You are here: Home / Communications and Information Systems Group / Research / Research Themes / Network Management and Operations Research

Network and Service Management Research

Autonomic Network Management

Emerging interactive applications and the continuing migration of telecommunication services over Internet-based networking infrastructure demand better quality, robustness and resilience. These emerging operational requirements have introduced additional challenges to the design of the next-generation network management systems, given that the current management functions that lie “outside” the managed network are, in this context, even more rigid and inefficient in dealing with complex arising conditions. As a result, there is a need for introducing self-management intelligence within the network in order to make the latter more flexible and adaptive to changing conditions through feedback closed-loop control solutions. This research discipline is known as autonomic networking / autonomic network management and envisages “systems and networks that can manage themselves given high-level objectives by administrators.”

Over recent years we have been developing a distributed in-network resource management infrastructure operating using autonomic principles. The decision process is distributed across the set of edge nodes which are embedded with dedicated management logic that enables them to perform reconfigurations based on instant feedback regarding the actual state of the network. These nodes area organised into a management substrate used to exchange information for reaching and enforcing management decisions. We have considered the use of different topology models for the substrate organisation in order to minimise the relevant latency and communication overhead. We have also been developing efficient knowledge sharing mechanisms to to better support the decision making processes. We have used the developed infrastructure to implement in-network load balancing, energy efficiency, and cache management applications.

In the context of simplifying the complex task of managing contemporary networks, another key research area is policy-based management. We use this technology to govern the network behaviour by enhancing the hard-coded functionality of managed devices and management applications with interpreted logic that can be changed dynamically to cater for changing requirements. This allows for a certain degree of programmability without the need to interrupt the operation of either the managed systems or the management system itself. In this context, we also investigate methods by which high-level objectives can be automatically decomposed to network-level configurations, and also methods with which inconsistencies (conflicts) between policy rules can be detected and effectively resolved.

Software-Defined Networking

<To be provided>

Traffic Engineering and Load Balancing

Network traffic has been growing substantially over the last decade, which is mainly attributed to the advent of media-rich applications, content distribution networks and the explosive use of smart devices (smartphones and tablets) that allow people to be connected while on the move. Given that network links have finite capacity, traffic management aims at controlling and optimising routing configurations and traffic allocation in order to avoid congestion and support good levels of service quality. Current practices mainly rely on offline traffic engineering approaches where the expected traffic demand is calculated from previous usage and specific routing configurations are produced to optimise resource usage over timescales of the order of weeks. Routing configurations are typically computed by a central management system based on the estimation of the traffic demand for the next provisioning period.

Due to their static nature, centralised offline approaches are unable to deal with dynamic conditions that cannot be reflected in the demand estimation and, as such, can be sub-optimal in the face of unpredicted traffic demand or link failures. At CISG we have been investigating novel online mechanisms that dynamically adapt the network settings in short timescales based on real-time information received from the network. This can be done either in a centralised manner, with the network management system that has a global network view enforcing decisions at the network edges, or in a distributed manner with edge nodes coordinating among themselves and making local decisions in an autonomic manner that avoids overall conflicts.

A key optimisation objective considered when performing traffic engineering concerns load balancing. This involves the computation of traffic splitting ratios by management entities located at network edge nodes and subsequently the routing of traffic over multiple source-destination paths realised through explicit path technologies such as MPLS or through multi-topology interior gateway protocol (IGP) routing. In order to avoid inconsistent reconfiguration actions that can result to network instability when performing load balancing in a distributed manner, the local edge node managers coordinate their decisions through the intelligent in-network substrate developed described in the autonomic network management section. We have obtained excellent results through both our adaptive centralised and distributed autonomic approaches in comparison to other approaches in the literature.

Energy Management

Energy awareness has been the subject of technological developments over the past decade, ranging from simple energy-saving techniques for battery-powered computer devices to more sophisticated techniques applied to data centres. The increasing power consumption of modern networks, driven by bandwidth-hungry applications, in conjunction with the rising cost of energy and increasing environmental consciousness, has led researchers to investigate methods by which the carbon footprint of network infrastructures can be reduced. At CISG we investigate ways to improve energy efficiency in the context of both communication networks and data centres.

The most prominent method to reduce energy consumption of communication networks is by powering off links/routers. However, powering these off may result in a partly disconnected network topology which may result in service degradation. At CISG we work on novel energy-saving approaches, which exploit the fact that many links in core networks are bundles of multiple physical cables, and adapt the link capacity at run time by switching off individual line cards (LCs). By controlling traffic flows in the network, some LCs can be offloaded and subsequently driven to sleep mode to save energy. Key issues we investigate involve determining the links in the network that traffic can be removed from or assigned to, but also the volume of traffic that can be re-directed to offload LCs, so that the energy optimisation objectives can be achieved.

In the context of data centres (DCs) we investigate methods to both optimise the use of energy within a single DC or across a set of co-operating, federated DCs. This optimisation is achieved by placing or migrating virtual machines between servers. At a single DC, we can optimise the energy consumption by exerting control on virtualisation systems in order to achieve an optimal distribution of virtual machines. We use a novel monitoring framework we developed to measure the energy consumption per server, networking or storage component and activate low-power states on devices which can be freed of their load. At a group of DCs, the goal is co-operation to optimise cumulative energy consumption by reaching an optimal distribution of virtual machines across all of the servers belonging to the group of DCs.

Smart Grid Control and Management

The energy sector is undergoing major transformative changes in recent years to address some pressing concerns in improving energy efficiency of the grid and reducing overall carbon emissions. The increasing penetration of distributed renewable energy sources (e.g., solar/wind farms), rising deployment of electric vehicles (EVs) and active consumer participation into power grid operations (e.g., interactive consumer applications) have all pushed today’s primitive and centralised grid information infrastructure to the limit. The dynamicity of power generation and consumption necessitate fundamental changes to the entire infrastructure.

In this dynamic environment, today’s centralised control is no longer sufficient to provide timely reaction (e.g., fault recovery, power load balancing). A distributed control and management system is needed. The emergence of intelligent devices / appliances also points to the need for bi-directional communication as oppose to the current one which is primarily one-way. In addition, the increased variability of power distribution (especially due to time-varying, non-deterministic distributed energy sources) calls for massive use of smart monitoring tools, e.g., phasor measurement units (PMUs), smart meters, etc., to improve the reliability of the power grid by enhancing the observability of the system. The next-generation power grid, known as the smart grid (SG), is envisioned to be resistant, extensible, scalable and secure to accommodate these new power grid requirements.

Our key activities within the smart grid communication research area take place under the auspices of the EU FP7 C-DAX project and are as follows:

  • SG communication infrastructure design and provisioning. The information infrastructure will play a central role as future power grids cannot be supported by centralised supervisory control and data acquisition (SCADA) systems. In this context, this activity investigates an integrated communication and information infrastructure with resilience, scalability and flexibility in mind for efficient support of massive integration of renewables and a heterogeneous set of co-existing SG applications. This activity targets the optimisation of network resource provisioning taking into account issues related to performance and QoS constraints imposed by the numerous smart grid applications, which present considerably different characteristics in comparison to Internet-based applications.
  • Information-centricity in Smart Grids. On top of the smart grid communication infrastructure, a wide set of applications necessitate the exchange of information between multiple devices and users, e.g., smart metering devices, power grid monitoring equipment, power grid operators, users, energy providers. In this context, this activity focuses on the application of the information-centric networking paradigm, as a means to reduce the excessive complexity of communication between a multitude of communicating endpoints through a public-subscribe ICN-like infrastructure. Our work builds on our experience in information centric networking research – see Network-layer research themes.
  • Integration of Electric vehicles. With sustainability in mind, various governments are already providing incentives for the use of more environmentally friendly electric vehicles (EVs). For instance, the UK government offers a Plug-in Car Grant while the U.S. federal government provides tax credits to EV buyers. This activity addresses the communication aspects of EVs including both Vehicle-to-Grid (V2G) and Grid-to-Vehicle (G2V) scenarios as well as the more advanced on-the-move EV communication for energy reservation and scheduling.

Robustness of SG – The demand for real-time responsiveness in the power grid has drastically strengthened the interdependence of the smart grid with communication networks. However, while their strong coupling may enhance their functionalities, it also significantly increases the vulnerability of the system as a whole. This is due to the fact that failures in one network may spread to the other and vice versa, resulting potentially in catastrophic cascading failures which may eventually bring down the entire grid infrastructure. This activity follows two main lines of research: (1) systematically quantifying the resilience of the interdependent systems against both natural disasters and malicious attacks and (2) the design of optimised protection schemes to minimise the effect of cascading failures.

Academics involved in the theme:

Representative projects:

      Representative publications:


      K.V. Katsaros, B. Yang, W.K. Chai, G. Pavlou, Low Latency Infrastructure for Synchrophasor Applications in Distribution Networks, Proc. of the 5th IEEE International Conference on Smart Grid Communications (SmartGridComm'2014), Venice, Italy, November 2014.


      K.V. Katsaros, W.K. Chai, N.Wang, G. Pavlou, H. Bontius, M. Paolone, Information-Centric Networking for Machine-to-Machine Data Delivery - A Case Study in Smart Grid Applications, IEEE Network, special issue on ICN Research Advances and Implementation, Vol. 28, No. 3, pp. 58-64, IEEE, May-June 2014.

      J. Araujo, R. Landa, R. Clegg, G. Pavlou, Software Defined Support for Transport Resilience, Proc. of the IEEE/IFIP Network Operations and  Management Symposium (NOMS'2014), Krakow, Poland, May 2014.

      J. Araujo, R. Landa, R. Clegg, G. Pavlou, Software Defined Support for Transport Resilience, Proc. of the IEEE/IFIP Network Operations and  Management Symposium (NOMS'2014), Krakow, Poland, May 2014.

      M. Charalambides, D. Tuncer, L. Mamatas, G. Pavlou, Energy-Aware Adaptive Network Resource Management, Proc. of the IEEE/IFIP Integrated Management Symposium (IM'2013), Ghent, Belgium, May 2013.

      D. Tuncer, M. Charalambides, G. Pavlou, N. Wang, DACoRM: A Coordinated, Decentralized and Adaptive Network Resource Management Scheme, Proc. of the 13th IEEE/IFIP Network Operations and Management Symposium (NOMS'2012), Hawaii, USA, April 2012.

      N. Wang, K. Ho, G. Pavlou, AMPLE: an Adaptive Traffic Engineering System Based on Virtual Routing Topologies, IEEE Communications, Vo. 30, No. 3, pp. 185-191, IEEE, March 2012.

      M. Charalambides, G. Pavlou, P. Flegkas, N. Wang, D. Tuncer, Managing the Future Internet through Intelligent In-network Substrates, IEEE Network, Vol. 25, No. 6, IEEE, November 2011.

      D. Tuncer, M. Charalambides, G. Pavlou, N. Wang, Coordinated, Decentralized and Adaptive Network Resource Management, Proc. of the 7th IEEE/IFIP International mini-Conference on Network and Service Management (mini-CNSM'2011), Paris, France, October 2011.

      S. Clayman, R. Clegg, L. Mamatas, G. Pavlou, A. Galis, Monitoring, Aggregation and Filtering for Efficient Management of Virtual Networks, Proc. of the 7th IEEE/IFIP International mini-Conference on Network and Service Management (mini-CNSM'2011), Paris, France, October 2011.

      L. Mamatas, S. Clayman, M. Charalambides, A. Galis, G. Pavlou, Towards an Information Management Overlay for Emerging Networks, Proc. of the IEEE/IFIP Network Operations and Management Symposium (NOMS'2010), Osaka, Japan, IEEE, April 2010.

      M. Amin, K. Ho, G. Pavlou. M. Howarth, A Closed-Loop Control Traffic Engineering System for the Dynamic Load Balancing of Inter-AS Traffic, Journal of Network and System Management (JNSM), pp. 343-370, Vol. 17, No. 4, Springer, December 2009.

      K. Ho, G. Pavlou, N. Wang, M. Howarth, Joint Optimization of Intra- and Inter-Autonomous System Traffic Engineering, IEEE Transactions on Network and Service Management (TNSM), Vol. 6, No. 2, pp. 64-79, IEEE, June 2009.

      UK EPSRC KCN Knowledge Centric Networking, 1/1/2015 – 31/12/2017, £396K