| Some Considerations on Resource Management in Wireless Networks | ||
|
by Jean-Paul M.G.
Linnartz
|
||
|
The Scarce Resource
In a communication system using cable, the spectrum efficiency is typically expressed in hr: the number of bits per second per Hz. For a wireless system, for instance a cellular telephone system, an optimization for hr alone never leads to satisfactory system efficiency. In fact modulation methods that allow the transmission of many bit/s per hertz, e.g. multilevel QAM, are sensitive to inference from other transmissions. Reuse of the radio spectrum in different geographic areas is the key to successful design of wireless systems but it leads to mutual interference. Therefore a more appropriate measure of performance is bit/s per hertz per unit of area or bit/s per hertz per base station. The scarce spectrum resource can be measured in the dimension time bandwidth space.
|
||
Anticipating on the results to come:
Efficient mobile packet data transmission, as needed in networks for multimedia traffic or in an ambient intelligence setting, requires entirely different spectrum reuse than telephone nets. Several examples have been covered to support this insight. To optimize spectrum efficiency, user capacity and network performance, presumably a new class of access schemes is needed that dynamically combine random access within one cell with protection against interfering signals from other cells.
We will address three schemes, namely Spatial collision resolution, STRMA, and the Virtual cellular network. each system is based on very aggressive frequency reuse, and resolution of access conflicts at a higher protocol layer.
Dynamic Channel Allocation (DCA) is known to provide a means to share bandwidth-time-space resources in a more dynamic and efficient way. However, DCA primarily works on a session by session basis, whereas bursty teletraffic is presumably best supported through access schemes that assign radio resources on a packet by packet basis, or at least on a burst by burst basis.
Under certain assumptions, Contiguous Frequency Assignment (CFA), i.e., cluster size C = 1, can support approximately 0.4 bit/s/Hz/cell. This appeared substantially more efficient than cellular frequency reuse with C = 3, 4, 7, ... . CFA also provides the smallest packet delay at a given spatial packet throughput intensity. This suggests that, in order to ensure minimum delay at maximum user capacity, mobile radio data networks should be designed with much denser frequency reuse than typically used to ensure an outage probability on the order of 10-2 or 10-3. This, however, results in low signal-to-interference ratios and large packet loss probabilities, which have to be addressed by appropriate data-link protocols.
In extreme cases, that transmissions to a particular terminal always see the same interfering base stations, so the number of required transmission attempts may become prohibitively large for certain terminals. This could have a detrimental effect on the entire network performance and it has motivated new 'spatial' random access protocols.
next: analytical framework