Section: New Results

Data collection and aggregation

Participant : Nathalie Mitton.

Named Data Networking (NDN) is a new promising paradigm for content retrieval and distribution in the future Internet. NDN communication is driven by data consumers that broadcast Interest packets to require named contents. The requests are forwarded towards the source(s) by directly using content names (instead of IP addresses), while in-network caching is used to improve delivery performance. NDN shows many similarities with data-centric models defined for wireless sensor networks (WSNs), e.g., directed diffusion. In addition, NDN defines a new complete communication framework with innovative naming and security schemes and novel routing and transport strategies. This clearly opens new perspectives in the design and development of sensor networks, which can benefit of the NDN framework to better support different kinds of applications and services. In [16] , we explore the potentialities of NDN applied to WSNs and propose enhanced delivery strategies inspired by the directed diffusion scheme to be deployed in the NDN framework. Performance of a plain NDN scheme and of our enhanced solution is evaluated through the ndnSIM simulator. Achieved results confirm the viability of a NDN-like approach over WSNs and the better efficiency and effectiveness of the proposed solution compared to a plain NDN.

[38] considers the Slepian-Wolf coding based energy-minimization rate allocation problem in a WSN and propose a distributed rate allocation algorithm to solve the problem. The proposed distributed algorithm is based on an existing centralized rate allocation algorithm which has a high computational complexity. To reduce the computational complexity of the centralized algorithm and make the rate allocation performable in a distributed manner, we make necessary modifications to the centralized algorithm by reducing the number of sets in calculating the average energy consumption cost and limiting the number of conditional nodes that a set can use. Simulation results show that the proposed distributed algorithm can significantly reduce the computational time when compared with the existing centralized algorithm at the cost of the overall energy consumption for data transmission and the total amount of data transmitted in the network.