Volume 10, Number 2, July 2007
Fifteen Constraints On The Capability Of High-Capacity Mobile Military Networked Systems
- 1 Carlo Kopp, Clayton School of Information Technology, Monash University, Clayton, 3800, Australia.
Abstract
The network centric warfare (NCW) model represents one of the defining trends in information age military technique. Its aim is to improve situational awareness and ‘accelerate the observation-orientation-decision-action (OODA) loop’. While much literature exists which extols the virtues of NCW, the problem of what constraints exist on the capabilities of such systems has been explored much less frequently. This paper identifies no less than fifteen constraints on the capability of networked military systems, implemented with tactical datalinks, and explores their respective causes and implications.
Introduction
Network centric warfare (NCW) and the less ambitious model of network enabled operations (NEO), represent defining trends in the development of military systems and capabilities, for this decade. The aim of both is to reduce the time required to perform engagements, or ‘accelerate the observation-orientation-decision-action (OODA) loop’. Less exact statements of aim include the notion of ‘dispelling the fog of war’ by using the network to rapidly gather and distribute information [1,3,5,17].
A key consideration for planners and architects of such systems is that of what constraints exist on the achievable capabilities of such a system, which is typically implemented using highly mobile tactical datalink technology. Such constraints do exist and the aim of this study is to identify these constraints and establish how they constrain the capability or utility of the network.
In the broadest sense, these constraints can be divided into two categories:
- Hard limits imposed by the physics of radio signal propagation and the mathematical properties of networked systems.
- Impairments resulting from hostile actions, and human constraints on the system.
The potential for any specific network to be implemented and used successfully will be determined by these constraints, which become increasingly relevant as the intended network approaches the hard limits imposed by physics and the mathematical properties of networks.
No less than fifteen specific constraints or limits have been identified. We will analyse each in turn, and where applicable, the implications of not considering their impact. This is especially important in terms of identifying differences in constraints which apply to cabled non-military networks. It is implicitly assumed that such networks are implemented using wireless radiofrequency digital datalinks.
The power-aperture constraint
The channel capacity of any wireless radio link is limited by the radio frequency power output of the transmitter, the bandwidth available within the radio spectrum, and the size of the antennas used, for any given distance between stations, in free space. This is the Friis power-aperture or path length loss equation, which constrains all radio-frequency communications [22].
Most established mobile military networking equipment, such as JTIDS/MIDS/Link-16 terminals, employs omni-directional low-gain antennas, and transmitters with power ratings of the order of ten of watts to kilowatts. For instance, a time division multiplexed JTIDS/MIDS/Link-16 network provides capacities between tens of kilobits/sec to Megabits/sec, subject to configuration. Characteristically, most such systems are implemented with omni-directional antennas to facilitate high mobility, without the cost impediments of a steerable antenna main lobe. Subject to constraints in bandwidth and modulation, achievable capacity is thus bounded by Friis [20].
In the broadest theoretical sense, unlimited bandwidth and power, and antenna aperture size permit unlimited growth in capacity. Pragmatic constraints will inevitably set hard limits on capacity. A good comparison is that of establishing what kind of radio-frequency link would be capable of competing in channel capacity with an optical-fibre link of Gigabit/sec class capacity. An extensive study by Kopp [12,15], in part since empirically validated by US industry [11], exploring high speed airborne ad hoc networks established that to achieve Gigabit/sec class capacity in an X-band or Ku-band radio-frequency network, over distances of tens to hundreds of kilometres, requires the use of power levels, antenna aperture sizes and receiver performance comparable to that of a contemporary active electronically steered antenna (AESA) radar on a fighter aircraft.
To provide an aircraft, warship or vehicle with 360° very high speed network link coverage with such performance thus requires three to four such antenna/transceiver systems. The cost of such an installation would be dominated by the AESAs employed, which at current radar costs would result in an installation cost of millions of dollars per platform. For many military platforms such installations are not feasible for reasons of size and weight, regardless of cost.
This problem is often described in terms of the ‘range/capacity/mobility’ trade-off, where the need for mobility and reasonable transceiver/antenna costs conflicts with the demand for power-aperture performance to affect high channel capacities [27].
The often-asserted expectation that Moore’s Law driven miniaturisation seen in computer hardware will apply to high-speed radio-frequency computer network hardware is false, as it fails to account for the physics of antenna aperture design, constrained by the Friis power-aperture equation. Put simply, Moore’s Law does not apply to antenna apertures [13,14].
The propagation physics constraint
The channel capacity of any wireless radio link with given transmitter power, receiver performance, bandwidth and distance between stations will be limited by the impairments experienced in propagating the radio signal through the atmosphere, regardless of terrain and antenna elevation constraints.
In the microwave bands of interest and relevance to military networking, rain, cloud, fog, haze, atmospheric gas molecules, reflections from terrain, solar noise, and unintended interference by commercial or military equipment can all impair transmission [2,7,10,17,23].
In the upper microwave bands these impairments can be severe enough to make connections unusable. A detailed treatment is provided in [16]. The atmospheric propagation conditions best suited to high-speed wireless networking may be absent due to weather conditions for much of the year in equatorial or tropical geography, which for the foreseeable future will remain a critical area for military operations.
The notion that the propagation environment will permit round-the-clock, all-seasons, high-speed, long-range network connectivity is predicated on the assumption that radio propagation conditions will always be ideal, and physics of radio propagation impairment mostly do not apply.
Communications links at shorter wavelengths, whether millimetric or infrared, suffer even more severe impairments than microwave links suffer because the scattering and absorption effects are more pronounced.
In summary, the constraints imposed by propagation further constrain achievable network capacity, in addition to the constraints imposed by Friis path-length loss.
The radio spectrum constraint
The available radio spectrum for military networking is limited by severe and worsening global congestion of the radio spectrum. This congestion has arisen as a result of rapid and sustained growth in satellite communications, broadcast usage, mobile telephony and networking, and dedicated commercial radio-frequency link usage.
Moreover, the radio-frequency bands which are least impaired by natural radio propagation impairments are by far the most heavily congested. This is a result of competitive pressures, as service providers seek to minimize the cost of hardware installations. The use of bands which suffer impairment, such as the 28 GHz band, incurs additional costs due to link power margins and spatial diversity [16].
A good example of this problem taking effect is the replacement of the APQ-181 Ku-band radar on the B-2A stealth bomber with an X-band design, as the radio spectrum used by this radar was reallocated for commercial use. Existing US datalink equipment for the AGM-130 and EGBU-15 smart weapons will also have to be replaced due to the loss of L-band spectrum used by this equipment. Many other examples exist [26].
The reality of this century will be increasing pressure upon the radio spectrum, with declining availability of spectrum which is well suited for military networking applications.
Limited availability of spectrum thus is an important constraint, especially for jam-resistant wideband spread-spectrum links.
The footprint constraint
High-speed radio networking requires that line-of-sight exists between the platforms carrying networking equipment, because microwave links (or laser links) cannot propagate through the earth or water. The curvature of the Earth therefore sets hard limits on what range is achievable. While atmospheric refraction can slightly ameliorate this problem, any gains so achieved are likely to be overwhelmed by low-altitude propagation effects such as fading and interference [15,16].
Whether the platform is an aircraft, unmanned aerial vehicle UAV), warship, land vehicle, fixed ground station, or satellite, the two stations at either end of the link must be connected by direct line-of-sight. While multi-hop networking schemes, such as JTIDS/MIDS relaying or ad hoc networking, permit a larger geographical footprint across the network, every routing or relay node introduces an additional point of failure in the network, and incurs a penalty in propagation delay and quality of service (QoS).
For a mobile transceiver on a surface based platform, antenna elevation and surrounding terrain elevation constrain the direct line of sight. For a low-flying aircraft or helicopter, platform altitude and surrounding terrain elevation constrain the direct line of sight. For a man-portable system, surrounding terrain elevation and obstacle height becomes a critical constraint. Achievable line-of-sight distances in such circumstances range between tens of metres, in extremis the ‘urban canyon’ scenario, and tens of nautical miles.
For an aircraft at the typical cruise altitude (tropopause) and surface based station, this distance is at best of the order of 200 nautical miles. High-flying UAVs or ‘pseudolites’ can do slightly better, and satellites much better, but both suffer increasingly from the preceding three constraints, and become increasingly expensive as achievable link speed is increased.
In practical terms, the footprint constraint forces the use of expensive network relays at high altitudes, or in orbit, or it requires a constrained density of network relays within the area of interest. In effect it is a topological manifestation of the ‘range / capacity / mobility’ trade-off.
The channel capacity growth constraint
The problem of channel capacity growth is inherent in all networks, whether civilian or military. Demand for capacity always exceeds installed capacity. In civilian networks, especially the Internet, this demand has been the driving force in the large-scale deployment of optical-fibre and copper-cable networks, and has resulted in the ‘bandwidth law’, which states that ‘available bandwidth doubles every two years’ [24].
Military networks do not obey the ‘bandwidth law’. This is because they are primarily wireless radio networks, where communications link capacity is strongly constrained by the Friis power aperture constraint, the footprint constraint, and the physics constraints of radio propagation. The ‘range/capacity/mobility’ trade-off imposes economic and physical size constraints which are quite different from those observed in civilian networks, where the latter can exploit the fixed cabled infrastructure. A multi-hop military wireless network incurs more queuing delays, and more stringent demands on rejection of interfering signals, particularly hostile jammers.
Therefore the expectation that military networks can achieve the same growth rate in capacity over time as seen in civilian networks is predicated on the validity of the ‘bandwidth law’ in wireless networks, which is a false assumption. Put simply, the ‘bandwidth law’ as stated does not apply to military wireless radio frequency networks.
The network congestion constraint
Computer networks are susceptible to congestion problems, which usually arise when the load on the network exceeds available capacity. Congestion impairs Quality of Service through dropouts, breakdown in services which are sensitive to time delays in transmission or even ‘congestion collapse’ whereby the network is unable to carry any useful traffic. Services which are especially sensitive to congestion related effects include live video and voice, or real time targeting updates, all of interest in NCW environments [4,8,9,25].
For instance a targeting datalink used to guide smart bombs may not require much bandwidth, but is QoS sensitive, as link delay must be constant and stable to permit multiple updates per second per weapon.
In a time critical and survival sensitive environment such as military operations, congestion problems can be fatal, yet the conditions in which congestion is most likely to arise are at times of peak activity or rapid situational change, when the demand for new information is greatest.
Moreover, congestion most frequently arises in ‘bottleneck’ stations or links in any network, which as a result of physical location concentrate traffic from a large number of other stations – in graph theoretical terms these are cut vertices or bridges, respectively. It is almost impossible to guarantee that in a mobile or mobile ad hoc wireless radio network no stations will ever become ‘bottlenecks’ to network traffic throughput, as the instantaneous network topology is driven by factors other than optimal radio-frequency coverage [16].
The preferred mechanism used for QoS management with real time traffic loads is bandwidth reservation, which is inherently problematic where the available capacity along any route dynamically changes at a high rate. At any given time, achievable capacity performance along a given routing path will be limited by the performance of the slowest links or routing nodes [4,8,9].
The sensor bandwidth mismatch constraint
Modern sensors gather data at rates which are mostly significantly higher than the rates at which data can be transferred, in real time, across wireless radio datalinks of contemporary technology.
Good contemporary examples of network capacity demand are intelligence, surveillance reconnaissance sensor platforms which can demand in excess of 50 Megabits/sec of link capacity per platform, if all sensors are active. A single live video link demands around 2 Megabits/sec of link capacity per platform. A single reconnaissance image of high quality may be 50 Megabytes or greater in size. Frequently such data formats do not yield good lossless compression ratios, and usually demand high QoS.
A reconnaissance camera or imaging radar will capture multiple images of 50 Megabytes or greater size per second, bounded by sensor integration and dwell times, yet the best contemporary datalinks provide a fraction of the bandwidth required to effect a real-time transfer. The growth in sensor data capture capability is set to continue to exceed available network capacity, as imaging chip and radar technology evolves.
In many important applications, especially those where human interpretation of sensor output is required, on-platform processing of the data is not feasible, unless the platform is large enough to carry specialized personnel. The expectation that sensor output processing can be used generally to reduce capacity demand is not realistic.
For the foreseeable future, the capacity of sensors to capture data will exceed by a large margin the capacity of networks to transfer that data quickly, let alone at real time rates [13].
The human bottleneck constraint
Shannon’s models show that information is not the same thing as raw data. Raw data must be understood and interpreted to produce information, whether this understanding and interpretation is by man or machine is an issue of specific implementation for a specific system.
The reality for the foreseeable future is that human interpretation will mostly be required to extract information from raw data, or validate machine interpretations of raw data.
In a network centric system this reality imposes hard constraints on the speed at which information can be collected and distributed via the network, as humans must interpret and validate this information at one or more points along the path from source to consumer.
Humans are slow and can be prone to error. Until artificial intelligence (AI) technology can replicate the cognitive reasoning of human beings, this ‘wetware bottleneck’ will constrain what can be achieved by automation in a network centric system, or ‘system of systems’.
The AI debate remains unresolved. Many theorists, over recent decades, have argued that such a breakthrough in technology was imminent, but it has yet to occur. There is no evidence to prove that such a breakthrough will necessarily occur over the next two decades [15].
The information integrity constraint
Computers and computer networks are the most efficient means devised to date for storing and distributing en-masse large volumes of data. The integrity or validity of this data is another matter, since a networked system can and will distribute erroneous data just as efficiently as valid data.
Whether errors arise from limitations or faults in sensors, incorrect human interpretation, or simple typographical errors, defective data can be quickly distributed across a warfighting system. Good examples are typographical errors in GPS coordinates used to program smart bombs in recent conflicts.
The result is that a single individual could produce large-scale damage effects quickly, if erroneous data is widely and quickly distributed.
The result of this effect is that there will always be a finite error rate in the information flow within a networked military system. The frequency of and severity of impact arising from this effect will be specific to the network in question, the sensors and skills of the human element within the network [15].
The rules of engagement constraint
The decision to fire a weapon against a target is mostly one which remains in the hands of human beings. No military force globally has yet adopted a policy which permits fully autonomous detection, acquisition, tracking and weapon delivery against a target by a machine.
The reason is the risk of a machine inadvertently killing innocent bystanders. A similar problem can arise in a highly networked system, where an operator might elect to launch a weapon on the basis of data gathered by multiple sensors, with the operator and sensors being relatively remote to the target. Unless a high confidence level can be achieved in identifying a target as conclusively hostile, there is a genuine risk that limitations in sensors, and human interpretation of sensors, will result in friendly fire or civilian casualties.
Limitations and faults in identification equipment can contribute to such situations, as well as human errors on the part of the operators prosecuting a potential target. As a result, rules of engagement imposed upon a military force in times of crisis or war may severely restrict the extent to which the advantages offered by a network can be exploited in combat.
Numerous case studies of this problem exist, including the USS Vincennes incident, the Kosovo tractor bombing incident, and the more recent bombing of Canadian troops in Afghanistan. It is important to observe that in every incident multiple human errors usually compounded by sensor limitations contributed to tragedy.
Rules of engagement can be expected to remain a constraint on the usable capabilities of networked warfighting systems [15].
The fleet economics constraint
Networking equipment is not expensive compared to other military equipment, examples including radar equipment, optical imaging equipment or gas turbine engines. A capable state of the art JTIDS/MIDS network terminal, such as the MIDS-LVT, including the cost of software and amortised integration on a combat aircraft costs of the order of 15–25% of the cost of a radar or targeting pod.
However, when equipping a fleet of platforms, this can add up to a significant cost burden. To this cost must be added the cost of through life hardware and software upgrades, and replacement equipment with every generation of networking equipment, which might need to be replaced two or three times over the life of a new platform.
Finite budgetary resources and fleet sizes thus represent a constraint on the rate at which any new networking technology, or upgrade, can be retrofitted to a fleet of platforms.
The compatibility problem
Networking hardware and software is inherently complex, and often evolves relatively quickly. A problem observed in both civilian and military networks is that of incomplete or partial compatibility, where ostensibly compatible equipment cannot fully interoperate as some features or options in the network definition are not supported by all items of equipment in use.
This problem becomes more pronounced if the fleet of platforms is equipped with networking equipment from different vendors, or of different generations or configurations. The result of such compatibility limitations is that the network can only reliably provide a subset of its total functionality, to the detriment of its users.
The enemy jamming constraint
Wireless radio networks are susceptible to jamming by an opponent, to a greater or lesser extent, no matter how good the jam resistance measures designed into the network [6,28].
The idea of an ‘unjammable’ communications link has been popular since the 1940s, but cannot be implemented due to physics of communications links. Design measures used to reduce susceptibility to jamming include spread spectrum modulations (frequency agility and power/bandwidth trades), controlled reception pattern antennas, and robust forward error correction codes [18,21,25].
A well designed network will cope with jamming by reducing link capacity, effectively slowing itself down, to resist jamming. A not so well designed network will collapse or drop out. The effectiveness of any jamming technique will depend on how closely it is designed to exploit weaknesses in a network, and upon how much jamming power is used. In general, it is unsafe to assume a technologically oriented opponent will be unable to jam a network [21].
Network capacity and reliability will thus be constrained by the effects of hostile jamming.
The enemy attack on network and sensor nodes constraints
Any opponent with reasonable competencies in electronic warfare will be able to identify stations in an opponent’s network by their radio frequency emissions. Platforms participating in a network and carrying traffic on behalf of other platforms will by necessity transmit more frequently than other stations, and stations which are bottlenecks in a network which is heavily loaded or nearing congestion, will also transmit more frequently.
Targeting attacks on such platforms, for instance by using long range missiles, can cripple a network, or force a shutdown.
Network availability and reliability will thus be constrained by the effects of hostile attacks on network stations, or the platforms carrying them [21].
Platforms equipped with sensors are the ‘eyes and ears’ of a networking warfighting system. Remove them from operation, and the network become blind and deaf, as what remains is the ability to move data around, but no data of value is provided.
Most networked systems in use today rely to a large extent on large long range airborne sensor platforms, such as the E-3 AWACS, E-8 JSTARS or RC-135V/W Rivet Joint. Such sensor platforms typically emit datalink signals, and in the instance of active sensors, high power microwave emissions. Therefore they can be easily detected and tracked [18].
Destroying such a platform with a long-range missile, or forcing the platform’s withdrawal, cripples the networked system. Examples of such weapons include the Russian Vympel R-37 (AA-13 Arrow) and Novator R-172 (AAM-L) missiles [18,29,30].
Network availability and reliability will thus be constrained by the effects of hostile attacks on network sensors, or the platforms carrying them.
The symmetrical opponent constraint
The notion that a networked military force will only ever confront opponents not equipped with networking technology is both foolish and shortsighted, yet remains a common idea. It effectively assumes that networked forces will only ever be used to deal with opponents of the technological sophistication of the Taliban. The reality of the digital age is that industrialised nations have been rapidly acquiring this technology [15,18].
It is no less important that the gap in military effect between an opponent using advanced networking and an opponent using less advanced networking is relatively small, compared to the gap between any two opponents, one of whom has networking and one of whom does not.
The availability of Russian networking and datalink equipment in the global market is public knowledge and has been so for some years.
Conclusions
These fifteen problem areas are well documented and many have been the subject of publications in the ongoing debate on military networking.
Each will affect to a greater or lesser extent all military networks, no matter how basic or advanced. These problems define the limitations of military networks, and are unavoidable consequences of the physics, mathematics and technology involved.
What more specific conclusions can be drawn, to aid planners, architects and theorists?
Clearly the constraints imposed by spectral congestion and power-aperture/propagation constraints, manifested in the ‘range/capacity/mobility’ trade-off and the resulting invalidation of the bandwidth law, are of critical importance. As networks proliferate and spectral congestion worsens over time, achievable capacity for military networks will not exhibit strong growth. There will be opportunities to improve modulation techniques, and exploit spread spectrum code sharing, but all incur losses in resistance to jamming, and resilience to natural propagation impairments. While the extant technology base has yet to hit hard bounds, these bounds exist and will become a consideration over coming decades.
Constraints arising from congestion are inherently specific to a network design, its protocols and its immediate topology at any point in time. Ad hoc networking techniques will provide some opportunities for reducing congestion, but all will be predicated on the absence of frequent cut vertices or bridges in the network topology, and relative stability in instantaneous network topology. The latter are inherently difficult to guarantee.
Constraints imposed on integrity, capacity and rules of engagement constraints are inherent by-products of the human element in such systems, where processing and data gathering limitations of equipment are not dominant.
Constraints on capability or utility arising from hostile action, as well as the use of networking by opponents, reflect the evolutionary nature of warfare and must be considered from the outset in the design of any such system. A sophisticated ‘peer competitor’ category opponent will have many opportunities to impair or cripple a network, and many opportunities to exploit their own networks.
In summary, careful consideration of the fifteen constraints on capability in networked military systems is essential if robust systems are to be implemented.
References
[1] D.S. Alberts, J.J. Garstka, and F.P. Stein. Network Centric Warfare: Developing and Leveraging Information Superiority. Washington, DC: National Defense University Press, 1999. See also 2d ed. Rev., 2001.
[2] L.V. Blake, Radar Range-Performance Analysis, Lexington Books, 1980, Ma.; Blake L.V. ‘Prediction of Radar Range’. in Skolnik M.I. (Ed) Radar Handbook, Second Edition, McGraw Hill, 1991.
[3] J.R Boyd, Patterns of Conflict, Briefing Notes, December, 1986—unpublished, Defense and the National Interest, URL: http://www.d-n-i.net/second_level/boyd_military.htm [Date accessed: 2/09/04].
[4] H.J. Chao and X. Guo, Quality of Service Control in High-Speed Networks, John Wiley & Sons, 2002.
[5] DoD, Network Centric Warfare, Department of Defense Report to Congress, 27 July, 2001, US Department of Defense, Washington DC.
[6] R.E. Fitts, Strategy of Electromagnetic Conflict, Peninsula Publishing, 1978.
[7] W.L. Flock, Propagation Effects on Satellite Systems at Frequencies Below 10 GHz, A Handbook for Satellite Systems Design, NASA Reference Publication 1108 (02), 2nd Edition, 1987, NASA Scientific and Technical Information Division.
[8] N. Giroux and S. Ganti, Quality of Service in ATM Networks, Prentice Hall, 1998.
[9] A. Grenville, Quality of Service in IP Networks, MacMillan Technical Pub, 2000.
[10] L.J. Ippolito, Propagation Effects Handbook for Satellite Systems Design, A Summary of Propagation Impairments on 10 to 100 GHz Satellite Links with Techniques for System Design. NASA Reference Publication 1082 (04), 4th Edition, 1989, NASA Scientific and Technical Information Division;
[11] J. Katzman, “Elec Tricks: Turning AESA Radars Into Broadband Comlinks”, Defense Industry Daily, URL http://www.defenseindustrydaily.com/2005/12/elec-tricks-turning-aesa-radars-into-broadband-comlinks/index.php, [Date accessed 3 May 2006].
[12] J. Katzman, “AESA Comlinks: DID Reader Has Done Prior Research”, Defense Industry Daily, URL http://www.defenseindustrydaily.com/2006/01/aesa-comlinks-did-reader-has-done-prior-research/index.php, last accessed April 2006.
[13] C. Kopp, “Sensors, Weapons and Systems for a 21st Century Counter Air Campaign”, Control Of The Air—The Future of Air Dominance and Offensive Strike Conference, ADFA, Canberra. November 15–16, 1999.
[14] C. Kopp, “Microwave Monoliths bring Moore's Law to RF”, Comms World, July 2000, Auscom Publishing, Sydney, URL: http://www.ausairpower.net/AC-0700.html , last accessed April 2006.
[15] Kopp C., “Moore’s Law and its Implications for Information Warfare”, The 3rd International Association of Old Crows (AOC) Electronic Warfare Conference, Zurich, May 20-25, 2000.
[16] C. Kopp, The Properties of High Capacity Microwave Airborne Ad Hoc Networks, PhD Thesis, School of Computer Science and Software Engineering, Monash University, Melbourne, 2000.
[17] C. Kopp and C.S. Wallace, TROPPO—A Tropospheric Propagation Simulator, School of Computer Science and Software Engineering, Monash University, Melbourne, Technical report 2004/161, 2004.
[18] C. Kopp, “Reflections on Information Age Air Warfare”, Journal of Information Warfare, Edith Cowan University, Perth, WA, Australia, ISSN: 1445-3312, Vol. 3, No. 3, pp. 11–28, November 2005.
[19] C. Kopp, Boyd, Metcalfe and Amdahl, “Modelling Networked Warfighting Systems”, 5th Australian IW&Sec Conference, November 2004. (Slides) http://www.csse.monash.edu.au/ ~carlo/archive/PAPERS/IWC5-Kopp-2004-Slides.pdf, accessed May 2006.
[20] C. Kopp, “NCW 101—Part 1—Information and Why it Matters”, Defence Today, May, 2005, http:// www.ausairpower.net/DT-NCW-101-1.pdf, accessed May 2006.
[21] C. Kopp, “NCW 101—Part 3—JTIDS/MIDS”, Defence Today, September 2005, http://www.ausairpower.net/NCW-101-3.pdf. , accessed September 2006.
[22] J.D. Kraus, Antennas, Chapter 12, McGraw-Hill, Singapore, 1988.
[23] M.L. Meeks, Radar Propagation at Low Altitudes. Artech House. 1982.
[24] J. Nielsen, Nielsen’s Law of Internet Bandwidth, 1998, http://www.useit.com/alertbox/980405.html, [Date accessed 3 May 2006].
[25] R.L. Peterson, R.E. Ziemer, and D.E. Borth, Introduction to Spread Spectrum Communications, Prentice-Hall, Englewood Cliffs, NJ, 1995.
[26] USAF, Case Study: Impact Assessment On Precision Strike Weapon Data Link Systems To Accommodate IMT 2000, 5 January 2001, Eglin AFB, Florida, http:// www.ntia.doc.gov/ntiahome/threeg/33001/usafcasestudy.pdf, Date accessed 3 May 2006.
[27] M. Ryan and M. Frater, Tactical Communications for the Digitized Battlefield, Artech House, 2002.
[28] L. Van Brunt, Applied ECM, Vol 1, 1978, Vol 2, 1982, EW Engineering Inc. P. O. Box 28 Dunn Loring VA.
[29] “The Aerial Sniper: Emerging Indo-Russian R-172 BVRAAM Cooperation”, Vayu Aerospace and Defence Review, URL: http://www.vayuaerospace.in/2006issue1/sniper.html, Date accessed 3 May 2007.
[30] “Air-to-Air Missiles—R-37”, Wartechnic.ru, URL: http:// www.wartechnic.ru/rus/armament/airtoair/r37/index_en.shtml, Date accessed 3 May 2007.