3.1 The Research Challenge
Smart Spaces
Many groups working on smart spaces have focused on supporting single applications in
intensively instrumented spaces, reducing many of the management problems, which arise
only from more general scenarios. The MIT Media Lab, for example, has conducted
extensive work on highly enabled spaces for children (KidsRoom); INRIA with smart
offices (INRIASmartOffice); the University of Reading with invisible building control
(WarwickImplants); and (interesting for being an outdoor system) a group at the
University of Lancaster with an electronic context-aware tourist guide (TouristGuide). All
the major corporations are also active in this applications area, with major initiatives
being announced by (amongst others) Sun, Microsoft, Ericsson, Nokia, Orange and Intel.
Such spaces have significantly simpler management requirements than the general case in
which we are interested, since it is possible to constrain the nature and behaviour of the
devices in the space very closely. The general case — where the population of devices in
a space is highly dynamic and largely unconstrained — opens up major research
challenges. Without a structured approach to management these challenges must be
handled ad hoc by individual architectures or applications — a phenomenon we may
observe in a number of current projects building smart-space infrastructures
(ContextToolkit and MetaGlue).
More interestingly, the behaviour of devices in a smart space is affected by a number of
factors not directly related to the devices concerned but rather determined by the use of
the space in terms of task allocations and architectural morphology. Knowledge about the
tasks undertaken within a space can be used to inform management decisions about what
devices are likely to appear and the anticipated requirements (CoenStopWorrying).
Equally, the way in which an architect designs a space (or set of spaces) radically affects
the movement potential, communication patterns etc. which are possible within it
(SpaceIsTheMachine). This makes managing the smart space a major challenge in holistic
systems engineering as well as in telecommunications.
Management
The development of an agreed set of standards to support the management of
telecommunications networks and services has been a major objective of the
telecommunications community since the early 1970s. Telecommunications management
may be defined as supporting the operation and maintenance of networks and services.
Today’s telecommunications management systems are required to:
A major problem with the interconnection of smart spaces is the lack of a Smart Space
management infrastructure. The choice of telecommunications management as the basis
for the development of M-Zone management solutions is based on the similarity of
requirements on the stability and maturity of today’s telecommunications networks.
Wireless Network Management
The state of the art of wireless communications management can be described along two
lines, that is management protocols and wireless network technologies. Two primary
protocols are currently used to implement network management functions, the Common
Management Information Protocol (CMIP) and the Simple Network Management
Protocol (SNMP). CMIP is an all-encompassing management standard with wide use in
telecommunication networks and demands significant processing capabilities restricting
its use to more powerful network nodes. The Internet-based SNMP is a lightweight
protocol, which makes it suitable for implementation on nearly all devices. Its success in
assisting the dramatic Internet explosion has lead to its adoption for next generation
wireless networks. Mixed management of SNMP and CMIP networks and achieving a
common view across these technologies is currently an area of significant research
interest. A more recent initiative from the Distributed Management Task Force (DMTF)
proposed the use of Web and XML technologies for the modelling and management of
networks and applications. However, this work is still maturing and has not addressed the
issues of smart spaces and wireless networks.
Wireless network technologies include cellular and personal communication networks,
wireless local area networks (WLAN), and short-range wireless mobile ad-hoc networks
(MANET). The management requirements between cellular networks and WLANs and
MANETs are quite distinct. Cellular networks are installed in a planned fashion and
controlled by a single network operator, whereas WLANs and MANETs are often
installed by their respective users in an ad-hoc fashion. This creates the design
requirement that these systems can operate and be managed in a distributed and selforganising
fashion. This requirement for self-organisation and self-management without
human intervention is a major focus of research in wireless systems management
(IETFmanet, IETFdesman). Within this context, radio resource management is of critical
importance as a system’s frequency spectrum is limited and therefore has to be managed
in an efficient manner in order to allow many users to use the system simultaneously
while still maintaining the guaranteed quality of service. Location management is a
further vital aspect in all wireless networks due to the mobility of users. It is therefore
significant to the objectives of the M-Zones programme as delivering telecommunications
services to mobile users across smart spaces is highly location dependent. Within this
context, location management makes use not only of the wireless communication system
infrastructure but will also utilise technologies such as GPS and sensors.
The scope for research in the area of wireless communication management within the MZones
programme will address the need for self-organisation and self-management of
wireless systems. The research challenges are to provide homogeneous management
across a range of functionally distinct wireless communication system architectures,
ranging from GSM and UMTS to Bluetooth, and wireless networking configurations from
planned cellular networks to self-organising wireless ad-hoc networks; to develop
management functions, which can facilitate end-to-end quality of service across
heterogeneous wireless communication systems while users are on the move as well as to
manage efficient sharing of the limited frequency spectrum among a large number of
wireless devices from smart cell-phones to wireless multi-modal sensor; and to develop
management algorithms that can be implemented on wireless devices with battery power
restrictions and limited processing capabilities.
|
