|Thesis abstract: |
The worldwide success and diffusion of content-rich applications, as well as the raising number of computationally powerful smartphones and tablets pose higher bandwidth requirements not only on the Radio Access Network (RAN) technology but also on the backbone network itself, since content-rich applications are extremely data-intensive. So far, in order to deal with the ever increasing customers¿ needs, the research community has mostly focused on new technologies that could reach higher data rates, while network operators have dimensioned their infrastructure with respect to peak traffic conditions.
However, since the new physical technologies have almost reached the theoretical channel capacity bound, it is extremely unlikely that this research path will obtain major performance gains in the near future. Not to mention, the redundant deployment of the infrastructure built to serve peak traffic conditions is having a negative impact on the environment and, at the same time, on the energy bill paid by the operators, since the RAN and backbone infrastructure are nowadays responsible for more than 80% of the energy consumption of the whole network.
Information Centric Networking (ICN) is a revolutionary design that is gaining momentum in recent years since it promises to benefit the user experience by leveraging universal in-network caching and nodes cooperation. Information Centric Networking proposes to change the basic networking paradigm by means of creating new protocols centred on the objects requested by the users rather then the locations where the information can be retrieved.
In the wireless context, a promising approach to further increase the network capacity while simultaneously reducing the energy consumption is represented by the opportunistic utilization of third party WiFi access devices, forming a Heterogeneous Mobile Network (HMN). This approach will allow operators to switch customer sessions from one access technology to another, in order to better manage their networks and accommodate more appropriately the QoS requirements of their users.
In order to jointly reduce the energy costs and improve the performance for data-intensive communications, we believe that radically new solutions are needed. In particular, our vision is to foster the opportunistic cooperation between nodes, since it is a viable way to preserve huge amount of energy while improving the user experience and it is also the distinctive feature shared by both HMNs and ICNs. Secondly, we recognize that new protocols are needed to better accommodate nowadays demands for communications. As a matter of fact, users are currently exploiting the Internet in a radically different way with respect to the original design goals that lead to the development of the worldwide network: people are mostly interested in accessing a given content rather than creating a point-to-point connection with two remote machines, and for this reason ICN approaches are extremely appealing.
This PhD thesis will investigate innovative policies and mechanisms to foster the deployment of the next generation of networks, with the precise aim to increase the capacity while saving large amounts of energy. In particular, the thesis will study how to integrate HMNs targeting wireless communications on the access network, and ICNs on both the backbone and the access network itself. Resource optimization will be performed to guarantee QoS requirements during network operation.
In this context we will:
1) investigate innovative mechanisms to foster the formation of Heterogeneous, Information-Centric Networks;
2) propose new optimization techniques to support the long and near-term planning strategies of the network, taking into account the particular features of the heterogeneous scenario as well as the availability of the distributed, universal caching;
3) design new algorithms and protocols for the dynamic management of Heterogeneous, Information-Centric Networks to simplify the reliable control of the system;
4) evaluate the proposed solutions using testbed implementations and real-life measurements.
Expected results and Evaluation criteria
The fundamental criterion to evaluate the research activities carried out within our project is the achieving of the project goals and the degree of dissemination generated by the project. More specifically the project activities should yield the following results:
- Design of auction and pricing mechanisms to jointly incentive network and caching resource sharing, in order to build globally efficient networks.
- Design of exact models and heuristics for planning optimal, heterogeneous networks, with a special attention towards the availability of the distributed caching mechanisms.
- Design of prediction techniques to forecast the behaviour of the overall system (traffic changes, user mobility) and dynamically reconfigure the network and redistribute cached content for improving the energy efficiency.
- On-field validation in dedicated network trials.
Of no lesser importance, exporting the proposed technology and solutions in the real world will brand the dissemination efforts with the mark of success. Further elements useful to the evaluation of the project results are the following: (1) Publication of the results in international journals and conference proceedings, (2) Discussion of the scientific results at international conferences and seminars, (3) Spreading of results and of the know-how acquired during the thesis through a project website.