Ad Hoc Management
The evolution of wireless devices and peripherals has been remarkable and heterogeneous networks can be formed in an ad hoc and arbitrary way. Therefore the need for managing mobile ad hoc networks exists. Ad hoc networks have been discussed for years among the research community and have received a lot of interest. Recently this interest has been reinforced and ad hoc networks have become an important research area for institutions worldwide. Departing from the traditional research paths which included routing and clustering , the field has been broadened and emerging problems like management paradigms and configuration techniques receive intense interest.
The task of network management has never been easy and dates back to the beginning of computer networks. Several factors such as unexpected link breakage, packet congestion or intruder attacks affected the effectiveness of management. In the past years, the dramatic growth of the Internet combined with the growing complexity of telecommunication networks has driven both the research community and the business interest in exploiting fixed networks management. The outcome of these efforts has been the development of effective and intelligent management systems which generally speaking have achieved a sufficient level of management quality. Today we witness a shift of the interest towards the management of ad hoc networks. Consequently after the maturity of routing protocols, which dominated the research interest for many years, the need for managing ad hoc networks has emerged and is currently receiving immense interest.
Managing an ad hoc network is fundamentally different from managing a fixed one because of its diverse nature. Links in an ad hoc network are by default unreliable and intermittently connected. Link breakage is frequent and normal; therefore the repair of a link is not an issue as in fixed networks. Nodes can be arbitrarily up or down since mobility and resource management factors are involved. Once again this fact is unavoidable and inherent in ad hoc networks. Thus we understand that in order to manage an ad hoc network efficiently, we have to address several unresolved issues and provide solutions tailored for ad hoc networks.
Critical issues which define the problem of ad hoc networks management are identified and are further explained in the next section:
- Organizational model
- Heterogeneity and Device classification
- User control and context awareness
- Management protocol
- Service discovery
This section of the Quality of Service Management Information Portal serves as a focal point for research related to the management of ad hoc networks which effectively creates the basis for the delivery of QoS.
The field of ad hoc networks has received intense interest both from industry as well as academia. For some years, research interest was focused on routing protocols with several proactive (OLSR, TBRPF) and reactive (DSR, AODV) variations arising. After the maturity of routing protocols and the establishment of sufficient performance at the network layer, today interest has been shifted to the management of ad hoc networks. The penetration of wireless devices in all aspects of our environment created the need for novel management paradigms. These paradigms need to be customised for the specific needs of a mobile ad hoc network and should consider the important differences between these networks and the fixed ones. Critical issues that should be considered in designing a management solution for mobile ad hoc networks are presented below.
As we have already mentioned, the diverse nature of ad hoc networks call for differentiation from traditional organizational models. The centralized model of manager-agent is dominant in the management of fixed networks for various reasons. However the adoption of this model for managing ad hoc networks is not efficient and other models should be considered. All views tend to adopt a distributed model and several variations exist.
Focusing our interest in ad hoc networks we can easily understand why a centralized organizational model is not suitable. In an ad hoc network we can not rely on a single central entity to manage the network because nodes are intermittently connected. Therefore nodes may appear and disappear any time for example due to radio environment variations or due to battery exhaustion. In the case that the manager node disappears then inevitably the network remains unmanaged. The major problem of a single point of failure introduces the need for a distributed organizational management model. Management intelligence should be spread among nodes making the network fault tolerant.
Beyond the specific to ad hoc problems mentioned previously, problems that apply to fixed networks are magnified in ad hoc ones. In large scale networks the task of centralized management requires a considerable message overhead which may cause congestion problems. Overprovisioning the network resources is a common remedy in this case. Conversely this solution is not applicable to ad hoc networks since they have very limited bandwidth and the high message overhead involved in management would consume the scarce node’s resources. So concluding, we come across to ad hoc networks’ properties, like intermittent links, sparse bandwidth and limited resources, which make the centralized model for management inapplicable.
Heterogeneity and Device classification
The essence of ad hoc networking is connecting heterogeneous devices via wireless interfaces. These devices could be either mobile or fixed. Therefore the interoperability of heterogeneous devices and networks is an important issue to investigate.
An ad hoc network is formed by multiple devices with different capabilities so issues of device interoperability and device classification emerge. Mobile phones, PDAs, laptops and other devices must be able to communicate with each other and most of all be managed in a consistent and unified way. However, the different capabilities of each device dictate different potential roles. The amount of memory, storage space and processor speed vary significantly among them which makes it difficult to implement a uniform programming interface on all nodes. The management system should be able to exploit for example the presence of a high-end laptop in the network. On the other hand, a mobile phone is not expected to host e.g. an LDAP directory. Therefore a device classification must be considered in combination with context aware information as discussed in the next sub-section.
Furthermore, issues of network interoperability could be considered since an ad hoc network can be seen as an extension of fixed networks. Stand alone ad hoc networks are possible, but it would be useful to consider the presence of an access point to either the Internet or a mobile operator’s network.
User control and context awareness
The wide adoption of ad hoc networking in various aspects of our communication needs has diversified the role of the managing and the managed entities. An individual user in an ad hoc network demands control over his/her device and is not willing to grand permission for reconfiguration of his/her personal settings. Such a reconfiguration might be implied from the enforcement of policies in a node which just joined the ad hoc network.
Moreover, advanced wireless devices can provide several data which can help ad hoc networks management, e.g. a GPS receiver providing location data for routing purposes. However data like user location are sensitive personal data and the user should explicitly permit or deny access to them. In general, context-aware management could exploit context information available on a user’s device, but a user’s permission must be granted. The option of context disclosure should also be given.
In order to integrate all the aspects of management a suitable management protocol has to be considered. The currently available management technologies are oriented towards the management of fixed networks. Therefore some of their features make them inappropriate for ad hoc network management. The centralized manager-agent approach of SNMP makes it ineffective for the management of ad hoc networks, as we discussed in previous sections. We are not considering the OSI System Management architecture because of its strict and heavyweight architecture which makes it virtually impossible to implement on resource-constrained devices, which usually form ad hoc networks. On the other hand it is also rather heavyweight and its effective applicability to ad hoc networks is questionable. The same limitations apply to emerging XCMS (XML-based Configuration Management System) e.g. WebServices platforms, which need to be further investigated.
The nature of ad hoc networks calls for an effective service discovery process, since the appearance and disappearance of nodes is arbitrary. This is an issue of great importance because different ad hoc nodes have different capabilities and different needs. Each node needs to know where it can retrieve or update its management information and on top of that it may require additional resources from the network. Different approaches have been adopted such as a publish-subscribe model or peer-to-peer paradigms. The solution should minimize the overhead cost and save the scarce resources of ad hoc nodes. At the same time it needs to be fast enough so as to effectively serve moving ad hoc nodes. Therefore node mobility adds further design requirements.
Last but not least, security is an issue which needs special attention. Once again, the nature of ad hoc communication adds to the problems complexity and requires significant effort to ensure secure networking. In ad hoc networking, the security issues are more difficult because the air interface is used and access to it can not be controlled. Security features should exist in the management system without making it resource demanding and hard to implement. Also the lack of a centralised coordinator makes this task more difficult, since we can not rely on a certificate authority for example.
Furthermore novel security techniques as the detection of misbehaving nodes could be devised in order to find and block nodes that may compromise system’s reliability and safety. Misbehaving nodes may be legitimate users who act in a malicious or a selfish manner. Malicious nodes may flood the network with useless data, similarly to a DoS attack, while selfish nodes may not forward packets for routing purposes. Some issues are unique to the nature of ad hoc communications and need further investigation.
- W. Chen, N. Jain, S. Singh, "ANMP Ad hoc Network Management Protocol," IEEE Journal on Selected Areas in Communications, Volume 17, Issue 8, pp. 1506 - 1531, Aug. 1999.
- R. Chadha, Y.H. Cheng, J. Chiang, G. Levin, L. Shih-Wei, A. Poylisher, "Policy-based Mobile Ad Hoc Network Management for Drama," IEEE Military Communications Conference (MILCOM), Volume 3, pp. 1317–1323, Oct. 2004.
- R. Badonnel, R. State, O. Festor, "Management of Mobile Ad-hoc Networks: Evaluating the Network Behaviour," Proceedings of 9th IFIP/IEEE International Symposium on Integrated Network Management (IM), pp. 17–30, May 2005.
- C. Shen, C. Srisathapornphat, C. Jaikaeo, "An Adaptive Management Architecture for Ad Hoc Networks," IEEE Communication Magazine, Volume 41, Issue 2, pp. 108 – 115, Feb. 2003.
- R. Chadha, H. Cheng, Y.H. Cheng, J. Chiang, A. Ghetie, G. Levin, H. Tanna, "Policy Based Mobile Ad hoc Network Management," Proceedings of 5th IEEE International Workshop on Policies for Distributed Systems and Networks (POLICY), pp. 35–44, June 2004.
- K.S. Phanse, L.A DaSilva, "Protocol Support for Policy-based Management of Mobile Ad Hoc Networks," Proceedings of IEEE/IFIP Network Operations and Management Symposium (NOMS), Volume 1, pp. 3-16, April 2004.
- M. Frodigh, P. Johansson, P. Larsson, "Wireless Ad Hoc Networking – The Art of Networking without a Network," Ericsson publications http://www.ericsson.com/about/publications/review/2000_04/files/2000046.pdf, Accessed Sep.2005.
- M. Burgess, G. Canright, "Scalability of Peer Configuration Management in Logically Ad Hoc Networks," e-Transactions on Network and Service Management, Vol.1, No.1, Second Quarter 2004.
- S. Sivavakeesar, G. Pavlou, C. Bohoris, A. Liotta, "Effective Management through Prediction-based Clustering Approach in the Next-generation Ad Hoc Networks," Proceedings of IEEE International Conference on Communications, Volume 7, pp. 4326-4330, June 2004.
- L.M. Feeney, B, Ahlgren, A. Westerlund, "Spontaneous Networking: An Application Oriented Approach to Ad Hoc Networking," IEEE Communications Magazine, Volume 39, Issue 6, pp. 176-181, June 2001.
- S. Gouveris, S. Sivavakeesar, G. Pavlou, A. Malatras, "Programmable Middleware for the Dynamic Deployment of Services and Protocols in Ad Hoc Networks," Proceedings of 9th IFIP/IEEE International Symposium on Integrated Network Management (IM)), pp. 3-16, May 2005.
- L.A. DaSilva, S.F. Midkiff, J.S. Park, G.C. Hadjichristofi, N.J. Davis, K.S. Phanse, Tao Lin, "Network Mobility and Protocol Interoperability in Ad Hoc Networks," IEEE Communications Magazine, Volume 42, Issue 11, pp. 88–96, Nov. 2004.
- G. Pavlou, P. Flegkas, S. Gouveris, A. Liotta, "On Management Technologies and the Potential of Web Services," IEEE Communications Magazine, Volume 42, Issue 7, pp. 58 – 66, July 2004.
- C. Mi-Jung, C. Hyoun-Mi, J.W. Hong, Ju Hong-Taek "XML-based Configuration Management for IP Network Devices," IEEE Communications Magazine, Volume 42, Issue 7, pp. 84–91, July 2004.
- N. Damianou, N. Dulay, E. Lupu, M. Sloman, T. Tonouchi, "Tools for Domain-based Policy Management of Distributed Systems," Proceedings of IEEE/IFIP Network Operations and Management Symposium (NOMS), pp. 203–217, April 2002.
- S.L. Keoh, E. Lupu, M. Sloman, "PEACE: A Policy-Based Establishment of Ad-hoc Communities," Proceedings of 20th Annual Computer Security Applications Conference, pp. 386–395, Dec. 2004.
- J. Schonwalder, A. Pras, J.P Martin-Flatin, "On the Future of Internet Management Technologies," IEEE Communications Magazine, Volume 41, Issue 10, pp. 90–97, Oct. 2003.
The list of journals, conferences and technical societies related to multicast communication does not mean to be exhaustive rather it is indicative. For additions/updates please contact the webmaster.