Technical Insight
Satellite and last-mile links offer mm-wave prospects
High-speed wireless broadband data communications is expected to create a diversified, multibillion-dollar market for millimeter-wave devices in the next five years. According to Ed Weatherwax, the key areas are cellular infrastructure, two-way satellite terminals and terrestrial broadband access.
Despite the current challenges in the telecommunications market, bandwidth needs continue to grow, and this in turn is creating opportunities for manufacturers of high-frequency communications equipment. Cellular operators deploying 2.5 and 3G mobile networks are using high-frequency (>10 GHz) wireless backhaul solutions and installing three times more cell sites than existing networks, with backhaul data rates that are two to four times higher.
Cost-effectively bridging the gap between the fiber infrastructure and paying customers is also an issue for fixed network operators. Although high-frequency radios have, for several years, provided data rates of more than 10 Mbit/s without the need to bury cable, attempts to provide access via this technology (for example local multipoint distribution service, or LMDS) have not been successful to date. The key issues are the need to own spectrum licenses, stringent performance criteria and high equipment costs.
However, the revitalization of the market for two-way satellite services, coupled with the recent reallocation of several GHz of spectrum above 70 GHz, is creating demand for high-frequency devices. Prices are already decreasing due to technology improvements, such as high-yield monolithic chips and automated assembly, tuning and test. This could push the total annual volume for >10 GHz radios to more than 1 million units.
Terrestrial cellular infrastructureNew cellular phones and the resulting new service offerings will require a large expansion and reinvestment in existing networks. Estimates suggest that more than 1.7 million backhaul and infrastructure high-frequency radios will be required by 2006. The widespread deployment of wireless "hotspots", with the corresponding need for backhaul connectivity, could drive growth in this market segment. The market in 2003 was around 250,000 units, with a value of $300-500 million. This segment represents growth from the traditional low-volume military uses for high-frequency components and subsystems, but is not currently a high-volume market.
Two-way satellite terminalsThe two-way satellite terminal or VSAT (very small aperture terminal) market is a recent entrant into the commercial high-frequency product market. Existing two-way services offered at Ku-band (12-18 GHz) have not achieved the broad market acceptance that was expected, despite benefits such as global access, rural connectivity, and integration with broadcast services. The most likely users of VSAT services in the US are the 25-40% of customers who live in rural areas not covered by terrestrial broadband alternatives. If only 1% of rural households require broadband connectivity, this still represents around 3 million potential subscribers. However, there are fewer than 500,000 Ku-band system users, mainly because of service problems and high service charges.
Next-generation Ka-band (26-40 GHz) systems are expected to offer better service at a lower cost by leveraging smaller spot beams, dynamic bandwidth allocation, smaller antennas, and other technological innovations. The result for the customer is that Ka-band offers significantly more overall system capacity than Ku-band, enabling lower service pricing.
Although deployment of Ka-band satellites has been delayed, current plans call for Ka-band satellite service deployments in 2004/2005 with a potential for 6 million subscribers by 2007. This represents a large market opportunity for the high-frequency community, but also some difficult challenges.
Although a number of license awards and orbital allocations have been made over the last 9 years, many ventures set up to provide bandwidth on demand, Internet access and HDTV services via Ka-band either failed or were delayed. Contributing factors included the general telecom downturn, acceptance of low-cost DSL and cable terrestrial access solutions, and the lack of success for Ku-band solutions. More importantly, the satellite industry was unable to construct profitable enterprise or consumer service offerings due to the high cost of infrastructure/bandwidth and terminals.
Renewed interestDespite the slow start of Ku-band services, systems such as DirecWay have now connected more than 250,000 two-way subscribers, which has sparked renewed interest in Ka-band services. Plans are now more modest, typically using hybrid Ku/Ka-band satellites, are driven by strong industry leaders, and are mostly based on capturing the enterprise market first. Planned Ka and hybrid Ku/Ka-band systems are shown in the table, including the Telesat system (figure 1).
Even with a stronger emphasis by these new players on enterprise and military users, there remains a major push to reduce overall system costs, which are driven by the space segment. The ground terminal cost is another area of focus. Historically, RF assembly in the ground terminal has been a major cost driver, particularly at Ka-band. High levels of integration and automation have driven a significant amount of labor and material cost from the module; however, a key cost driver remains the MMIC chipset, and in particular the need to achieve the desired output power. Terminal requirements are in the 1-3 W range, depending on the desired antenna size, data rate, modulation format and other factors. As shown in figure 2, MMICs represent 40-50% of the cost of the assembly, and this figure is expected to grow to 65-70%, following improvements to the remaining parts of the assembly.
Although current and planned improvements will allow RF assembly vendors to approach the price targets of terminal suppliers and service providers, all parties must work together to drive MMIC chip costs down through circuit compaction, improved yields, higher levels of integration and other approaches. Ultimately, terminal costs will determine whether Ka-band systems will be used for enterprise networks only (hundreds of thousands of units) or for consumers as well (millions of units).
Terrestrial broadband accessFor millions of businesses and even residential users outside a fiber ring, the only feasible high-speed broadband network connection option remains a "last-mile" (or "first-mile") fixed wireless link. Although DSL, T1/T3 and cable technologies offer solutions to many users, a large worldwide population remains unserved. It appears that low-frequency alternatives such as Wi-Fi (802.11) and WiMAX (802.16a) will solve the consumer/SOHO needs, but the deployment of higher-frequency networks, for example LMDS for access applications, is still very limited.
However, there is an increasing demand for enterprise-level connectivity to support Ethernet deployments, which are happening at a rate of more than a million 1 Gb Ethernet switch ports per month. 10 Gb ports are now being deployed, and projections call for 100 Gb Ethernet performance by 2007. Connection of these ports to the fiber infrastructure is moving very slowly, with fiber trenching virtually halted and current-generation wireless technologies limited to about 622 Mbit/s.
Alternatives that provide 1 Gb-plus data rates are available, but these have limitations. Radios in the 60 GHz band are unlicensed, increasing the probability of interference, and are limited by low power and atmospheric effects to very short-range applications. Free-space optics links have not received widespread acceptance due to weather effects, and solutions that improve availability (combinations of free-space optics and 60 GHz links) result in higher cost.
A new alternative has emerged with the release of rules for the 71-76, 81-86 and 92-95 GHz (70/80/90 GHz) bands. The industry-driven rules provide broad bandwidth consistent with the development of high-capacity bridging products without the limitations of the current licensed spectrum. Because broad-bandwidth frequency blocks are available in the 70/80/90 GHz bands, and because atmospheric effects are conducive to the development of radios with appropriate transmission ranges, the communications industry saw an opportunity to create rules that would enable simpler, lower-cost radios.
There is a general consensus that the major market for devices in these bands will be access links to the 8000 fiber-connected points-of-presence (POPs) in the US. There are also approximately 750,000 business buildings with more than 20 employees, of which only 5% have fiber connections. However, approximately 75% of these buildings are located within 1 mile of a POP. The business opportunity is to bring each of these buildings online with fiber-like capacity by leveraging the benefits of the spectrum available at 70/80/90 GHz.
Transceiver costsThe transceiver cost will be a key issue in developing cost-effective equipment at 70/80/90 GHz. Cost goals for links in these bands are in the $2000-10,000 range ($1000-5000 per radio) in quantities of 10,000 or more, depending on the data rate required. Figure 3 provides top-level specifications and cost targets for the MMIC chipset for these radios. Some of the parameters are set to enable upgrades to higher radio data-rate performance, and can be relaxed for low-cost, relatively low data-rate applications. The performance of the resulting chain permits operation with 1-10 Gbit/s data rates at >1 mile range with 99.999% availability.
This emerging segment presents a challenge as well as an opportunity for MMIC and subsystems suppliers. There are only a limited number of MMIC chipsets available in these bands today, and they are not optimized for high-performance communication links. Estimates suggest that shipments of more than 1 million units are possible if performance and cost targets can be reached, although one must be mindful of the previous predictions of similar volumes for access products from the prior generation of systems at 20-40 GHz. However, growing demand and the adoption of user-friendly licensing and rules compatible with lower-cost systems, combined with the interest of major manufacturers, makes this an extremely viable market segment for high-frequency equipment manufacturers.
Cost-effectively bridging the gap between the fiber infrastructure and paying customers is also an issue for fixed network operators. Although high-frequency radios have, for several years, provided data rates of more than 10 Mbit/s without the need to bury cable, attempts to provide access via this technology (for example local multipoint distribution service, or LMDS) have not been successful to date. The key issues are the need to own spectrum licenses, stringent performance criteria and high equipment costs.
However, the revitalization of the market for two-way satellite services, coupled with the recent reallocation of several GHz of spectrum above 70 GHz, is creating demand for high-frequency devices. Prices are already decreasing due to technology improvements, such as high-yield monolithic chips and automated assembly, tuning and test. This could push the total annual volume for >10 GHz radios to more than 1 million units.
Terrestrial cellular infrastructureNew cellular phones and the resulting new service offerings will require a large expansion and reinvestment in existing networks. Estimates suggest that more than 1.7 million backhaul and infrastructure high-frequency radios will be required by 2006. The widespread deployment of wireless "hotspots", with the corresponding need for backhaul connectivity, could drive growth in this market segment. The market in 2003 was around 250,000 units, with a value of $300-500 million. This segment represents growth from the traditional low-volume military uses for high-frequency components and subsystems, but is not currently a high-volume market.
Two-way satellite terminalsThe two-way satellite terminal or VSAT (very small aperture terminal) market is a recent entrant into the commercial high-frequency product market. Existing two-way services offered at Ku-band (12-18 GHz) have not achieved the broad market acceptance that was expected, despite benefits such as global access, rural connectivity, and integration with broadcast services. The most likely users of VSAT services in the US are the 25-40% of customers who live in rural areas not covered by terrestrial broadband alternatives. If only 1% of rural households require broadband connectivity, this still represents around 3 million potential subscribers. However, there are fewer than 500,000 Ku-band system users, mainly because of service problems and high service charges.
Next-generation Ka-band (26-40 GHz) systems are expected to offer better service at a lower cost by leveraging smaller spot beams, dynamic bandwidth allocation, smaller antennas, and other technological innovations. The result for the customer is that Ka-band offers significantly more overall system capacity than Ku-band, enabling lower service pricing.
Although deployment of Ka-band satellites has been delayed, current plans call for Ka-band satellite service deployments in 2004/2005 with a potential for 6 million subscribers by 2007. This represents a large market opportunity for the high-frequency community, but also some difficult challenges.
Although a number of license awards and orbital allocations have been made over the last 9 years, many ventures set up to provide bandwidth on demand, Internet access and HDTV services via Ka-band either failed or were delayed. Contributing factors included the general telecom downturn, acceptance of low-cost DSL and cable terrestrial access solutions, and the lack of success for Ku-band solutions. More importantly, the satellite industry was unable to construct profitable enterprise or consumer service offerings due to the high cost of infrastructure/bandwidth and terminals.
Renewed interestDespite the slow start of Ku-band services, systems such as DirecWay have now connected more than 250,000 two-way subscribers, which has sparked renewed interest in Ka-band services. Plans are now more modest, typically using hybrid Ku/Ka-band satellites, are driven by strong industry leaders, and are mostly based on capturing the enterprise market first. Planned Ka and hybrid Ku/Ka-band systems are shown in the table, including the Telesat system (figure 1).
Even with a stronger emphasis by these new players on enterprise and military users, there remains a major push to reduce overall system costs, which are driven by the space segment. The ground terminal cost is another area of focus. Historically, RF assembly in the ground terminal has been a major cost driver, particularly at Ka-band. High levels of integration and automation have driven a significant amount of labor and material cost from the module; however, a key cost driver remains the MMIC chipset, and in particular the need to achieve the desired output power. Terminal requirements are in the 1-3 W range, depending on the desired antenna size, data rate, modulation format and other factors. As shown in figure 2, MMICs represent 40-50% of the cost of the assembly, and this figure is expected to grow to 65-70%, following improvements to the remaining parts of the assembly.
Although current and planned improvements will allow RF assembly vendors to approach the price targets of terminal suppliers and service providers, all parties must work together to drive MMIC chip costs down through circuit compaction, improved yields, higher levels of integration and other approaches. Ultimately, terminal costs will determine whether Ka-band systems will be used for enterprise networks only (hundreds of thousands of units) or for consumers as well (millions of units).
Terrestrial broadband accessFor millions of businesses and even residential users outside a fiber ring, the only feasible high-speed broadband network connection option remains a "last-mile" (or "first-mile") fixed wireless link. Although DSL, T1/T3 and cable technologies offer solutions to many users, a large worldwide population remains unserved. It appears that low-frequency alternatives such as Wi-Fi (802.11) and WiMAX (802.16a) will solve the consumer/SOHO needs, but the deployment of higher-frequency networks, for example LMDS for access applications, is still very limited.
However, there is an increasing demand for enterprise-level connectivity to support Ethernet deployments, which are happening at a rate of more than a million 1 Gb Ethernet switch ports per month. 10 Gb ports are now being deployed, and projections call for 100 Gb Ethernet performance by 2007. Connection of these ports to the fiber infrastructure is moving very slowly, with fiber trenching virtually halted and current-generation wireless technologies limited to about 622 Mbit/s.
Alternatives that provide 1 Gb-plus data rates are available, but these have limitations. Radios in the 60 GHz band are unlicensed, increasing the probability of interference, and are limited by low power and atmospheric effects to very short-range applications. Free-space optics links have not received widespread acceptance due to weather effects, and solutions that improve availability (combinations of free-space optics and 60 GHz links) result in higher cost.
A new alternative has emerged with the release of rules for the 71-76, 81-86 and 92-95 GHz (70/80/90 GHz) bands. The industry-driven rules provide broad bandwidth consistent with the development of high-capacity bridging products without the limitations of the current licensed spectrum. Because broad-bandwidth frequency blocks are available in the 70/80/90 GHz bands, and because atmospheric effects are conducive to the development of radios with appropriate transmission ranges, the communications industry saw an opportunity to create rules that would enable simpler, lower-cost radios.
There is a general consensus that the major market for devices in these bands will be access links to the 8000 fiber-connected points-of-presence (POPs) in the US. There are also approximately 750,000 business buildings with more than 20 employees, of which only 5% have fiber connections. However, approximately 75% of these buildings are located within 1 mile of a POP. The business opportunity is to bring each of these buildings online with fiber-like capacity by leveraging the benefits of the spectrum available at 70/80/90 GHz.
Transceiver costsThe transceiver cost will be a key issue in developing cost-effective equipment at 70/80/90 GHz. Cost goals for links in these bands are in the $2000-10,000 range ($1000-5000 per radio) in quantities of 10,000 or more, depending on the data rate required. Figure 3 provides top-level specifications and cost targets for the MMIC chipset for these radios. Some of the parameters are set to enable upgrades to higher radio data-rate performance, and can be relaxed for low-cost, relatively low data-rate applications. The performance of the resulting chain permits operation with 1-10 Gbit/s data rates at >1 mile range with 99.999% availability.
This emerging segment presents a challenge as well as an opportunity for MMIC and subsystems suppliers. There are only a limited number of MMIC chipsets available in these bands today, and they are not optimized for high-performance communication links. Estimates suggest that shipments of more than 1 million units are possible if performance and cost targets can be reached, although one must be mindful of the previous predictions of similar volumes for access products from the prior generation of systems at 20-40 GHz. However, growing demand and the adoption of user-friendly licensing and rules compatible with lower-cost systems, combined with the interest of major manufacturers, makes this an extremely viable market segment for high-frequency equipment manufacturers.
