Chapter 2: Technical Viability¹

I. Introduction

The telecommunications landscape is changing rapidly, in technical, economic, and institutional arenas. This section concentrates on technical options that could strengthen the Pacific.

Many considerations should be addressed when designing new, comprehensive, systems. This is certainly true in the Pacific, where investments have already been made in such technologies as (a) over-the-air and/or cable broadcasting, (b) two-way radio communications for outlying areas, (c) C-band satellite communications infrastructure between metropolitan areas and internationally, (d) analogue or digital microwave mobile telephone services, and (e) copper and limited fibre-optic cabling. How much of such infrastructure should be expanded or at least maintained – long-term or for transitional periods? How much of such infrastructure should be replaced – quickly or in the longer term?

To dispel such apprehensions, the (admittedly early) experiences reported by the Eastern Caribbean Telecommunications Authority, and by other small States such as Mauritius, is that new technical opportunities, coupled with new business/service models, offer unprecedented opportunities for small island economies to benefit from improved telecommunications.

II. Technical Options for the Pacific

Findings from this study result from a synthesis of previous studies on different aspects of telecommunications infrastructure development, strengthened by additional context from socio-economic analysis of all Pacific island economies; compilation of information from sources on connectivity in the Pacific; relatively new perspectives on redeploying surplus cable capacity; and new developments in satellite communications and wireless networking.

Connectivity infrastructure may be broadly categorized as wired, terrestrial wireless, and satellite. Each of these different types of “pipe” has traditional analogue, narrowband digital, and broadband digital capabilities, and can provide broadcasting, telephony, Internet and conferencing services. Technical options include those shown in Table 2-1.

Wired connectivity has evolved from copper to fibre-optic cable, though new technologies have prolonged the life of copper in the “last mile” through the use of digital subscriber line (DSL) technology. Although previously considered unaffordable for small Pacific island economies, undersea fibre-optic cable is nevertheless considered highly desirable for its cost-effective and relatively reliable bandwidth. Although relatively few small Pacific island economies formerly benefited from broadband cable connectivity, this situation is improving as several cable projects are underway. Several more countries are near to existing or planned cable nodes – offering the potential to build short, potentially affordable, extensions to serve their people.

Terrestrial wireless connectivity used to mean two-way radio, shortwave and AM/FM radio and television. It now adds mobile phones, and in some countries, between-island microwave backbone and/or WiFi/WiMax end-user connectivity. Mobile devices are growing in popularity at the fastest rate among ICT tools in much of the world, and should be seriously considered for their lower cost and greater reliability, in comparison with personal computers, and rapid enrichment as end-user ICT devices.

Satellite communications provided a leap forward in Pacific connectivity about three decades ago. Since then, costs have dropped somewhat, and C-band connectivity has been adapted for digital as well as broadcast capabilities. However, the recent revolution in cost-beneficial satellite broadband, coupled with great improvements in terminal performance/cost, has not yet substantially benefited the Pacific. A no-frills satellite communications service model, designed to maximize service/cost for the Pacific, now appears possible and may be able to significantly improve the financial sustainability of such a service. In the short term, Pacific States may wish to form user consortia to negotiate grouped bandwidth (at lower rates for the consortium than may be achievable individually by the States).

Table 2-1. Technical options for Pacific connectivity “pipes”: A summary

Pipe Media	Media Details	Generalized Description
Wired	Fibre-optic cable	This is the most desired and cost-beneficial connectivity for sufficiently large markets. The market size necessary for cable appears to be dropping, making cable possible for increasingly small communities. First-generation cable is being retired due to excess capacity in the marketplace. Redeployment of such cable may be an affordable opportunity for small markets.
Wired	Copper cable	This might be largely obsolete for international connectivity. Where already installed, this might be maintained, awaiting replacement by fibre, wireless or broadband satellite. In some communities, DSL over copper may justify installing or upgrading the copper infrastructure for such a purpose.
Terrestrial wireless	Microwave, wireless phone, WiFi, WiMax, Wireless LAN, etc.	The revolutionary connectivity brought by WiFi/WiMax has been extended to several kilometres from each transmitting node. Such approaches enable a single transmission node (satellite terminal or broadband cable) to cover larger dispersions of people – thus making them more cost-effective. Microwave, limited by its relay requirement for every 50 km, could be used as backbone connecting neighbouring areas.
Wired	Terrestrial two-way radio	This “pipe” may be the least capable and reliable option today, but has long served distant, dispersed populations with basic telecommunications. Where other means become cost-effective, this might be replaced. However, modernized two-way radio, including amateur radio, may continue to serve communities for years to come.
Satellite	Communication satellites	For an appropriately located satellite, this can bring “universal service” more practically than cable where populations are widespread. Existing C-band services are reliable, and may be worth continuing for high-reliability needs, though a satellite phone might also provide basic backup at lower cost in some locations. Newer Ku-band may be lower-cost for higher bandwidth yet adequately reliable for Pacific island environments – for all but the most essential basics. New approaches suggest that a dedicated satellite for the Pacific can now improve performance, yet be cost-effective. New approaches to jointly negotiating for one or more existing satellite transponders (by collectives of Pacific economies) may bring economies of scale to several countries.

III. Communication Services

A. Basic Internal and International Connectivity

Communications for Pacific island economies can be broadly divided into local communications within a community, nearby communications within an island, national trunking communications farther away in a single country, and international communications.

Local communication within a community may be accomplished by wired or wireless telephony and/or Internet. In many cases, wireless networking may be ideal for communities with little legacy wiring to support – or even when such wiring is in place. WiMax may serve cities and their surroundings, supporting flexible connectivity/networking with more types of device than traditional cellular wireless.

Nearby communications within an island may be similar if the island is sufficiently compact and flat (e.g. an atoll), but might involve trunking similar to inter-island communications if line-of-sight communication is made more challenging by distance or terrain.

National long-distance trunking is likely to involve local networks at each end similar to local communication within communities. Wired, wireless or satellite trunking can serve between communities. The choice of trunking may be influenced by pre-existing infrastructure. For example, legacy satellite communications may be expensive, low-bandwidth (adequate for current use, but perhaps inhibiting growth), and beset by latency that is inconvenient but can easily be adjusted to. Legacy copper cable with analogue switching, or two-way radio, may be of limited use for modern applications. New approaches may be worth encouraging.

International trunking may be similar to national trunking, with national networks at both ends linked by wired, wireless or satellite trunking between communities and nations. Established Pacific pipe is likely to be C-band satellite communications, other than in the few countries with undersea cable. C-band trunking may be worth replacing or supplementing with greater capacity and cost-efficient Internet Protocol trunking by cable or satellite – at least in the long term.

B. Applications

Applications include radio and television broadcasting, audio and video conferencing, including business, government and education (such as international lecturing); voice (including local and trunking services); Internet (including distributed data management and file sharing); direct-to-home/office services; e-learning including distance education; e-health including tele-medicine; e-governance/government and public services; e-commerce and business services (including locally or globally collaborative open-source software development and support); and community services such as Internet cafés and e-centres carrying social/enabling services, including disaster management and entertainment.

Such applications may be grouped into the broad categories shown in Table 2-2, and may often be suitably delivered by whatever international connectivity pipe (in Table 2-1) is available.

Note that services formerly associated with specific media (such as over-the-air for radio and television, copper cable for telephony and Internet) now may be delivered by a diversity of “pipes” from satellite, cable and terrestrial wireless – or combinations of these. In the end, revenues, and the delivery of compelling telecommunications products and services, will depend on creativity resulting from competitiveness and focus on customers.

Table 2-2. Generalized types of telecommunications services

(ordered by current popularity)

Broadcasting (radio, television) is traditionally received via over-the-air broadcasting, but may also be accessed over the Internet (as streaming audio and video), via new mobile wireless, and via satellite. This may be via direct-to-home or cable subscriber systems. Much broadcast content is fed from content developers to service providers via satellite.

Telephony was traditionally circuit-switched over fixed telephone lines, though new wireless mobile services have surpassed fixed lines in many markets. Rapidly increasing in popularity, Voice-over Internet Protocol (VoIP) services are often economical, and may be delivered to users through fixed and mobile services, or through users’ computers. Satellites have been supporting the expansion of both circuit-switched and VoIP telephony services to less-developed and remote areas. Satellite phone services cover many areas that lack alternatives. Though costs of the latter have been dropping, they are considered expensive for some users. Many parts of the Pacific are still served by two-way radio telephone, which has little opportunity to go beyond traditional voice calling or conferencing.

The Internet is normally delivered by copper or fibre cable to service nodes, and then to users by many approaches, such as cable, DSL, terrestrial wireless and satellite.2

Audio and video conferencing have been implemented in various forms, from simple conference telephone calls, to ISDN and Internet-based video conferencing. Satellite-based systems have also been used, as have two-way radio-based audio conferences.

IV. Findings

A. Wired “Pipe”

Though many submarine cables cross the Pacific, most are clustered across the North Pacific between Asia and North America, or run between south-eastern and north-eastern Asia, or between Asia or Hawaii and Australia/New Zealand. To date, such cabling has bypassed most Pacific island economies, and it recently demonstrated its vulnerability when the earthquake of 26 December 2006 damaged several cables at once, severely constraining connectivity in many countries. For this reason, the cabling industry might consider diversifying its routings (a) to better protect itself against such catastrophic breakage and (b) to better diversify the markets it serves. Such diversity of routing might pass through the northern, central, or southern Pacific. Such strengthening of routes can also be routed to facilitate connections to Pacific island economies. Figure 2-1(a-c) illustrates such a view of current and possible next-step cabling³.

It is notable that several storms, as well as the recent unexpected complete failure in November 2004 and January 2005 of Intelsat communications satellites IA-7 and IS-804, have also cut off peoples, including Pacific island users/economies, from the telecoms world. Therefore, a single satellite or cable is not yet the “unique” solution. Rather, affordable, integrated alternatives of different characteristics may today provide the best “fail-safe” system against complete isolation in the case of single-system failure. An ITU-PITA project has addressed this issue, emphasizing the necessity to establish and implement regional strategies for contingency planning, including the development of systems to rapidly switch between satellites in case of a difficulty or failure in one option.

Increased partnering is resulting in some cabling projects, such as those announced or underway to link the Federated States of Micronesia and the Marshall Islands to Guam, and New Caledonia to Australia. In addition, the redeployment of a portion of the PacRimWest cable to Papua New Guinea (discussed later in this chapter) offers an exciting first example of such an approach – which could potentially benefit several other countries. The cost savings of such redeployments, and the attention that it might gain, can be significant strategic assets to countries that may be prepared to truly benefit from connectivity.

Figure 2-1a. Illustration of the Japan-USA bottleneck attributed to imbalanced trans-Pacific capacities
Source: Barney, Bill 2007. Crisis, Opportunity and the Submarine Cable Industry. Proceedings PTC'07, Honolulu, USA. Mr. Barney is president and CEO of Asia Netcom.

Figure 2-1b. Proposed cable routing bypassing the Luzon-Taiwan Strait
Source: Barney, Bill 2007. Crisis, Opportunity and the Submarine Cable Industry. Proceedings PTC'07, Honolulu, USA. Mr. Barney is president and CEO of Asia Netcom.

Figure 2-1c. A multi-ring architecture, reducing dependency on any one potential failure route³
Source: Barney, Bill 2007. Crisis, Opportunity and the Submarine Cable Industry. Proceedings PTC'07, Honolulu, USA. Mr. Barney is president and CEO of Asia Netcom.

In the 1990s, great enthusiasm developed for cabling “solutions”, mostly across the North Atlantic and Pacific oceans. Around the turn of the millennium, however, this enthusiasm turned into a potential crisis, as the “bubble” burst and many pioneers wound up in bankruptcy, and their systems wound up in the hands of new operators, who acquired them for pennies on the dollar. These operators can provide services for a much lower cost point, yet still make a profit. Enthusiasm returned to the cable industry by 2007, with several recent announcements on new trans-Pacific cable initiatives. The earthquake of 26 December 2006, which severed several cable links, resulted in a call for a more fail-safe cabling infrastructure (as partly illustrated in Figure 2-1), and the realism that grew out of the 1990s boom and bust may help the industry develop improved sustainability and improved marketing (including the serving of additional markets) in its next build-out.

1. Local Cabling Initiatives in the Pacific

Guam and Fiji have undersea cables. Several additional countries have cabling projects underway or under discussion – and the approaches taken by those countries may set an example for other Pacific States, for both international and internal connectivity.

The Marshall Inter-Island Cable was installed in 1992, and the Palau Inter-Island Cable was installed in 1996, dispelling the view that small economies cannot develop a domestic fibre-optical backbone. The MTC Marianas Cable connects Saipan,³ Guam and intermediate islands, dispelling the notion that small economies cannot lay domestic and international backbones. Though the MTC is only 240 km long, spurs or loops of slightly greater length could link several Pacific island economies to current or envisaged trans-oceanic cables (Figure 2-3, Tables A-2 and 2-3). The governments of the Federated States of Micronesia and of the Marshall Islands have begun construction of a cable between Guam, Majuro and some islands between these end points. On the drawing board are plans to extend this system. Papua New Guinea’s new APNG-2 cable connection is described just below.

These changes bode well for other Pacific economies. Several economies are relatively near to the Southern Cross Cable Network landfall in Fiji, and could build local or subregional loops from Fiji to serve themselves. Early discussions about another major cable between Australia-New Zealand, the United States of America, Guam and Asia might be an opportunity to pursue additional local loops. Redeployment of first-generation fibre-optic cables (labelled “Reuse?” in Table A1, Appendix A), along the lines of APNG-2 to Port Moresby, is another opportunity, sketched later in this chapter. Other plans, involving French Polynesia and several other states are in development. Such activities merit attention by Pacific states (who should probably not “sign the first contract offered” but negotiate aggressively, perhaps with knowledgeable and unbiased help, to ensure that they do not jump irrationally at the first offer by overpaying, or over-committing of state resources, for such connectivity).

The cost for deployment and maintenance of current-generation fibre-optical cable is quite high, so it is suitable for high-traffic situations, including connecting Pacific Rim economies to one another. Two options appear viable for the Pacific:

(a) Partner with trans-Pacific cable projects, to include spurs to one or more Pacific islands, as Fiji was connected to the Southern Cross cable. This may be good for the public relations of cable operators, and such operators may seek additional diversity of routings as they respond to the cable cuts of 26 December 2006. Indeed, Palau, Guam, the Marshall Islands and perhaps others are situated along the path of the Proposed AsiaNetcom cable illustrated in Figure 2-1;

(b) Consider redeployment of first-generation undersea cabling as noted immediately below. This may be much more economical to deploy and reasonable to maintain.

2. Redeployment and Reuse of Undersea Cables

First-generation electro-optical fibre-optic submarine cables, installed between about 1988 and 1995, pioneered a revolution in connectivity. Such cable was replaced in the mid 1990s with second-generation purely optical cable technology with much greater capacity. Because of in-place upgradeability, innately greater capacity, and a surplus of overall capacity installed in the past decade, operators of first-generation cables have retired them prematurely, or are considering doing so.

Such retirements, however, are occurring long before such cables become technically unserviceable. These cables may be redeployable for Pacific island economies with suitable requirements.

Cable Case History #1: The academic community reuses undersea cables

There are plans to reuse first-generation submarine fibre-optic cables for two purposes. One is for research, and the other for operations supporting research. Butler (2003)⁴ discussed a study by the Incorporated Research Institutions for Seismology (IRIS) on the potential of economically repositioning first-generation cables for new uses. Used cabling can be redeployed for about 10 per cent of the raw materials costs, and about 20 per cent of former peak-demand ship time. The report notes that Pacific cabling laid in the late 1980s and early 1990s might be decommissioned as surplus,⁵ and might be negotiable for redeployment and reuse. Such cabling has nominal bandwidth between 280 and 560 Megabits/second (as opposed to 1,000 times that bandwidth or more for recent-generation cabling). IRIS has formed IRIS Ocean Cable Inc. to negotiate the assumption of management of several such cables upon their decommissioning. It has taken over some of these systems and is using them for scientific purposes.

Clearly, if the scientific community can negotiate to take over such systems, such developments indicate a potential for redeployment for Pacific islands, as well. Additional possible cables that could be redeployed to connect Pacific island economies include GPT, PacRimEast, the remainder of PacRimWest after part was taken to make APNG-2 (see below), TPC-3 and TPC-4 (see Appendix A). Pacific Island economies that could be connected in such ways could include American Samoa, the Cook Islands, the Federated States of Micronesia, French Polynesia, Kiribati, the Marshall Islands, Niue, the Northern Mariana Islands, Palau, Samoa, the Solomon Islands, Tonga, Tuvalu, and Wallis and Futuna. An article aimed at scientific⁶ use could be adapted as an overview of how Pacific island economies could be connected using such cables. Indeed, a pioneering initiative of that type has taken place (see below).

Cable Case History #2: Papua New Guinea – A new pro-activity

Until recently, connectivity in Papua New Guinea was via the 897-km long Australia-PNG (APNG) cable, a copper coaxial analogue telephone cable laid between Cairns and Port Moresby in 1967, and by Intelsat Earth stations at Port Moresby and Lae. However, not long ago APNG was retired from service, leaving the country and Telikom PNG with only satellite connectivity. Recently, Telikom PNG announced a partnership effort to acquire 1,800 km of the recently retired PacRimWest cable, previously linking Sydney and Guam, and redeploying it between Sydney and Port Moresby, naming it APNG-2 (see redeployment map in Figure 2-2), anticipated to be serviceable for at least 15 years. Its cost has been estimated at US$11 million, compared with an estimate of US$60 million to acquire and lay new cable. Its anticipated capacity of 560 Megabits/second is considered roughly 70 times that of the old copper analogue cable. A trial recovery of 300 km of PacRimWest cable, conducted by contractor Alcatel, found that the retired cable and its repeaters were in good condition. Completion of the laying of APNG-2 was announced in October 2006, with initial services commencing soon thereafter. Telikom PNG also has plans for an additional 18 Mb/s of satcom capacity in the near future,⁷ of which 8 Mb/s has already been installed.⁸ In addition, broadband VSAT terminals are being eyed for connecting about 25 additional communities, with new VSATs already installed at Tufi, Bogia and Pangia.

In this estimate, a cable landing could be established in most Pacific island economies for a total of roughly US$50 million, considerably less than previously estimated – by redeploying first-generation cable that has been or is soon to be retired.

Telikom PNG is coming to the end of its current regulatory environment, during which it enjoyed a virtually monopoly situation. It is thus positioning itself to act more pro-actively, in preparation for an envisaged climate of freer competition. One move has been to invest in management training for its staff. Another has been to aggressively expand mobile phone service coverage. Another move in late 2006 was to make unlimited Internet use available to universities, colleges and schools for US$275/month – a service it plans to extend to other non-profit organizations such as hospitals. Telikom states that such access is being provided as part of its community service obligation programme, and is being timed to coincide with increased capacities anticipated from APNG-2.

Figure 2-2. Redeployment map
Source: Hibbard, John, Recovery and Re-Lay of Submarine Cables. Presented at PTC'07, January 2007

3. Recommendations on Cable

Given the opportunities, and in spite of uncertainties inherent in the task of laying cable, the Pacific should look seriously at potential partners toward acquiring their own international cable landings – especially in light of the cost savings from the redeployment of first-generation cable, and the potential interest of developmental partners in seeing improved connectivity for Pacific islands. Figure 2-3 illustrates, and Table 2-3 lists, Pacific island economies, ordered by population but also showing the Human Development Index.⁹.

Figure 2-3. Sketch of possible cabling spurs (dotted red lines) to reach Pacific Island economies

The following example may illustrate the possible use of Table 2-3 for strategizing cable deployments: Tonga might be connected to Fiji, which is on Southern Cross, with about 700 km of cable – costing perhaps US$3 million if first-generation cable is redeployed as exemplified by APNG-2. Niue, Samoa or American Samoa could be next, connected to Tonga (after Tonga was connected), with a few hundred kilometres of cable costing perhaps US$2 million per State.

Table 2-3. Present and possible future international cable connectivity in the Pacific ¹⁰

State	Population 1,000s	HDI	Cable Date	Length (km)	Sequence?	Cost Estimate US$M	Notes
Papua New Guinea	6,002	0.523	2006	3000	Operating	11	Pioneering redeployment of PacRimWest
Timor Leste	1,063	0.513		~800	Next?	3??	From Darwin?
Fiji	906	0.752	2000	hub	Operating	20?	Short loop, on Southern Cross
Solomon Is.	552	0.594		100	Next?	2??	Redeploy PacRimWest – as spur to APNG-2?
Fr. Polynesia	275	0.78		1000	4?	4??	Redeploy PacRimEast?, extension beyond Cook Is., or direct from Hawaii?
New Caledonia	239	0.79	2008?	2000	Announced	53	Gondwana-1 cable announced from Sydney
Vanuatu	218	0.659		300	Next?	2??	Extension from New Caledonia using redeployed 1st-generation cable?
Samoa	183	0.776		500	2?	2??	From Wallis and Futuna or Tonga, from Fiji?
Guam	171	0.9	1989		Operating	10?	Hub with many cables
Tonga	115	0.81		700	Next?	3??	From Fiji, redeployed 1st-generation cable?
Micronesia	108	0.61	2007?	2175	Being laid	67.4 (shared)	From Guam, combined with Marshall Is.
Kiribati	105	0.61		600	Next?	3??	From Majuro, redeployed 1st-generation cable?
N. Mariana Is.	83	0.84	1997		Operating		From Guam
Am. Samoa	63	0.81		500	2?	1??	Extension (with Samoa?) from Wallis and Futuna?
Marshall Is.	60	0.62	2007?	2175	Being laid	67.4 (shared)	From Guam, combined with Micronesia
Palau	22	0.76		800	Next?	4??	From Guam
Cook Is.	21	0.72		1000	3?	5??	Extension from Niue, redeployed 1st-generation cable?
Wallis & Futuna	16	0.71		700	Next?	3??	Side loop from Fiji, extensions to Samoa, Tuvalu, etc; redeployed 1st-generation cable?
Nauru	13	0.71		700	2?	3??	Extension beyond Kiribati from Majuro, redeployed 1st-generation cable?
Tuvalu	12	0.67		700	2?	3??	Extension beyond Wallis & Futuna from Fiji, redeployed 1st-generation cable?
Norfolk Is.	2	0.93		?
Niue	2	0.78		300	2?	2??	Extension beyond Tonga from Fiji, redeployed 1st-generation cable?

Note: Completed, underway or announced projects cost ~US$160 million (not including N. Mariana Islands).

B. Terrestrial Wireless “Pipe”

Terrestrial wireless is invaluable for last-mile user connectivity in various parts of the world. In some Pacific island economies it may be used for some domestic trunking.

Mobile phones and related services (short messaging service, Internet and broadcasting via mobile phone, etc.) is the fastest growing telecommunications product/service in most developing countries. In the Pacific, it has an advantage over the Internet (which may get more publicity despite its lower growth and penetration in many markets), due to the relatively lower cost and greater ease in purchasing and servicing a phone, compared with computer equipment – in most Pacific markets at present.

In addition to the desirability of delivering popular and enabling mobile phone services to people of the Pacific, terrestrial wireless may be useful (in the form of WiMax or other wireless networking) in delivering the Internet and IP-based services, such as low-cost Voice-over IP telephone services. In this mode, connectivity brought by VSAT or cable to a community could serve an area up to several kilometres from the connectivity point – by means of wireless networking.

Terrestrial microwave wireless has been used to provide a “backbone” between islands, where they are sufficiently close together for line-of-sight transmission between microwave towers.

All of these approaches to terrestrial wireless have been implemented in the Pacific, as noted in Appendix B, and offer the potential to serve additional Pacific communities.

Wireless Case History #1: Niue – Gratis non-commercial WiFi, Internet etc.

With fewer than about 2,000 residents on a single low-lying coral atoll of 259 square kilometres, vulnerable to tsunamis and other disasters, Niue is located 400 km from its nearest neighbours, which are also small island economies. Most of the country’s citizens are living overseas because opportunities are scarce in the country. Providing telecoms infrastructure to such a small nation has been a challenge. The country outsourced the licensing of its .nu domain name to an overseas company. The .nu domain name is considered “trendy” in several other countries (primarily in Scandinavia, where it is associated with “new”).

Niue has been offering free local email to residents since 1997, free non-commercial dial-up Internet supplemented by international satellite communications link from 1999, and free non-commercial Wi-Fi wireless broadband Internet to all residents and visitors (including owners of yachts in the harbour) from 2003. The country was severely damaged by 300-km/hr winds from Cyclone Heta; but quickly restored the Internet service, thanks in part to help from its neighbours in the Pacific. This is arguably the first nation with full Internet connectivity, free (and license/contract-free) for all residents and visitors with Wi-Fi capability. A gratis Internet café is operated in Alofi, for the use of those lacking their own computers.

People involved in the process may describe¹¹ Niue’s situation as unique, but some aspects appear adaptable to other communities, and it demonstrates that high technology can be appropriate for small, isolated communities, if they can combine local conditions with capabilities developed in the global ICT community – and develop a supportive policy environment. The business model – revenues from domain name licensing helping to underwrite connectivity improvements – might be applied to a few other countries in the Pacific. The service model of “universal” access to wireless Internet, is being taken up elsewhere, with New Orleans installing free wireless city-wide after hurricane Katrina, Rhode Island and Vermont both aiming to become the first American state to have complete statewide wireless networking, the Former Yugoslav Republic of Macedonia reportedly covering 95+ per cent of its population with wireless broadband networking, and Tonga Communications Corporation pursuing national WiMax coverage.

Wireless Case History #2: Microwave LAN: Connecting Tongan islands with broadband

Comprising 36 inhabited rugged volcanic and low coral islands, plus about 140 additional islands, Tonga faces a challenge in building communications infrastructure. However, recently a public-private partnership brought GSM wireless networking to several islands, giving them mobile telephony and broadband Internet. Using existing commercial capabilities, several islands have been connected with scalable bandwidth. Each connected island has a satellite link with about 1 megabit per second bandwidth. The project recently received the GSM Association Award for best infrastructure or network solution product for its delivery of Internet, GSM telephony, and entertainment services around the country. Schools have been connected, to bring educational opportunities to the country’s people. Shoreline, a Tongan ISP, has been rolling out such wireless network assess¹² across the islands, with satellite or wireless linkages between islands. Tonga is also getting a WiMAX network to coordinate with the GSM network of Tonga Communications Corporation, currently serving part of the country.¹³

C. Satellite Communications “Pipe”

1. Satellite for the Pacific

Because of the traditional views of satellite communication, people may overlook the improved cost-effectiveness of the technology in addressing the ICT requirements of rural and remote areas and islands, so alleviating problems associated with inadequate connectivity. In the long run, the combined commercial and non-commercial benefits of expanded telecommunications with a satellite platform should far exceed the direct cost involved. The resulting connectivity could result in significant national developments, thus reducing the perceived obstacles of isolation, rural poverty, general social inequity and increased urban/overseas migration.

Satellite communication by its very nature can provide “universal” connectivity for urban and rural areas. The biggest advantage of satellite is that it can provide instantaneously deployable (and redeployable) infrastructure, assuming a suitable satellite is available, so that it can be used for point-to-point, point-to-multipoint and for broadcast applications. Solutions exist for sparsely distributed areas such as the Pacific, and suitable purely satellite, or satellite/cable/wireless approaches may be implemented in several combinations. Depending on the amount of traffic, one or two satellites can be designed to meet the requirements of Pacific islands. This might include deploying a small or medium-sized satellite to meet first-phase projected requirements and adding a co-located satellite later for enhanced capacity. Such an approach, though slightly more expensive at the outset, in comparison with a single larger satellite, in the long run will be more commercially viable due to greater flexibility and the increased security of redundancy. Other solutions may involve a range of contracts with satellite service providers, with a number of satellites being used over time according to capacity, pricing and demand considerations – and of hybrids with cable perhaps being built to high-traffic areas as a satellite approaches full utilization.

Given the nature of the Pacific climate, it is noteworthy that satellite connectivity is increasingly independent of weather and local conditions.¹⁴ In recent times, satellites have been used for societal development applications like telemedicine and tele-education, all complemented by dramatic reductions in mobile phone and computing costs. A single satellite platform can be designed to have communication, entertainment and societal applications such as telemedicine and distance learning, since all applications use digital data and thus can share storage capacities. Recent trends in satellite communications, which involve higher modulation and coding techniques and use of multiple spot beams supported by a shaped beam, make satellite communication increasingly competitive with terrestrial systems.

Satellite Case History #1: Internet Protocol communication satellites

Anik F2 was launched on 17 July 2004, marking a new generation of broadband Internet Protocol communications satellites. It has 45 downlink Ka spot beams covering Canada (except the Arctic islands) and the 48 lower states of the United States (not including Alaska or Hawaii). It has six uplink beams. Total Ka bandwidth is 9 Gbps, while C and Ku bandwidths are each about 1 Gbps. The satellite weighs about 5,950 kg, sits over 111.1°W longitude, and has an anticipated service life of 15 years. One estimate is that about 750,000 individual users can be served by the system. In the United States, Anik F2 bandwidth is being marketed by wildblue.com, and in Canada by telesat.ca. Wildblue packages reportedly start at US$50/month for 512 kbps download and 128 kbps upload maximum. A 30-cm (nominal) dish and hardware normally cost about US$300; a 60-cm dish for fringe areas will cost somewhat more, not counting the cost of required professional installation. In late 2006 there was a promotion offering gratis installation, and a reduced price of $200 for the terminal. Wildblue 1, the first exclusively Ka-band broadband satellite, weighing about 4,735 kg, was launched on 5 December 2006. It is also designed to serve dispersed populations in rural North America.¹⁵

IPStar was launched a year after Anik F2. Estimated prices are reported by Shin Satellite to be similar to those for Anik F2, with 90 Ku-band spot beams over the most populated portions of southern and eastern Asia and Australia. Unfortunately, there is only a broad beam covering some Pacific island economies.

Satellite Case History #2: Edusat

The Indian Space Research Organization (ISRO) launched Edusat in September 2004, the world’s first communication satellite dedicated to educational applications. With five Ku-band spot beams, one Ku-band and five C-band shaped beams with national coverage, Edusat weighs 1,950 kg and cost US$20 million to build. The Indian launch vehicle cost US$33 million to build, with the total launch priced at US$45 million.¹⁶ This is an example of how costs can be contained when affordable service delivery is a main objective.

Satellite Case History #3: French Polynesia and Cook Islands “licensing consortium”

OPT, the French Polynesia telecoms operator, and Telecom Cook Islands have joined in a form of licensing consortium, to jointly lease satellite communications transponder capacity. OPT had some spare capacity, in its arrangements for transponders for Internet, direct-to-home/office, and telephone traffic for its 275,000 people. Telecom Cook Islands arranged to lease some of that capacity for its 21,000 residents, at rates reported as a 50 per cent savings to customers, yet affording OPT a modest mark-up to its cost price. “The result was a huge saving for Telecom Cook Islands. And OPT also got revenue that they otherwise would not have had. It was a real life win-win situation,” according to Stuart Davies, chief executive of Telecom Cook Islands.¹⁷

.

2. VSAT Technology for Communication

Satellite communications can be facilitated on the terrestrial end by satellite Earth stations, very small aperture terminals (VSATs) or satellite phones.

Satellite Earth stations exist in most Pacific States, as noted in Appendix A. Satellite phones could serve all Pacific island economies, depending on the phone system – with no other infrastructure than a means of charging batteries in the phones, and the battery charger may be solar. With a retail cost of about US$1,000 for the phone, this may be the lowest-cost current modern telecommunications infrastructure option available to the Pacific – probably beating the lowest-cost VSAT by a small amount. Air/bandwidth is still relatively expensive, but a “bulk” agreement for airtime might reduce costs somewhat, and make possible universal coverage at least for urgent communications.

Very small aperture terminals have been made possible by improved technology. Unlike the several-metre-wide Earth stations that currently exist to provide the most reliable (but expensive and relatively low bandwidth) “fail-safe” gateway for many Pacific communities, sub-metre VSAT terminals are almost as reliable, and, when used for Ku- or Ka-band broadband satcom, they can deliver true broadband capabilities. The terminals themselves are much more economical than previously – on the order of US$1,000 per installation (antenna, electronics, installation, and short communication line). VSAT offers improved cost-effectiveness implementing high-quality, reliable communications to regions that are not well served by terrestrial networks.

VSAT-based solutions can service a wide range of population densities and be flexible enough to grow as user requirement change. Various solutions include (a) connectivity to subscriber lines to serve a scattered population (1 to 20 lines), (b) connectivity to wired or wireless/cordless local loop for clustered populations (20 to 300 lines), and (c) connectivity to macro-cellular networks for medium-density populations uniformly distributed (> 300 lines).

Advantages of VSAT networks include costs that are independent of distance and terrain,¹⁸ expansion costs that are predictable; a high degree of customizability; a high level of security; control and network management; rapid installation and relocation; and unattended and maintenance-free operation. As a result of such features, VSATs provide customers with significant gains in productivity efficiency, cost control and profitability. Satellites have advantages in networking broad, sparsely populated areas, as infrastructure is less a function of distance than when one must lay cable or build wireless towers.

VSAT services are available to parts of the Pacific Ocean area “from US$160 for 512 Kbps/128 Kbps download/upload”¹⁹ using New Skies and Asiasat C- and Ku-band satellites. By using dedicated satellites and adopting new commercial modes, the cost may be greatly reduced.

To illustrate the importance of VSAT in 21^st-century operations, every Wal-Mart (considered a global leader in efficiency in inventory management and resupply distribution) has a satellite terminal, which it uses to minimize overhead costs, while maximizing information and communications management, delivery and efficiency.

3. Satellite Capacity Estimation

Table 2-4 computes the estimated bit rate needed for rural satellite services. In addition to rural connectivity, capacity can utilize direct-to-home (DTH) broadcast and community-development-oriented applications such as telemedicine, tele-education, and disaster management. A total bandwidth allocation of 100 MHz is earmarked for such applications between all service areas. This will be apportioned among the various nations being covered.

The total bandwidth estimate is as follows:

(a) Bandwidth for rural connectivity (assuming 2 bits/Hz) is 463,230 MHz;

(b) Bandwidth requirement for services like DTH, etc. is 100MHz.

The above indicates an approximate total bandwidth requirement of 540 MHz.

Note that, though this computation is in terms of a satellite option, additional considerations should be made here. If satellite communications were to play a stronger role in urban areas – and as new, bandwidth-hungry products and services develop²⁰ – total bandwidth used by the Pacific would potentially grow to greater volumes, thus justifying longer-term plans for greater total bandwidth (satellite and cable combined).

Table 2-4. Estimated capacity requirements for rural connectivity

Type of Connectivity	A Population to be Connected	B Connection Ratio	C No. of Connections (Col. A x Col. B)	D Activity Ratio	E Actual Connectivity (Col. C x Col. D)	F Data Size (Kbps)	G Total Data Size (Kbps)
Wired (voice)	348,602	0.20	69,720	0.050	3,486	8	27,888
Mobile	2,371,874	0.20	474,375	0.020	9,487	16	151,800
Internet	1,555,774	1.00	1,555,774	0.020	31,115	4	124,462
Internet (download)	1,555,774	1.00	1,555,774	0.020	31,115	20	622,310

4. Satellite System Design Considerations

Frequency Band of Operation

For different applications using satellites, various frequency bands such as L, S, C, Ku or Ka are used. Sometimes the types of applications envisaged require a hybrid satellite using more than one frequency band. For the applications envisaged for Pacific islands, S- and L-band frequencies are not suitable from the viewpoints of coverage, equivalent isotropically radiated power (EIRP) or orbit-frequency coordination. C band is already well-deployed in the Pacific. However, most such systems covering the Pacific are low-powered, which require more expensive and difficult-to-operate user terminals than newer models might require. New broadband services built on Ku and Ka band are a major consideration. Ku band is considered more appropriate to conditions in the Pacific than Ka band, because of the latter’s susceptibility to signal attenuation or drop-out in heavy rains experienced in many tropical areas. Cost considerations for ground equipment also suggest that Ku band be considered the preferred frequency band option for the Pacific. A dedicated C-band satellite with shaped beam may also meet requirements for Pacific connectivity, and so should also be considered.

The selection of frequency bands can be from planned or unplanned bands. In the case of planned bands, the cost of ground equipment will be costly, whereas for unplanned bands, ground equipment is cheaper and technologically more advanced. In the case of unplanned bands, the frequency of operation should be decided by consideration of coordination constraints, in addition to satellite complexity, data rate requirements, user terminal size and cost of acquisition of ground equipment.

Satellite Orbit Location

The ideal orbital slots for satellites to provide coverage over Pacific islands are between 120°W and 170°E longitude. However, there are already around 10 satellites parked between 120°W and 150°W providing services over the Pacific and other regions, as shown in Figure 2-4.

Much of the remaining 20° of arc of geo-stationary orbit is vacant. Use of such slots can avoid coordination issues and can provide coverage over Pacific islands within a 10º elevation angle of the horizontal. If the slot between 120°W and 150°W were to be used for Pacific Island coverage, coordination issues would need to be resolved before finalizing any candidate orbit. A few orbital slots earmarked for the Pacific island countries are in the planned bands. But the ground equipment for planned bands would be costlier than equipment for unplanned bands. With such considerations, choice of orbital slot may be made judiciously considering all traditional selection factors.

Satellite Technology: Bent Pipe vs On-Board Processing

Satellites with on-board processing reduce the complexity and size of the hub but increase the complexity of the satellite system. Moreover, satellites with on-board processing cannot utilize advancements in modulation and coding techniques that occur after launch. Over the 15-year life span of a communications satellite, bent pipe transponders are a better choice for lower satellite costs, and for benefiting from possible technological advancements during the lifetime of the satellite.

Modulation and Access Schemes

Modulation and accessing schemes can be based on traditional or IP-based platforms. Selection should be made in the detailed network design phase. Use of advanced modulations like 8PSK turbo coding and of adaptive modulation and access technologies are assumed here in estimating the bandwidth, i.e. 2 bits/Hz. System cost and flexibility will depend on the technology selected.

Figure 2-4. Commercial communications satellites in geosynchronous orbit
Source: Boeing - Commercial Communications Satellites Geosynchronous Orbit 30 June 2006

5. Recommendations on Satellite

Four candidate configurations are presented here:

Option 1

A first option would be a satellite with 10 spot beams covering each Pacific Island nation and two transponders with a global beam encompassing all the nations. The details of the payload are as follows:

Satellite type: Three axes stabilized

Payload: 10 spot beams with 54 MHz transponders @ 51 dBW EIRP operating in Ku band

Two transponders of 54 MHz for two wide beams with 39 dBW EIRP in Ku band

Power amplifier: Linear TWTA

Power source: Sun tracking solar panels with advanced multi-junction cells, 100 per cent eclipse support using lithium iron batteries

Life: Orbit manoeuvre 12 years, design life 15 years

This proposed satellite would provide a total capacity of 540 MHz in spot beams and 108 MHz in wide beams, to satisfy the projected requirement derived above. It would have a total power of 2.5 KW and weigh 2-2.5 tonnes. The total estimated cost for the spacecraft, including launch and spacecraft insurance, would be US$110-120 million. Coverage is shown in Figure 2-5.

Figure 2-5. Sample 10-spot beam system covering the Pacific

A first satellite could be launched for the first phase of a satellite project, with a similar unit co-located later – or a larger satellite designed to replace the first satellite if its capacity is fully used or it becomes old enough for retirement.

Option 2

As a second option, a satellite named PIS-2 with 24 spot beams, with each beam dedicated for each Pacific nation, and two global beams encompassing all the nations is proposed. The details of the payload are as follows:

Satellite type: Three axes stabilized

Payload: 24 spot beams with 54 MHz Ku-band transponders with 54 dBW EIRP

Two transponders of 54 MHz for two wide beams with 39 dBW EIRP in Ku band

Power amplifier: Linear TWTA

Power source: Sun tracking solar panels with advanced multi-junction cells, 100 per cent eclipse support using lithium iron batteries

Life: Orbit manoeuvre 12 years, design life 15 years

Figure 2-6. Sample 24-spot beam system covering the Pacific, the “diaspora” and other neighbours, while also providing redundancy in case of failure of cables or other satellites

The system as configured also provides coverage to islanders who have emigrated to Australia, New Zealand and Hawaii, and provides services (including partial backup in case of another 26 December 2006-type cable outage) for countries around the Pacific Rim.

Option 3

The third option is to launch a more traditional C-band system, with coverage for all Pacific island economies, plus perhaps neighbouring areas of Australia, New Zealand and mainland Asia. The cost of such a package might be about US$205 million, including satellite, ground station for control and management, launch, and insurance. A sample coverage map is given in Figure 2-7.

Satellite type: Three axes stabilized

Payload: One shaped beam (2-4 transponders) covering 13 Pacific developing countries with EIRP > 40 dBW, for television and audio coverage; 3 spot beams (9-12 transponders, frequency–reused) covering 13 Pacific developing countries in three groups, with EIRP > 44 dBW for connectivity within and among these countries, and broadband based services; 1 shaped beam (5-12 transponders) to cover 13 Pacific developing countries, some South-East Asian countries, east Australia, New Zealand and part of south-east China, with EIRP > 40 dBW, for connection of the 13 Pacific developing countries with the outside world and for commercial services of these non-Pacific developing countries, so that the income may used for the operation of the satellite system

Power amplifier: Linear TWTA or solid amplifiers

Power source: Sun tracking solar panels with advanced multi-junction cells, 100 per cent eclipse support using lithium iron batteries

Life: Orbit manoeuvre 12 years, design life 15 years

Figure 2-7. Sample C-band shaped wide beam system covering developing countries of the Pacific, and part of Australia, New Zealand and South-East Asia
Source: Sino Satellite Communications Ltd. Presentation to Asia-Pacific Business Forum, 2006

Option 4

Another option would be to partner with one or more operators to lease partial capacity of a commercial satellite. This might be an existing satellite with spare capacity, or a future satellite whose design could be modified for maximum cost and performance efficiency for the situation of the Pacific.

This fourth option complements other options and might yield relatively quick benefits if pursued quickly. The Pacific Islands Telecommunications Association is currently studying such a course of action. Its understandings to date, and partnering potential, should be assets in the pursuit of this option.

6. Realization of the Selected Satellite

Options discussed here are summarized in Table 2-5.

Considering the population and traffic required for providing connectivity to all the countries in Pacific developing countries – plus providing financially rewarding services to neighbouring areas (including the nearby diaspora), a small or medium-sized satellite will suffice. It takes about 24-30 months to fabricate and operationalize such a satellite.

Because of the steady advances in technology and cost savings, it is suggested that establishment of infrastructure for ground systems begin one year before launch of the satellite.

Cost Estimation of the Satellite System, Including Major Application Supporting Systems

A refined cost estimate of a desired satellite would depend on the following factors:

• Application

• Coverage

• Data rate

• Satellite power

• Satellite construction

• Launch

• Insurance

• Operation and maintenance

Costs would also depend on other factors:

• Whether a single bigger satellite, or a number of medium-sized satellites, would be required;

• Redundancy option (e.g. potential for backup by another satellite or other means);

• Time duration for building and launching a satellite.

These points require detailed analysis regarding services to be provided, coverage area, and other factors.

However, unlike commercial telecom traffic in highly populated areas of the world, in this case it is to be expected that the break-even point in terms of revenue earned may not be reached for quite a long time, or in some cases may not happen at all. It is expected that an element of capital subsidy will be sought, plus some capacity to subsidize remote connections. Alternatively, service sharing, with beams added to serve partner areas, such as eastern Australia, New Zealand, Hawaii and eastern Asia. The overall costing in such cases may thus take into account intangible benefits such as improving the standard of living of the people, establishing a robust communication infrastructure, and other external or non-commercial benefits. The desired satellite systems would need to be built to achieve such objectives.

Cost Estimation of the Control Facilities and User Terminals and Facilities Individually, as well as Community-Based Facilities, for the Suggested Applications and Services

The cost of telemetry, tracking and control (TT&C) and satellite control facilities for options 1 and 2 will be US$9 million. For user terminals under options 1 and 2, the cost estimation takes into account the following factors:

• Population to be served;

• Service establishments that need connectivity (hospitals, schools, civil service offices, Internet cafés, security agencies and private establishments);

• Type of service (always-on broadband, dial-up etc.);

• Emergency services;

• Fixed or mobile or television broadcast;

• Communication terminals at personal level or at community level.

Table 2-5. Summary and comparison of the four satellite options

Option1	Estimated Cost in USD2	Orbital Slot	Service Supported	Advantages3, 4	Disadvantages
1. To build and own a Ku-band satellite with 11 spot beams	110-120 m satellite only	170ºE – 150ºW. Easy to find a slot.	10 spot beams for broadband-based services; 2 wide beams for television	High power and reused frequency may provide affordable bandwidth. May meet the bandwidth need in normal ICT development by 2015. May be able to provide additional economically useful services to Pacific Rim countries.	If applications develop quickly, second satellite may be needed before its retirement. Connection to outside relies on other means.
2. To build and own a Ku-band satellite with 26 spot beams	130-150 m satellite only	170ºE – 150ºW. Easy to find a slot.	24 spot beams for broadband-based services; 2 wide beams for television	High power and reused frequency may provide affordable bandwidth. More bandwidth than option 1. May be able to provide additional economically useful services to Pacific Rim countries.	Capacity might not be fully used. Connection to outside relies on other means.
3. To build and own a C-band satellite, with part of resources for commercial services	200 m., including satellite and main Earth station	145-180ºE. Not easy to find a slot, but easy to find partner and commercial opportunity.	2-4 wide beam for television; 9-12 spot beams for broadband-based services; 5-12 mid-wide-shaped beams for commercial service and interlink to outside	High power and reused frequency may provide affordable bandwidth. Commercial services may reduce subsidiary. Connection to outside world through the same satellite.	Covering only 13 Pacific economies, not all Pacific island countries and territories
4. To own or lease partial capacity of a possible commercial satellite	75-85 m., part of satellite for 12 years	~ 180ºW. Easy to find a slot, but not easy to find a partner.	1 wide beam	The partner provides all technical support. Connection to outside may be through the partner.	Lower power requires more expensive user terminals. Fixed satellite operator.

1. All satellites have a lifetime of 12-15 years; construction requires 2-3 years.

2. The estimated cost includes satellite manufacture, launch and insurance.

3. Options 1-3: May provide broadband services, accessible with similar low-cost user terminals.

4. Options 1-3: Operation of satellite may be contracted to any satellite operator under satellite footprint; it is not necessary to be in a Pacific island economy.

Consideration of these factors will lead to an estimate of how many ground terminals would be required in each service category, and cost estimates need to be arrived at appropriately. For the suggested application and services, recurring monthly subscription of the poorer communities should be kept to the minimum, financed by direct subsidies where possible and needed, in order to deliver affordable benefits of ICT to the rural poor people of the region. General access and user charges should, however, be set at commercially viable levels, which will depend on fund sources and costs. While estimating the cost, this point is to be kept in mind and all possible techniques should be adopted to reduce the cost, consistent with sustainable finance.

Pricing of service access is discussed in Section 4.III.B.