Summary:
Cooperative systems and radios, which are capable of intelligently forming mutually cooperative entities, are a promising way to achieve optimal increase in capacity for current and future applications. Technological and social barriers have prevented this so far. The capacity offered by the wireless medium is unlimited. Cooperative, distributed relaying systems have received significant attention in the past decade. The authors respond to the question of whether such a technology is ripe for deployment or only the result of academic hype.
Cooperating with other individuals or entities, usually means short term losses, but may also mean long-term gain - something children do not yet recognize and adults rarely understand. Any cooperative technology depending solely on human decisions is hence a priori doomed to fail. Some recent failures: ad-hoc networks in general; 3rd generation partnership project (3GPP) on opportunity driven multiple access (ODMA); and Ricochet networks.
Ad-hoc networks have been researched for more than 30 years. So far no commercially viable product has appeared. Two reasons: Design degrees of freedom have turned out to be too large to reach commercialization because of a psychological barrier of both manufacturers and service providers. This has not been mastered for much simpler cellular systems; Data relaying process requires users to give away battery power and bandwidth, and possibly jeopardize the security of their own data, with no obvious instantaneous gains - that is, a psychological barrier prevailed at the user side.
3GPP uses in an opportunistic manner the 3G terminals within a cell to give coverage to terminals our of base station range. Suitable protocols have been proposed and overall system gains have been demonstrated. However, it is now withdrawn, because the users needed to authorize the sue of their own resources to facilitate the cooperative relaying process. Also ODMA is not stand-alone. Richochet (www.richochet.com) , developed a decade ago, was a broad band wireless network that formed a mesh network by means of relay-capable nodes attached to lampposts. Technology was ahead of its time. Today, Ricochet seems to be back and is likely to grow. These and other failures are rooted in the reluctance of users to relay other people's data for no apparent short-term gain and the inability of cooperative systems to guarantee constant and reliable availability.
Such failed projects contain lessons. Packethop (www.packethop.com) has applied the now affordable cooperative relaying technology to federal projects, as with Homeland security, where security needs may trump the user concerns. Commercial operators have other needs to meet: A minimum QoS (qualify of service) to the end user --can't use mobile terminals as their availaibility is not reliable. Hence the use of fixed relays, clearly cheaper than base stations, but more expensive than user terminals for relaying. Timus has shown that fixed relays do not yield significant cost benefits.
Such cooperative techniques are likely to survive where machine applications, not humans, make the decision to cooperate. An example: WSN (wireless sensor networks), where cooperation improves data reliability, energy savings, and network longevity. In the long run, incentive schemes will be needed - users who sacrifice their resources to cooperatively aid other users should be credited and those users who benefit from such cooperation must be charged. Software-controlled radios are in position to sense the radio context and knowledgeably draw a decision whether cooperation is beneficial and which steps have to be taken to set up a cooperative scheme.
IEEE P1900 is a standardization effort. An example of pouring funds into research areas without any quantified benefits is ad-hoc networks. Research entities need to synthesize and simplify the large degrees of freedom in cooperative systems. Development will be better optimized, energizing the business units.
For the everyday end user, cooperation should ideally go unnoticed - underlying technologies should make their own knowledgeable decisions in a transparent fashion. An example is France Telecom's seamless handovers between its cellular Global System for mobile communications (GSM) technology and its home livebox Wireless LAN once in communication range. A likely beneficiary of the widespread acceptance of cooperative technologies will be intelligent transportation systems (ITS) for road safety and driver information. Car-to-car cooperation could significantly improve the efficiency and safety of modern transportation systems. For example, vehicles could communicate about detour, traffic accident, etc. This cooperation could be used for cooperative driving. Another beneficiary is the environment. Cooperative techniques allow for great savings in the energy needed to transmit and support communications.
Allow users to be kept out of the decision loop as much as possible to get the most of this technology. 4G will most likely be a single cooperative entity.
Design concerns and design constraints of such large-scale cooperative networks will need to be better understood. Kumar and Gupta establish that the network throughput scales with 1/square root (M log M), where M is the number of terminals. Hence, no matter what we try, we cannot design a scalable protocol. This requires that topologies different from pure ad hoc have to be invoked. Recent studies show that network capacity can increase linearly by using cooperative communication hierarchies (circles of friends, with decreasing importance). Practical test runs: see EASY-C project (www.easy-c.de), in Dresden,Germany.
