ELEKTRO INDONESIA                     Edisi ke Dua, April 1996

A Concept of Advanced Technology Satellite Systems for Asia Pacific : Rural Communications Applications

By : Arifin Nugroho

Common problem perpetuating in the Developing Countries, or in the rural areas, is how to develop communications infrastructure as rapidly as possible to cover the need of its populated regions at lowest costs possible. Apparently the satellite will be able to do the job, but the question is on which way. Many scenarios are revisited, starting from traditional VSAT, Mobile Satellite Systems (MSS), Baseband Processing, and finally High Power C-band satellite systems. Analysis shows that use of massive capacity technology is inevitable, although some simplifications are in order due to the cost contraint. Classical technology of bent-piping is still worthed so long as it is working in the lower bit rate environment combined with high satellite EIRP.

POTENTIAL PROBLEMS

Since "The Missing Link Report" of Maitland Independent Commission for World Wide Telecommunications Development established, the rate at which the gap is narrowing between high and low income countries in the telecommunications development has been slower than expected. More than two-thirds of all households worldwide still have no telephone. [MDIS]. The ITU noted in its World Telecommunications Development Report a growing gap is prevailing although some 200 million lines have been added since the publication of the Maitland report. "High Income economies" hold to just 15% population, have 71% of the world’s main lines. Teledensity increases in these countries increased from an average of 38/100 population in 1983 to 49/100 at the end of 1992. By contrast, the teledensity in the rest of the world increased from 2/100 to 3.5/100 on average. According to the ITU, some countries experienced, over the ten-year period, a decrease in teledensity due to fact that main line installation was outgrown by population.

Table 1 shows some 1994 statistics of countries in Asia Pacific.

Country         Population (million)      Direct Exch Lines(000)   Teledensity

Indonesia	     187.6	                 2,280	              1.5
India	             900.543                     8,037                1.1
Malaysia	     18.6	                 2,411	               13
China	           > 1000	                 23,000	              2.3
Thailand	     57.2                        2,542	              4.7
Vietnam              70.881                      260                 0.37
Cambodia	     9.633	                 5.9                 0.06
Myanmar              44.704                      3.8                 0.27
Laos PDR	     4.5	                 8.6	             0.19
Papua New Guinea     4.15                        39.8                0.96
Australia	     17.7	                 8,540	             48.2

TOTAL                2315	

Table 1 : Asia Pacific Statistics on Teledensity

Out of the above list of teledensity, there must be some "hidden" associated parameters with regards to the rural parts of each installation. Unfortunately, it is very difficult to acquire such data, and yet qualitatively it can be easily conjectured that those parts of regions are still suffering a very defficient provision of telecommunications facilities. If the imbalance is still prevailing in the telecommunications development in the developing as well as in the least developed countries, as compared with the high income countries, same remark should then apply for the development of telecommunications in rural areas compared with its concurrent urban areas in each developing and least developed countries. If the total rural population accounts for 70% of the total country polulation, then there will be a potential market for more than 1.6 billion rural lines in Asia Pacific.
The deficiency of rural telecommunications development is often associated with questions whether or not it is economic and profitable for extending services into rural areas. Many countries in Asia-Pacific region are also experiencing a growth of urbanisation. The average percentage of urbanisation of this region at 1969, 1990 and 2020 respectively are 18, 30 and 54%. The phenomena will further impede the rural development since the mass urbanisation will end up with higher population density in the urban area, which leads this area to be more lucrative than the rural one. Privatisation and competition scheme which affect most of the Asia Pacific countries will certainly further shift the old paradigm that telecommunications are parts of the government’s universal service obligation (USO) , toward more as business opportunity for the investor and operators. Since USO must be executed anyway, the rural development due diligence will finally be blamed to the governments. For developing countries this obligation will necessite certain percentage of GDP allocation for the telecommunication development.
The challenge would lie on what extend the committments are of those governments to spend some of their GDP for telecommunications, especially in rural areas. In view of less portion of the national budget allocated to the rural telecommunications development, many thoughts suggested cross subsidy should be encouraged between urban areas and rural areas. However, most of the impediments for this problem are due to the costs involved for installing the rural networks. GAS-7 of CCITT (ITU-T) has made extensive studies on various technology alternatives which can be possible solution for this problem. Those possibilities are : In particular very much dictated by the fact that copper technology are prohibitively expensive to serve those areas, in view of extensive geographical territory to be served with, in comparison to the radio-base system technology (wireless).
Wireless technology applicable for rural telecommunications spanning from wireless local loop, cellular, cordless and satellites. The paper will focus on the satellite system, especially for serving the remote area in which the rural population reside, and because the other wireless system will be too expensive to develop in such an area.
The satellite or transponder cost will not cheap. With the present technology, it takes up to $250 million to build and launch a modern telecommunications satellite today. Such high costs of acquiring satellite system will lead to to an expensive transponder costs, and finally to circuit cost that the rural people should pay, otherwise some subsidies have to be born. Many satellite had been launched, but this condition will not guarantee less transponder-leasing fees. The three primary cost escalators are costlier launch vehicles, more complex satellites and higher insurance rates. The high cost of having satellite in orbit will refrain the rural community of accessing to the satellite channel. In addition they still to spend more money for ground terminal equpments.
In view of helping the rural community to get about their needs for telecommunications access, the planner and the industry has a social responsibility to make the costs affordable to such potential users. This paper will address some methods for lowering the satellite costs.

NEEDS FOR AFFORDABLE SATELLITE SYSTEMS FOR RURAL COMMUNITY

Any satellite system can give adequate service to the rural community. What kind of service? Hudson suggested [TEL] that at least the following service shall be provided to rural areas as priority : electronic mail, computer conferencing, electronic meeting, data services, electronic transactions, training and distance educations. The suggestion would translate to the following : telephone, faximile, low-bit rate data communications, and video conference.
The "classical" solution of geosynchronous satellite operating at C and Ku band will make use of VSAT technology, delivering data, telephony, facsimile and perhaps some video signals related with teleconference and educations. Because of still high costs prevailing for acquiring such terminals, (VSAT could cost at higher than $20,000, plus tax, can go up to $30,000-40,000 costs) it is however unlikely if the end user will be ready to spend such a big amount of money. In addition, someone has to take care of the maintenance of the whole terminals plus carrying out the network management. In the VSAT technology domain, the so called Hub station will perform such a function. It basically consists of a big terminal plus the associated rf, baseband, controller, which represent quite a big budget to acquire them. Since this will involve a fairly big investment, such an investment will be born by a VSAT operator, and the end user can either : i) lease a VSAT terminal, in case the end user is a business company, or ii) to adhere to a wireless terminal of a WLL distribution networks, attached to the VSAT. A hybrid solution such as the latter will be limited to certain geographical and morphological size of the area to be served, otherwise the WLL would not be anylonger attractive. A recent study [Hughes-Telkom] shows this trend. In any event, it seems that current VSAT technology can be a solution to the rural community telecommunications in somewhat limited extent.
Since the WRC92, some space industries -- mostly those from the US -- have had taken initiatives for constellation of "Big" Low Earth Orbiting Satellites, called LEO which will give a opportunity of anyone within the satellite coverage to access to the satellite based network with a handheld type of terminal. This technology falls under the MSS (Mobile Satellite System) realm in view of the fact that most application to this technology would be for mobile user. Other market share which influence their business viability is the fixed wealth users, especially those who are at the rural areas. However, at the present technology, it would be very difficult to have a "mass production" circuits, due to the limitation of the spectrum bandwidth availability at L and S bands.
Compared with the terrestrial cellular systems, we conclude that the LEO satellite system would not compare with the cellular networks in providing the access capacity per km2. As a result, the "air time" of such LEO systems becoming so high which represent barrier entries to less income people to adhere to them. To some wealth spots, the advent of this technology can be a great help to them in accessing to the telecommunications networks. But from the massive aggregate of rural community, this alternative would hardly be an affordable system.
The discussion leads to a necessity of designing a satellite system which will carry a massive volume of (telephone)circuits, which tranlates into its capability of providing much bigger Erlang/Km2/Hz.

CAPACITY-COST ANALYSIS OF VARIOUS ADVANCED GEO SYSTEMS

One way of designing "massive" satellite would be, first of all, utilizing the higher spectrum, for example in the Ka band. A total of bandwidth of 2.5 Ghz are at least available on a sharing basis. Following the classical sharing spectrum of 500 Mhz at C band per satellite, than one can proceed the use of any 500 Mhz out of the available Ka spectrum. The design can then follow the ACTS experimental satellite [Gedney]. Among the followers of this method in the US are Spaceway (Hughes), VoiceSpan (AT&T), Cyberstar (Loral), and Millenieum (Motorola). These are GEO based satellites global systems, whilst the Teledesic is proposing a quite revolutionary system consisting of 840 Low Earth Orbiting Satellites also working at Ka band. In fact the last WRC-95 has granted to Teledesic the exclusive band at Ka spectrum. The advent of those advanced satellite system seemed to be inspired by experimental systems of the NASA called ACTS. Many technological heritage are drawn up from the said experiment.
In any case the answer seems will be a Ka band satellite. The rationale behind this direction would be : Ka band will give opportunity for multi-beam satellite at very high gain (around 1-3 deg), which enable the smaller earth station to access to it, thus it will enable less cost earth stations, a condition which the rural areas are expecting; besides, it will also accomodate more frequency-reuse to certain geographical areas, thus more channels can be obtained. In fact, following the concept of the ACTS’s heritage, the relevant concepts applicable and advantageous to the rural users are:

1. the satellite shall be totally digital, which payloads include sophisticated all-digital processors that demodulate and decode the uplinked pecketized traffic, and then route the traffic to the appropriate downlink or crosslink beam according to the packet header address for recoding, remodulating and onward transmission.
2. the satellite would provide a lower rate data transmission service ( at 16 kb/s) serving extremely small apperture terminals (0.6m in diameter).
3. Fixed receive spot beams, in conjunction with Baseband Processor (BBP) technology will provide up to 6000 simultaneous users per beam, or 3000 duplex links. SpacewayTM system, for example, does plan for 144.000 duplex links for up to 48 beams per satellite.

The last figure will translate to a number of 2,000,000 users can be serviced by the satellite. Comparing the circuit capacity rendered by the 80’s bent pipe C-band satellite technology (for the class of HS 376, etc), this figure represents some 2 to 5 times higher more capacity.
The ACTS’legacy technology has also shown that the use of Ka-band will augment the communications capacity per unit on-orbit weight to approximately 3 times conventional system [Gedney].
It is however felt that for immediate applications of such heritage, some adaptation seems necessarily be considered, in view of complexities, markets, and economy as applied to the benefit rural community, and to the developing countries.
This paper will suggest some design of Ka band satellite using some principle given by the ACTS technology, but adapted to the rural requirements.

RURAL SATELLITE CONCEPTS

The logical solution for the problem would be a design of Very Large Mesh VSAT Fixed Networks, applicable for rural areas. First of all, one should determine the most probable throughput that the terminal will be transmitting. We choose the throughput rate of 64 kb/s to be adequately representing voice data or video conferencing signals, in view of current technology available. Tradeoff design would be presented as follows :

Multi Access Design

To significantly reduce the costs in providing a very large mesh VSAT networks for rural areas, it is necessary to keep terminal cost minimum. To do so, terminal diameter should be kept very small. First, consequently, the downlink was assumed to be TDMA so as to operate the satellite transmitter may be operated at maximum efficiency and provide maximum RF power for a given available DC power , hence this will facilitate the less cost receiving terminal. Cost optimization, however, is still to be pursued in the uplink portion of the network, in particular with regards to its multiple access techniques. Criteria having impacts on the system cost and hence user cost would be capacity, timing and frequency accuracy, transmit power and ease of operation. Current study has shown that [MITCHEL] that there are only four candidates of multiple access techniques qualified to be a better solution for optimizing the system costs. These are : a) FDMA, b) MF/TDMA, c) MF/CDMA and d) MF/CPDMA. FDMA is the simplest system with low cost ground terminal, since it does not require extremely accurate timing at terminals. However this alternative merits highest capacity compared with the other candidates. In fact it can be shown that for a single uplink beam, the relative capacity, measured bu K factor, which is equal to the number of users sharing the same bandwidth, would be N, asN, N/1.5, and N/1.5 for FDMA, synchronous MF/TDMA, MF/CDMA and MF/CPDMA respectively. N is the number of spreading code symbols per channel information symbol. MF/TDMA, MF/CDMA as well as MF/CPDMA require only moderate timing technology compared with the full-band TDMA, full-band CDMA and full-band CPDMA hence less cost.

The situation would be different if we are to design a multi-beam uplink technology, instead of single beam. The capacity comparison among the above four candidates would be given by comparing the frequency reuse-factor FR resulting from each candidate analysis. It can be shown that the analysis will result to FR figures of FRC, FRC , 0.48 FRC , 0.54FRC, for MF/CDMA, MF/CPDMA, MF/TDMA and FDMA respectively.
Thee above results depend on certain assumption regarding the number of uplink beams ( 8 in this case). If we increase the number of uplink beams, the results are still true but the difference between various candidate will tend to diminish.
There are however operational burdens with regards to the need of power control function for both CDMA and CPDMA. In view of implementing least costs of ground terminal, the logical candidate for the up-link multibeam environment would be FDMA, MF/TDMA, MF/CDMA and MF/CPDMA.

On-board switching or not

In the "analog" satellite realm, the purpose of satelite transponder is merely to amplify and broadcast any signal received by satellites. With the present of VSAT technology, this bent-pipe type of satellite has extended its function such as networking, along with the software and other intelligent implemented in the ground. In the totally digital era of satellites, the transponder will expand its capability in just to amplify and broadcast signals, it will also process, regenerate, route and switch the signals. These methods, in one way or another are seeking optimum solution of networking to the benefit of multiple type of users; it was always a connotation that the users in this case are "modern" users with various bit-rates signals requirements. The ranges of bit-rate would, for example, encompass the 16 kb/s to say above 10 Mb/s. Such variability in types of signals will fit to the ATM technology, a broadband networks which switch various type of signals by first converting them into a cell-type of packets at constant length. This requirement would lead unevitably to a switching of higher data rate throughput, and to a memory hungry situation. In addition, the flexibility in bandwidth assignment carries also a heavy processing penalty [Cangiane]. This problem should not necessarilily appear if we are to limit the range of signals, say from 16 kb/s to 2 Mb/s. Publication of Cangiane et.al [Cangiane] propose a Multi-channel Demultiplexer/Demodulator suitable for this simple range of application. This lower bit rate uplink throughput will lead to less cost terminals. In other word, implementation of the same technology to only certain subset of signals for rural communications application, will invite some economic justification. A burst of ATM signals at higher bit rate will necessite higher bandwidth, and physical law will dictate higher "system value" of the link. In other words, we must have higher terminals’G/T, higher Terminals’ EIRP and higher costs.
The Ka operation is subject to rain attenuation, especially for satellite links at lower latitutudes. Countermeasures [YIM] may be possible to surmount the effects, among others are i) up-link power control, ii) signal redundancy, iii) reduce data rate, iv) adaptive modulation and v) adative TDMA. The difference in availabilty requirements will translate to its associated fading margin; in a lower latitudes region, 99.5% requires 63 dB, while 98% requires 11.8 dB. It is then unavoidable that the power control or other measures should be utilized to compensate such a fading, which lead to further increase the terminal cost. A way to alleviate the problem is to loosen the availability requirement, from, say, 99.5% to 98% or less.

The rain adverse effects, on the other hand, will virtually not exist in lower spectrum as in the C or X bands. Furthermore, a simple bent-pipe transponder could also give a networking solution to VSAT system, and after appropriate choice of spectrum and improving the satellite EIRP, the system can be exploited further in giving USAT networking capability as well. The Table 2 is showing typical down link budget for a USAT terminals application for such a system.

Table 2 : Satellite Down-link for mobile type quality links at Extended C-band using bent-pipe payload

	                      C-band transp      Ext C-band transp

Transponder EIRP, dBW	        39                   45
Output Back-off, dB	        -3                   -3
Number of carriers,dB	        -39.8 (9600)         -39.8 (9600)
Voice Activation, dB	        4                    4
Space Loss, dB	                -196                 -196
Antenna gain, dB	        36(2m)               30(1m)
Receive carrier, dBW	        -159.8               -159.8
K (Boltzman Const), dBW/K.Hz	-228.6               -228.6
B (bandwidth), dBHz	        36                   36
T (temp), dBK	                21                   21
N (noise), dBW	                -171.6               -171.6
(C/N)ul, dB	                11.8                 11.8
(C/N)IM, dB	                16                   16
(C/N)dl, dB	                11.8                 11.8
Eb/No, dB	                10.4                 10.4

 

CONCLUSION

The developing countries, the rural or remote areas, or the hard to visit regions have a common problems of telecommunications infrastructure. Such a chronicle can be terminated among others by implementing a suitable satellite technology. One way of doing it would be an advanved Ka band satellite with Base Band Processing (demod, mux, atm-switch, mod, etc), along with suitable choice of multiple access techniques. The bottom line should be how to reduce the terminal costs, which is always proportional to power and bandwidth (throughput), and antenna diameter requirements. Perhaps some relaxation to the availability is conceivable for rural applications. Other solution would be a bent pipe simple satellite systems at C band with higher satellite EIRP.

Acknowledgement:

I should convey my deepest gratitude to the CEO of PT TELKOM, Mr Setyanto P Santosa, for initiating and encouraging me to carry out this work and allowing me to share TELKOM’s perspective. I would also indebt many thanks to Dean Olmstead, Jeff Pugay, Victor Barajas, Dr. Klaus Johannsen, all of them are from Hughes for such fruitfull discussions during preparation of this work. Finally I should owe Dr R Kularajah from Lockheed Martin Aerospace and RA Gafar from PT CSM for sparing their precious time in reviewing this paper.

REFERENCES

[Cangiane] : P. Cangiane, G.Caso, H.Courtois, R.Rey, TRW Space & Electronics Group, Multi-channel Demultiplexeer/Demodulator, AIAA-94-1030-CP
[Gedney] : R.K.Schertler, R.T.Gedney, M.J.Zernic, ACTS’ legacy : The Ka-band explosion, Aerospace America, February 1996
[Hughes-TELKOM]: Unpublished Internal Exchanged Documents
[Mitchel]: W.Carl Mitchell and Kent M. Price, Space System LORAL, Capacity Comparison of Uplink Multiple Access Techniques for Very Large Mesh VSAT Networks, AIAA-94-0924-CP
[MDIS]: MDIS Publications, September 1995 : Rural Telecommunications: The Challenges, Technologies and Players.
[YIM] : W.H.Yim, F.P. Coackley, Centre for Satellite Engineering Research, University of Surrey, Guildford, UK, Multicarrier Demodulator for On-board Processing in Ka-band, AIAA-94-1029-CP
[Tel] : Dr. Hellen Hudson : Telecommunications - A Vital Element in Asia’s Economic Development , Telecommunications Challenge in Asia : a Roundtable Dialogue, Hongkong June 1994, Sponsored by BT, Northern Telecom, Salomon Brothers.