MVCE Strategic Partnership: Core 5 Flexible Networks

点击翻译按钮获取中文摘要

Today's fluid society is reflected in the choices we make with respect to relationships (such as to service providers or to other users), choices of identities, information we seek and many more issues. End user demands are met in a way that hides an increasingly complex ecosystem of service and facility providers, each of which is driven by their own business needs. In future networks will need to cope with increasing conflicts of interest between users, network operators and service providers (and more) - the concept of tussle. This will require a collaborative negotiation approach that enables simple and flexible arrangement of the components needed to the requirements of the stakeholders, performed in a way which is invisible to the end user.Traditional network design is done in a planned and static manner, reflecting the high investment and long lead times of new network deployments. It involves predicting future requirements, and designing to cater for that vision, following a prescriptive engineering approach. This static design approach is based on incomplete, a priori, information. Outdated solutions can thus reach the market which fail to meet (unpredicted) user needs; market drivers that motivated the initial design may have changed by the time the network is deployed. This has in the past caused certain features and services (that have been invested in and provisioned) to fail, due to lack of post-deployment customer take-up (e.g. mobile video calling). Conversely, other services (e.g. SMS and social networking) have emerged unexpectedly as hugely popular, mainstream applications. Introducing a degree of autonomy into network operation is seen by many as part of the solution, but on its own this may not address operator concern over network reliability and may not support increasing tussle between users, service providers, etc.This provides the motivation to bring together engineering and computer science to explore network technologies which can support the required properties of flexible networks. These include virtualisation, self-*, network coding and multi-homing. , which can provide structural flexibility in a future network, as well as technologies such as overlaying, auto-configuration and dynamic market approaches, which can provide the required control flexibility. Any network which embodies such powerful flexibilities must be trustworthy, in terms of its network composition and future self-evolution, hence novel network verification techniques are also investigated. This ensemble of technologies must be combined, using a novel design methodology, to enable the provision of future flexible, agile, self-evolving, robust and efficient networks that combine an ability to support the somewhat unpredictable needs of the future. A networking paradigm is required that is less prescriptive and enables network self-evolution.The project aims to enable such new networking paradigm and design methodologies for future flexible networks, which need to cope with unpredictable demand patterns, are likely to be self-evolving and will require different properties from those of today. It will create design methodologies that enable flexible networks (networks that are flexible, agile, dynamic, and self-evolving and accommodate tussle).In its first phase the project team will investigate, analyse and develop tools to model the conflicts of interest between stakeholders in future networks. Baseline metrics for network virtualisation and market based resource allocation will be identified and means to ensure robustness and increase efficiency across heterogeneous networks will be investigated. Finally requirements and tools for network verification tools capable to ensure trust and accountability will be defined.This project forms the first phase of the Flexible Networks project, part of the strategic partnership between EPSRC and Mobile VCE.