as Diginet, Winstar, Formus and Convergence are all building out point-to-multipoint networks and are in various stages of deployment.
TDD in PTM Systems
The non-uniformity of band allocations in Latin America as well as other areas of the world, has been one of the key factors driving the development of time division duplexing (TDD) technology in the design of PTM networks. TDD is an airlink protocol that allows both send and receive transmissions to occur on a single channel, which is a capability that allows a PTM system to operate efficiently in any size spectrum band. The traditional alternative, FDD (frequency division duplexing), requires two channels (one for transmit and one for receive) and a guard band between the two channels. FDD has proved problematic and expensive for some of the smaller slices of spectrum as well as for some band allocations where duplex pairing is not designated.
In Argentina, for example, there are two allocations, the C-Block (26.85 GHz to 27.35 GHz and 31.075 GHz to 31.15 GHz) and the E-Block (25.85 GHz to 26.35 GHz and 29.175 GHz to 29.25 GHz). In both of these allocations, there is not enough room in the upper allocations (only 75 MHz) for the guard band, so a guard band would have to be carved out of the 500 MHz lower band. Since FDD requires at least 1 per cent spacing, doing so would result in 270 MHz of wasted spectrum. A TDD system operates full duplex communications in a single channel, ranging in size from 8.33 MHz to 50 Mhz. The operator can make full use of and receive full value from all licensed spectrum -- every channel becomes a fully functioning revenue stream.
In FDD systems, two RF channels are in continuous operation at each site -- one for transmit from hub to remote, and another for the return from remote to hub. At each site, the transmitter operates simultaneously with the receiver. As the channels are typically close to each other, the transmit signal can overwhelm the sensitive receiver tuned to the nearby channel. FDD systems require channel separation and special filters, called duplexers, to prevent such overload.
With TDD systems, duplexers and isolation techniques are unnecessary since the transmit and receive channels within a sector are never active simultaneously and cannot interfere with each other. The required separation (guard band) between channels is also unnecessary, since only one channel is used for both send and receive.
With FDD, a unique transceiver and duplexer design is required for each band plan. The simplified TDD architecture, however, wherein the duplexer is replaced by a switch, operates without modification across an entire band. TDD transceivers operating at different frequencies can be deployed side-by-side on a single hub rooftop platform, allowing RF engineering design to be simplified for the service provider who may want to operate multiple band allocations.
Eliminating the need for paired channels increases the channel assignment flexibility in PTM cellular plans, resulting in a much higher probability that the deployment will be free from self-interference problems. FDD systems must be re-engineered for each new RF channel plan resulting in the need for customised equipment for each deployment, which varies significantly from country to country. With TDD equipment, companies operating in multiple countries will be able to benefit from uniform, volume purchase agreements along with reduced inventory management, complexity and customisation lead times. This simplicity also reduces the number of required spares since with TDD only one radio per band is used.
TDD systems have quantifiable performance advantages over FDD systems when transporting bursty data like Internet traffic. Voice traffic is constant and predictable, whereas data traffic is variable and unpredictable. Where FDD systems fix their upstream and downstream data rates, TDD systems can dynamically vary the amount of channel capacity allocated to upstream and downstream communications -- up to the total capacity of the channel. In other words, it is possible to ‘borrow’ time from the transmit side to increase the return side as necessary, or vice versa. Thus, time-varying bi-directional data traffic can be transported more efficiently using TDD. In multipoint systems, this results in a 100 per cent improvement in burst rate capability at the hub over FDD. The ability to dynamically allocate bandwidth translates into a more positive business case through more efficient network oversubscription.
TDD-Based PTM Equipment
A TDD-based PTM system includes hub and remote indoor and outdoor units and provides 4 X E1 service from a central hub to remote locations via short line-of-site link ranges. The hub outdoor unit consists of the horn antenna and transceiver (called a transceiver assembly or TRA), distribution electronics and a mounting frame. The hub provides dual DS3 or dual OC3 interfaces. The hub indoor unit consists of one shelf of common control circuit cards and up to eight modem control assemblies (MCAs) all mounted in a 19-inch free standing rack. One MCA and one TRA combine to form one sector. The hub indoor unit and outdoor units are connected by coaxial cables.
The remote outdoor unit consists of a single TRA and its associated 10.5-inch parabolic antenna. The remote indoor unit encloses an MCA, AC power supply and three I/O modules. The remote indoor unit can be configured over the airlink and a serial port allows the remote to dial into the network operations centre, if needed.
Time division duplexing (TDD), time division multiple access (TDMA) and variable index modulation (4, 16 or 64 QAM) combine to maximise spectral efficiency and throughput and create a simplified architecture. One radio operates across an entire frequency band, enabling simplified deployment, redeployment and sparing. A maximum of four symmetric, duplex connections (i.e., four remotes) are supported with each connection, independently operating in 4-QAM, 16-QAM or 64-QAM mode. The variable index QAM permits an efficient trade-off between spectral efficiency and link range.
The hub and remote units use a scaleable and flexible software architecture. This software is designed to provide open standard, simple network management protocol (SNMP)-based network management capability. It enables operators to configure, manage, and upgrade both hub and remote unit software from a central location without disruption to the remote sites.
PTM technology allows networks to grow in a building-block fashion. Service can begin with one hub and one or two remotes, and then, as demand warrants, additional remotes can be added which will share the sector capacity. Also, varying antenna widths can be used as it is decided whether to acquire additional remotes with a wider beam or accommodate fewer more distant remotes with a narrower beam.
Kimberly Tassin is director of marketing for Wavtrace Inc, based in Bellevue, Washington. She can be contacted on ktassin@wavtrace.com. |
