My personal Adaptive Global NET - MAGNET
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
WWRF8bis, Beijing, China, 26-27 February 2004
My personal Adaptive Global NET – MAGNET
Rasmus Løvenstein Olsen, Frank Fitzek, Ramjee Prasad, Juha Saarnio
Aalborg University
{rlo|ff|prasad}@kom.auc.dk
Nokia
juha.saarnio@nokia.com
The scope of this paper is to introduce the research project MAGNET supported within
the 6th framework of the European Commission. The project acronym MAGNET stands
for “My personal Adaptive Global NET”. As the name implies, the project has a very
strong emphasis on user-centricity, personalisation, adaptation, interoperability, personal
networking and interconnecting heterogeneous networks. The user-centric approach is
called for, since is it is widely accepted that future applications and services need to be
developed with the user in the driver’s seat and in equal pace with the corresponding core
and radio networks.
The MAGNET vision is that Personal Networks (PNs) will support the users’
professional and private activities, without being obtrusive and while safeguarding their
privacy and security. A PN can operate on top of any number of networks that exist for
subscriber services or are composed in an ad hoc manner for this particular purpose.
These networks are dynamic and diverse in composition, configuration and connectivity
depending on time, place, preference and context, as well as resources available and
required, and they function in co-operation with all the needed and preferred partners.
The project provides advances in the PN concept that include
• PN architecture
• Insight into business models for PNs and the related mobile data services in multi-
network environments
• Networking and interworking issues both at PN and PAN-level, in particular
resource and context discovery, self-organisation, mobility management,
addressing and routing, service discovery, and co-operation between public and
private, licensed and unlicensed networks
• Security and privacy issues in PNs
• Adaptive and (re)configurable radio access covering a wide range of data rates,
system capabilities and requirements, optimised for low-power and cost-
effectiveness
The PN Concept
PNs comprise potentially “all of a person’s devices capable of network connection
whether in his or her wireless vicinity, at home or in the office”. The work towards
enabling this vision transparently for users results in major extensions of the present
Submitted to WG3 (e) Page 1/8WWRF8bis, Beijing, China, 26-27 February 2004
Personal Area Networking (PAN) and Ambient Intelligence (AN) paradigms. PNs are
configured in an ad hoc fashion, as the opportunity and the demand arise to support
personal applications. PNs consist of communicating clusters of personal and foreign
digital devices, possibly shared with others, and connected through various suitable
communications means. At the heart of a PN is a core Personal Area Network (PAN),
which is physically associated with the owner of the PN. This is illustrated in Figure 1.
Unlike PANs, with a limited geographically coverage, PNs have an unrestricted
geographical span, and may incorporate devices into the personal environment regardless
of their geographic location. In order to extend their reach, they need the services of
infrastructure-based, and possibly also ad hoc, networks. A PN extends and complements
the concept of pervasive computing.
Remote personal
Home network
Local foreign devices
Communication Infrastructure
(Internet, cellular, WLAN, etc.) Corporate
Smart network
building
Ad hoc
Core PAN PAN
Remote foreign
Figure 1: Illustration of the Personal Network (PN) concept
The main components of a PN are:
• A core PAN consisting of personal devices in the close physical vicinity of a user
including devices moving around with him or her. The core PAN is an essential
component of the PN.
• Local foreign devices or clusters thereof, which are owned by other parties and could
either be reserved solely for the PN owner or be shared with others. They are linked
to the core-PAN via communication infrastructures.
• Remote personal devices, which are grouped into co-operating clusters and which are
linked to the core-PAN via communication infrastructures.
• Remote foreign devices or clusters thereof, which are linked via communication
infrastructures, and again can be shared with many users or be reserved for the PN
owner.
Submitted to WG3 (e) Page 2/8WWRF8bis, Beijing, China, 26-27 February 2004
• Communication infrastructures, in principle wide area networks (WANs) making use
of some sort of infrastructure-based resources (CN), which can be public (e.g.,
cellular, Internet) or private (e.g., leased lines), licensed or unlicensed (e.g., WLAN).
A simple example is a PN-based remote babysitting application (
Figure 2) [1]:
Consider the case of a mother visiting a friend’s house while her child is asleep at
home (supervised by a person not specialised in childcare). She might want to
remotely watch and observe the child. She does this by using a PN consisting of some
personal devices, e.g., a UMTS and Bluetooth-enabled PDA and a headset she carries
with her, and, a remote pair of eyes and ears in the child’s bedroom at home. The latter
consist of a digital video camera, a microphone and a UMTS phone, forming a cluster
of co-operating devices.
Alternatively, since the friend’s living room is equipped with a TFT wall display
including speakers, hooked up to a home network and accessible to authorised guests
via a Bluetooth the mother may want to use these instead to observe the child.
Remote personal
devices
Core PAN
(Internet, UMTS, WLAN, …)
Remote foreign
devices
Local foreign devices
Figure 2: Remote babysitting application [1]
State of the Art
Even if the PN concept above fits in the visions produced by different groups and from
different perspectives: the IST Advisory Group [2], WWRF [3], KTH [4], the UK
Government’s “Foresight” Initiative [5], the UK Mobile Virtual Centre of Excellence [6],
EURESCOM [7] and the Association of Computing Machinery (ACM) [8], to the best of
Submitted to WG3 (e) Page 3/8WWRF8bis, Beijing, China, 26-27 February 2004 our knowledge, there are no published results on PNs as envisioned by the MAGNET project. However, there are a number of approaches and research initiatives, which are of relevance to PNs, and we provide a brief survey of these below. User requirements The dominant approach to user requirements is that services and applications are shaped by the combined influence from terminals and networks developed according to the current technological possibilities, i.e., user requirements are not taken into account during initial conception. Service development then involves the PAN and a combination of networks: Public Switched Telephone Network (PSTN), cellular networks, digital broadcasting networks, as well as Bluetooth, WLAN, the Internet, etc., and the combination of these heterogeneous networks. MAGNET’s approach to user requirements is different. The methodology to describe and develop understanding for implementation of an efficient PN solution in a heterogeneous, multi-modal environment involves ‘technology’, ‘user needs’ and ‘economics’. A key element of ‘user needs’ is perceived quality of service (cf., Quality of Experience) associated with given private and/or business activities and its relation to the underlying technologies. The introduction of PN services along with the associated technologies will constitute a major paradigm shift. There are currently very few business models or scenarios in place for PNs, however, an enhanced understanding and knowledge of possible business model solutions as well as market and socio-economic aspects are necessary in order to achieve the full benefits of a heterogeneous communication model as proposed in the PN concept. Networking A number of system aspects are important when addressing the PN networking issues: middleware for mobile distributed systems, resource and context discovery, addressing and routing, self-organisation, mobility management of sub-networks and service discovery and provisioning in heterogeneous environments. Middleware for mobile distributed systems – Middleware architecture involving a PN- like architecture has been developed in project MOPED [9]. This architecture solves many problems by relying on a dedicated infrastructure-based proxy, like a home agent in Mobile IP. MAGNET on the contrary intends to avoid infrastructure-dependence. Resource, context and service discovery – Resource description system, e.g., Intentional Naming System (INS) developed at MIT, is an essential element of resource discovery. IETF Service Location Protocol (SLP), Sun’s Jini service directory, the Simple Service Discovery protocol (SSDP), universal plug-and-play, and Berkeley’s service discovery service are existing protocols, which are not suitable for PNs, because they do not support the highly dynamic context to be expected in PNs and preclude ad hoc operation since they involve infrastructure-based servers. IBM’s DEAPspace project has developed a resource discovery mechanism, which takes these considerations into account. Context- Submitted to WG3 (e) Page 4/8
WWRF8bis, Beijing, China, 26-27 February 2004 aware applications exploit information about, e.g., the geographical location, the time of day, the available equipment, the interaction history and the presence of other persons to provide the user with the service, which is best suited to the user’s circumstances. Self-organisation – IEEE 802.11 provides link level self-organisation. For Bluetooth networks, likely to be an important link technology for the device clusters constituting a PN, the self-configuration of so-called scatternets consisting of multiple pico-nets is still a research topic. For the network level the problem has been studied extensively. Here, however, the problem has a totally different dimension, and scalability becomes an issue. Addressing and Routing – Many routing strategies have been devised and analysed [10], e.g., long distance geographic routing is a technique, which relieves the nodes from keeping volatile network state information about distant nodes and links. Mobility management of sub-networks – new solutions are needed when dealing with the mobility of terminal devices and sub-nets. Worth mentioning in this context are the activities on mobile networks within the Mobile IP Working Group [11] of the IETF and the work on extensions of mobile IP for mobile ad hoc networks interconnection [12]. Ad hoc networks have received a lot of attention in recent years, in particular, the efforts around the MANET [11]. Even when this work provides valuable insight in the problem, it has mostly concentrated on homogeneous nodes and single parameter optimisation. For all this, IPv6 has a critical role to play. Adaptive and Scalable Air Interfaces for PANs The MAGNET concept involves the development and provision of a highly adaptive and spectrally efficient PAN air interface. It is envisioned to utilise and enhance existing and develop novel air interface technologies, and also to provide an interworking structure in the form of a Universal Convergence Layer to enable an adaptive and spectrally efficient solution across legacy PAN technologies. MAGNET will interact with project MATRICE as well as the FP6 project 4MORE, to provide expertise and results to complement the work undertaken. The overlap of participants in the projects will guarantee a close link. MAGNET is partially based on project PACWOMAN but it differs in the way that communication ranges from low data rates up to high data rate (e.g., multimedia) for the same type of air interface. Additionally, the PAN communication is limited to the immediate personal space (
WWRF8bis, Beijing, China, 26-27 February 2004 in close-to-the-body situations, channel sounding has to take place in order to provide access to the mechanisms of the inherently (ultra-)wideband channel and its dynamics. An appropriate and accurate channel model will take these into account and will be able to produce results that approximate real time varying multi-path channels. Interference of UWB systems or co-existence of any air interface is a major issue and is currently being studied for fixed access mobile architectures. The ad hoc nature of the system proposed in MAGNET for the PAN is innovative, and the study of the interference issues is required to ensure proper co-ordination as well as to minimise the impact on other systems operating in the same frequency bands. As an alternative, the millimetre-wave band shall offer similar bit rates as UWB in case of short-range systems. The very large available bandwidth available (3GHz on the up and down links) offers high bit rates up to 200 Mbit/s as it has been demonstrated within the ACTS MEDIAN and SAMBA and IST BROADWAY projects, as well as RNRT project COMMINDOR. Security Current security solutions for wireless technologies such as the one from 3GPP for GSM/UMTS based on (U)SIM algorithms, IEEE 802.11i [14] drafts for WLAN security, or the Bluetooth security recommendations, are all tailored for securing the traffic exchanged between user devices and access points. Applications running on those entities need to secure their traffic using application-layer schemes such as Transport Layer Security (TLS) [13] or IPsec. Here IPv6 will have a major role to play. While combining both transport and application layers would result in end-to-end secure communication, this approach was optimised for the Internet with a fixed infrastructure (i.e., very low data rate, fixed links which can only be either available or unavailable, and in the former case (temporarily) congested of not) and powerful end devices capable of decrypting the secure message. To achieve the same level of security in ad hoc networks, approaches based on co-operative authorisation and distributed key management are being discussed. However, these solutions usually incur considerable overheads in terms of signalling, and thus bandwidth usage, and processing needs. In the case of mobile networks, first standardisation efforts are being carried out in the MANET and NEMO groups at IETF. Other IST project, such as the SHAMAN and PAMPAS have contributed to this area and several FP6 projects, namely UBISEC and Ambient Networks, will cover this area. MAGNET will formally liaise with them. However, a general solution that is both adaptable to the network conditions and end system capabilities as well as enabling inter- domain AAA negotiation is by large still missing. For the latter, MAGNET will formally liaise with FP6 project E2R, which builds upon the work of FP5 IST project MOBIVAS. Flexible platforms PAN is an emerging paradigm that has attracted lately an increasing amount of activity not only in academia, but also in industry. The prevalent approach to PANs concentrates on short-range communications with limited communications capabilities (e.g., Bluetooth Submitted to WG3 (e) Page 6/8
WWRF8bis, Beijing, China, 26-27 February 2004 that is a master-slave system with scatternet extensions). WLAN developments have in the meantime made it possible to build more sophisticated services based on short-range communications. PAN research follows mostly an ad hoc network approach, with IETF MANET-type routing. There is also a large body of research conducted on self-configuration ad hoc or sensor networks – a good example is the TERMINODES project in Switzerland. MAGNET follows a different path. It extends the state of the art, in the context of PNs that interact in wireless fashion within the user’s “personal bubble”. The work is focused on optimising the short-range wireless protocols; instead of ad hoc routing, the focus is on interoperability and inter-working of routing mechanisms between core-networks and PNs (including addressing schemes, service discovery, etc.). Another strong novelty is the objective of having PN networks that are seamlessly interoperable not only with any access technology, but also, through any type of core networks, be able to connect to other PNs. This means that PN-to-PN communications links need to be established that are secured through automatic VPN tunnelling, namely in the IPv6 context. Finally, it is envisioned that some active networking and agent brokerage structure could be used at the network edges to provide better flexibility and optimisation. Concerning reconfigurability, and its role and impact on PANs and PNs, MAGNET will interact with the FP6 IST projects E2R and SIMPLICITY. Specifically, E2R plans to develop a large-scale technology demonstrator and an interchange of requirement specification and device capabilities in the early phase, as well as exchange of information on the demonstration platform in the final phase will be formalised. Conclusion MAGNET is an FP6 project with a focus on Personal Networks (PNs). A PN enables the user to access not only the devices moving around with her, but also her other devices at remote locations, and to negotiate the use of services provided by other persons’ devices situated either locally or remotely. This is facilitated through the means provided by different kinds of communication infrastructures interconnecting the various short-range networks or stand-alone equipment capable of network connection. MAGNET also ensures that all the technological development is carried out at the same pace and complemented by user-centric studies of the needs and requirements, as well as, economic studies of feasibility. MAGNET endeavours to actively liaise with other FP6 projects touching similar R&D issues, as well as, use relevant results of those public projects already finished or at their finishing stages, e.g., several FP5 projects. Interested readers are referred to the project web pages at www.ist-magnet.org or to contact Kjeld B. Olesen, Administrative Project Manager Aalborg University, Niels Jernes Vej 10, 9220 Aalborg East, Denmark, Phone: +45 9635 7343, Fax: +45 9815 9757, e-mail: kbo@aua.auc.dk Submitted to WG3 (e) Page 7/8
WWRF8bis, Beijing, China, 26-27 February 2004
References
1. I.G. Niemegeers and S.M. Heemstra de Groot, “Personal Networks: Ad Hoc
Distributed Personal Environments”, invited paper at Med-HocNet, IFIP Conference
on Ad hoc Networks, September 2002.
2. K. Ducatel et al., Editors, “Scenarios for Ambient Intelligence in 2010”, IST
Advisory Group (ISTAG), European Commission, Brussels,
www.cordis.lu/ist/istag.htm, 2001.
3. W. Mohr et al., Editors, “The Book of Visions 2000”, Version 1.0, Wireless Strategic
Initiative, November 2000. www.wireless-world-research.org/general_info/
Bookofvisions/BookofVisions2000.pdf
4. J. Zander et al., “Telecom Scenario’s in 2010”, Internal report, KTH, Sweden, 1999.
5. UK Foresight, http://www.foresight.gov.uk/
6. Core 3 Programme, Mobile Virtual Centre of Excellence, http://www.mobilevce.com/
7. “Systems beyond 3G – Operators Vision”, 7th WWRF Meeting, Einhoven, The
Netherlands, December 2002.
8. “The Next 1000 Years”, Special Issue of Communications of the ACM, Vol.44, No.3,
March 2001.
9. R. Kravets, C.Carter and L.Magalhaes, “A Co-operative Approach to User Mobility”,
Computer Communication Review, Vol.31, No.5, October 2001, pp. 57-69.
10. C. E. Perkins, “Ad Hoc Networking”, Addison Wesley, 2001.
11. IETF MANET Working Group, http://www.ietf.org/html.charters/manet-charter.html
12. U. Jönsson, F. Alriksson, T. Larsson, P. Johansson, and G. Q. Maguire Jr.,
"MIPMANET – Mobile IP for Mobile Ad Hoc Networks”, Proceedings of the
IEEE/ACM Workshop on Mobile and Ad Hoc Networking and Computing, Boston,
USA, August 2000.
13. T. Dierks and C. Allen. The TLS Protocol — Version 1.0. IETF RFC 2246, January
1999. http://www.rfc-editor.org/rfc/rfc2246.txt.
14. IEEE Std 802.11i/D3.0, November 2002 (Draft Supplement to ISO/IEC 8802-
11/1999(I) ANSI/IEEE Std 802.11, 1999 edition), Part 11: Wireless Medium Access
Control (MAC) and physical layer (PHY) specifications
Submitted to WG3 (e) Page 8/8You can also read
