COMPARATIVE APPROACHES IN THE ECONOMICS OF BROADBAND SATELLITE SERVICES
by Mark Dankberg, President & CEO, ViaSat, Inc and John Puetz, President, MasterWorks Communications
(courtesy of ViaSat)

There are a number of economic, business, and technical considerations in bringing broadband services to the marketplace using satellite based facilities. The basic concept that "one network can satisfy all broadband markets and applications" is challenged in this paper and shown to be too simplistic. In light of these three considerations, the basic satellite systems concepts (GEO and non-GEO) are evaluated, along with the requirements of the various market segments (consumer, direct-to-home, enterprise, VSAT, SOHO, and mobile), and the various system approaches being deployed or under construction. The intent of this paper is to inform the reader concerning market needs, economic drivers, system performance and service costs trade-offs and considerations. In addition some new concepts are presented that address what broadband service users and operators are really looking for.

return to top

Introduction

Historically satellites have been most successful in distributing information over very large geographical areas using a single transmission. With services such as television broadcasting, data broadcasting, digital messaging, enterprise virtual private networks (VPNs) and point-to-point telecom-datacom services, traditional "bent-pipe" satellites have played a significant role in our daily lives. A new generation of application needs, higher throughput requirements, and communication demands are changing the way satellite systems are designed, implemented and operated. New architectures and system networking concepts are being implemented to make satellite systems capable of addressing these new market demands. The progressive idea of making satellite systems that are optimized for highly in demand services (e.g., Internet access, VPNs, personal access) opens entire new market opportunities that go far beyond the traditional viewpoint of selling services only into markets that where satellite services excel (e.g., broadcasting, multicasting and content delivery).

While technology is an important and very necessary ingredient to success, equally important, if not more so, is the need for a viable business model that can withstand the rigors of the marketplace and provide earnings within a reasonable time frame. To that end, all aspects of a new broadband service must be carefully considered; market demands researched, user needs profiled, implementation and operational costs analyzed, service uptake rates accurately estimated and service revenues and margins realistically forecasted.

From a user's perspective, consumers have different service requirements than do corporations and small to medium enterprises (SME). Service speed, throughput capacity and connectivity are very different-and much more demanding in an enterprise environment. The mobile broadband market adds yet another dimension to system capability and design.

To illustrate the wide variety of market needs and user expectations we've formulated a service and market requirements matrix as presented in Table 1. As can be seen, there is a very wide range of service capability, performance expectations and pricing. Thus the concept that one system can address the diverse needs of the consumer, business and mobile marketplace is just not realistic.

Table 1 - Service / Mareket Requirements Matrix

To be successful in any particular market segment (or even any two adjacent segments) the system must be customized to meet the particular segment demands. For example, to provide upstream data rates of 1 to 4 Mbps the satellite terminal needs to have a much larger antenna, significantly more transmit power or the satellite must have a high G/T factor (e.g., spotbeam operation) than that for a consumer Internet access service capable of 64 to 256 Kbps. Another example is that a $300 to $400 terminal price is incompatible with a 4 Mbps upstream transmit speed. Direct peer-to-peer connections needed for enterprise networking applications (and potentially future consumer applications) require mesh connectivity, not hub-spoke as in other systems. Small dish, high-speed mesh connectivity is currently only achievable using specialized on-board satellite processing techniques in conjunction with Ka-band spot beams.

From the service operator's (and investors') perspective, the business' return on investment (ROI) must be attractive and compelling-service revenues need to be maximized and the operational costs minimized. Take for example the first generation of satellite broadband service (e.g., StarBand) that operates with upstream rates of 30 to 60 Kbps and much larger downstream capacity (150 to 500 Kbps per user, 30 Mbps total capacity) to the user terminals that number 10,000 to 20,000 per transponder. However, system operational limits keep the number of concurrent online users to below 8,000 per transponder. Yet the ROI economics for the service provider require many more subscribers per transponder-and the end users demand much higher data capacity as interactive broadband applications and services become more widespread.

return to top

What Does Broadband Satellite Really Mean?

Broadband satellite systems both receive and transmit rich-media content to and among network end-users whether at home or in the office-these systems are not intended to supply huge amounts of bandwidth for backbone infrastructure purposes. The market need is great for two-way broadband network access across large geographical areas where infrastructure has not been built out, or would be too costly to implement. In short, satellite will become the broadband "local-loop" in such communities.

Forecasted broadband satellite service revenues are projected in Figure 1 over the next eight years, growing from $2.2B this year to over $40B and contribute 30 percent of broadband service revenues worldwide. Thus, there is considerable economic motivation for today's heavy investment in next-generation broadband satellite systems by a number of players.

return to top

Satellite System Approaches

There are four basic technology categories that form the basis for the various satellite broadband service offerings: Ku-band FSS, bent pipe Ka-band, on-board processing Ka-band and L-band MSS. These approaches and representative service offerings are summarized in Table 2. The first generation services that are now in place use existing Ku-band fixed satellite service (FSS) satellites for two-way connections. Using FSS, a large geographical area (e.g., the United States or all of North America) is covered by a single broadcast beam.

Figure 1 - Global Broadband Satellite Service Revenue Growth
Source: IEC Study (Dr. J.N.Pelton)

Table 2 - 2-way Broadband Satellite Technologies

The new Ka-band systems use more focused beams that cover a much smaller area (hundreds of miles across, rather than thousands of miles with FSS) that form coverage cells like the illustration below. Adjacent cells use different frequency ranges but a given frequency range can be reused many times over a wide geographical area. In this way there is a large increase in overall capacity because of frequency reuse; the spot beam frequency gain is analogous to the difference between a direct-to-home broadcast signal and cellular phone coverage. From a practical standpoint, Ka spot beams provide 30 to 60 times the system capacity of the FSS approach. The increase system capacity to 30 Gbps plays a very significant role in helping to make satellite broadband services a long-term, economically viable business opportunity, as end-users' bandwidth requirements will only increase over the next five to ten years.

The Ka-band systems under development are being designed with two basic constructs: bent pipe and on-board processing. Bent-pipe satellites are essentially repeaters in the sky-they simply receive and retransmit signals without performing any additional functions like multiplexing, switching or routing. All waveform processing intelligence, like rain fade mitigation or data rate adjustment, is performed by the ground station terminal equipment. This bent-pipe approach is much less complex, less costly, and is less susceptible to obsolescence than the on-board processing approach.

Having said this, onboard processing has a number of benefits over bent-pipe technology and it will be deployed on three of the four forthcoming Ka-band systems as indicated in Table 3.

Table 3 - Global Broadband Service Offerings

On-board processing payloads act as intelligent signal routers and switches, directing traffic from one spot beam to another within the same satellite or to another sibling satellite to provide large regional or global single-hop connectivity. On-board processing enables very efficient full-mesh broadband connections that can adapt quickly to changing data throughput and system loading demands-all key attributes for enterprise networking and the increasing importance of supporting peer-to-peer networking applications.

The new Ka-band systems under development will be deployed in two varieties-regional and global. Four major global offerings are planned as shown in Table 3, with most scheduled to launch services in the 2003-2004 time period. Common to each of these systems are high-bandwidth transmit/receive capability and hefty system implementation price tags ($4B to $9B).

The regional broadband offerings, summarized in Table 4, will provide the first indications of business plan success for the satellite broadband markets as several have launched this year (e.g., StarBand and Astra-Net) and the remaining will be in service in 2002. These systems are much less complicated than their global counterparts and have greatly reduced system price tags ($500 to $900M). The regional systems appear much more likely to succeed because of less technical complexity, but more importantly they appear to have far fewer business risks-significantly lower infrastructure costs, less regulatory concerns, and fewer distribution and service channel issues.

Table 4 - Regional Broadband Systems

return to top

The Economics

Broadband means bandwidth, and in any media more bandwidth means higher transmission costs. Until now satellites shining glory has been delivering content (TV, movies or real-time data) to large numbers of content consumers using a single transmission. The economic gain for this point-to-multipoint distribution in terms of cost per user/receiver is phenomenal and easily surpasses any other media-fiber, cable/coax, copper or wireless local loop.

However, for two-way interactive connections, satellites require a return channel from the user location, which significantly impacts the economic equation. Equipment costs are much higher than for their receive-only DBS cousins and perhaps more importantly, satellite bandwidth costs quickly dominate. Additionally, for non-spot beam systems, throughput capacity can quickly become a bottle-neck and served subscriber density drops significantly. The following tables and figures illustrate the economic differences between Ku- and Ka-band systems given the stated assumptions.

Table 5 - Ku-band Economics

As shown in Figure 2, the primary cost categories in offering service are space segment, service and operational expenses, and end-user equipment. As subscriber volume increases, equipment costs will fall to within $300 to $350, which will enable subscribers to purchase equipment without subsidies from service operators. Current Ku-band service providers subsidize the terminal cost to their customers, as the true terminal cost is in the $800 to $1,150 range.

Table 6 - Ka-band (bent-pipe) Economics

Figure 2 - Service Costs per Subscriber by System Type

Key to reaching the $300 equipment cost level, is a strategy that uses key components that are already used in high-volume set-top consumer units, such as the DVB-S technology used for digital satellite TV or the DOCSIS technology used in the very large cable modem market.

The most notable economic difference between these two systems is the bandwidth per user cost basis (Mbps/subscriber/month). The bent-pipe Ka-band system approach enjoys a huge 82 percent savings over the Ku-band system. All of this savings can be attributed to the greatly reduced cost of air-time (space segment) for the Ka-band system. Thus broadcast (FSS) satellites are much more expensive than Ka-band spot beams for providing 2-way bandwidth intensive service.

Within the general telecom industry, studies show that a one percent decrease in costs results in a three percent increase in demand. Applying a similar model to two-way satellite broadband, the significant reduction in air-time costs with Ka-band systems could stimulate a two to four-fold increase in service demand. This greater demand yields increasing service revenues which in turn significantly increases the likelihood of business success.

The key to a successful service offering is attracting and keeping a satisfied subscriber base that is in excess of the breakeven points presented. The primary restriction with a Ku-band offering is the limited system capacity, which inherently limits the number of subscribers and therefore makes economic success more risky.

Both types of systems can further increase service revenues by augmenting basic Internet access with premium services, such as specialized content delivery and media-casting/streaming. Furthermore, it's likely that higher bandwidth service levels for power-users will be offered along with quality-of-service (QoS) guarantees.

return to top

Performance & Trade-offs

Internet access and networking services will have an increasing dependence on high-bandwidth capacity. And, as shown in Figure 3, the expected bandwidth capacity need will increase three to four fold over the next few years. This demand for continual performance increase requires that service providers plan accordingly and carefully evaluate their overall implementation approach so that their businesses have longer term viability.

The commercial success of a satellite broadband service offerings will be closely linked with three key factors:

All of the above factors are significantly influenced by the overall system design and implementation. The first is closely inter-related as illustrated by a hypothetical system loading curve illustrated in Figure 4. System throughput and user demand vary significantly as a function of time of day and the types of applications used by the end-users. The application type affects the amount of data to be transferred, the timeliness of information and the number of concurrent applications to be supported across the network or a geographical area (e.g., within a particular spot beam).

Figure 3 - Average Bandwidth Demand for Broadband Subscribers

Figure 4 - System Loading & Usage

By implementing a system that supports dynamic resource allocation in a controlled, easy to administrate fashion, a large number of users can be supported with a consistent quality-of-service experience. This approach relies on the basic premise that most subscribers don't use peak capacity all the time. In fact, dedicating a fixed amount of system bandwidth by data rate to each subscriber (or subscriber type) to address the quality-of-service need, is not only an inefficient use of overall system bandwidth, but it severely limits the number of end-users that the system can support. Furthermore, simple "peak-to-average" allocations are not sufficient to address the high demand for broadband access.

New advanced bandwidth-on-demand techniques allow for:

• flexible bandwidth allocation-e.g., high bandwidth applications get more system resource attention than web surfing

• prioritized bandwidth allocation-e.g., real-time applications over low-priority file down-loads

• "bandwidth pools" shared by many subscribers ensure large user population usage

• quality-of-service flexibility: priority based and connection based

• tiered service offerings:

  • enterprise customers: "metered" usage, by-the-minute or Mbyte

  • residential: flat rate usage, "all you can eat"

Many of these challenges are being addressed in the terrestrial broadband marketplace, especially by the hybrid fiber-cable system operators and equipment manufacturers. Within the United States the development of products based on Data over Cable Service Interface Specification (DOCSIS) 1.1, cable networks is moving from the best-effort service defined in DOCSIS 1.0 to the delivery of guaranteed service level agreements (SLA) for critical business applications. By implementing end-to-end Quality of Service (QoS) controls, cable system operators are:

• expanding their customer base by offering a wide variety of business and residential services

• building increased customer loyalty by offering bundled services supporting voice, data, audio, and video traffic

• creating multiple revenue streams from their HFC network

The second key factor in network design is addressing the largest cost in implementing broadband services, space segment. A partial remedy lies in implementing the latest in transmission channel modulation (e.g., 8-PSK) and coding techniques (e.g., turbo product codes) to maximize the bits-per-hertz signal density. These techniques produce considerable improvements. Even the DVB-S standard is in the process of adopting 8-PSK and turbo product codes. To further maximize system throughput and service availability, especially when dealing with Ka-band systems, many of the transmission channel parameters (e.g., coding rates, data rates, link power) need to be adaptive and adjusted in an intelligent manner on a dynamic basis. This capability is being built into the subscriber terminals and the associated network control system.

The third key factor in deploying a success broadband business is the most often overlooked within the satellite industry-provisioning the operating business functions. Figure 5 illustrates the system hierarchy for any broadband service offering, whether satellite or facilities based. Table 7 provides a brief explanation of each level.

Figure 5 - Broadband System Hierarchy

Table 7 - System Hierarchy Elements

Traditionally the satellite industry has focused only on the bottom three aspects (network management, media access and the physical {satellite channel} layer) and has ignored the operations and business support systems. However, service providers know differently, and the entire hierarchy must be implemented successfully to have a viable service offering. The early success of the initial digital cable modem rollouts in the United States was enabled by the emergence of the DOCSIS standard which addresses all levels of the hierarchy.

While the DVB-S standard has made possible the success of the global digital television broadcasting market, the underlying service is one-way. The Starband and DirecWay systems use the DVB-S standard for the outbound broadcast channel and use their own proprietary technology for the return channel.

In early 1999 an ad-hoc group was formed to facilitate a standard for a return channel via satellite, DVB-RCS. The DVB-RCS specification provided definition of the various network independent layers (e.g., physical and MAC) only and left the network management and offered services for the network operators and service providers to define. The DVB-RCS is emerging as one of the baselines for broadband satellite services as recently deployed by Astra-Net in Europe. Standards are important for a number of reasons:

• enable multi-vendor participation and adaptation

• facilitate high volume production in a competitive environment, ultimately leading to mass-market (low-priced commodity) products

• enable market growth as services/products become widespread.

WildBlue has based its Ka-band system design on a satellite-enabled version of the DOCSIS 1.1 standard. This approach provides four primary advantages: 1) low cost subscriber terminals by leveraging very high volume chip sets; 2) fast time to market through minor modifications to existing chipset design spins; 3) immediate availability of a very mature set of infrastructure products for network control, system management, subscriber management, and billing systems and 4) the ability to leverage the huge investment in advanced networking features in DOCSIS 1.1 that support QoS and other advanced networking features.

In sharp contrast to the consumer market based WildBlue system is the Astrolink system which is focused on the corporate enterprise networking and Internet access market. Astrolink plans to provide:

• data, video and voice services that support business applications;

• interactive or two-way high-speed connections

• point-to-point service, as well as multicasting content delivery services

The Astrolink system approach puts advanced packet switching technology onboard the Ka-band satellite. Using the asynchronous transfer mode (ATM) protocol, the Astrolink network will be able to accommodate multiple types of data, video or voice traffic. ATM's ability to guarantee quality-of-service levels has led to its widespread adoption by the telecommunications industry. In addition, ATM makes it possible to bill customers for their actual network usage if they so desire. By paying only for the bandwidth they use, when they use it, Astrolink will be able to offer customers significant service cost savings. Ka-band operation ensures that network terminal antenna size remain attractively small and low-cost.

return to top

Summary

Globally there is a strong need for new two-way broadband systems to reliably deliver IP-based services to large numbers of residential subscribers and enterprise users. Current infrastructures do not provide the necessary capacity, reach capability and service price points to satisfy this growing demand in all geographical areas-and satellite based systems are being deployed to provide the needed broadband "local-loop" service. Because of the diverse market requirements-service and network terminal costs, types of services, throughput performance and service quality levels-no single broadband system can address multiple market segments.

To achieve commercial success, service providers must tailor the network implementation to fit the market needs. The limited success of the first generation satellite broadband systems is restricted in part by the high-cost of satellite space segment and the less-than-optimal network throughput and operational performance. The new second generation systems are customized to address their target markets. With significantly reduced bandwidth costs and greater subscriber population capabilities, these second generations systems have a significantly greater chance at achieving commercial success.

return to top

Web satelliteonthenet.co.uk