Dinh Nguyen

Efficient multipath forwarding and congestion control without route-labeling in CCN

One advantage of content-centric networking lies in its built-in support for multipath forwarding. In this paper, we propose a multipath forwarding strategy in order to fully utilize available bandwidth in a content-centric network. Using the proposed strategy, each intermediate router sensibly adjusts the amount of traffic forwarded over a given face (among a set of faces) associated with a content name based on the estimated available bandwidth it observes receiving the content through the face. Available bandwidth is sampled and estimated from the actual Data packet flow received at the face. The advantage of our method is that forwarding decision closely matches the congestion status of the forwarding face which results in efficient bandwidth utilization without biasing short round-trip time routes. What is more, by aggregating congestion control states at end users, our method requires neither route labeling nor any other kind of cooperation among routers on the delivery paths. Simulations confirm the effectiveness of our proposed scheme.

Peer-to-Peer Content Distribution in Clustered Topologies with Source Coding

Network coding has been applied successfully in peer-to-peer systems to shorten the distribution time, particularly in extreme conditions where peers in the network are clustered and separated by physical links with limited bandwidth. In this paper, we focus more closely on the use of source coding, i.e. encoding is done only at the source, as an alternative to network coding to facilitate content distribution under such limited bandwidth configuration. We observe analytically and experimentally that, in this specific case, with appropriately chosen expansion factors, source coding can have comparable performance to network coding in terms of distribution time, yet consuming much less computational resources than the latter approach does.

Cache-Friendly Streaming Bitrate Adaptation by Congestion Feedback in ICN

HTTP adaptive streaming is an attractive solution to the explosion of multimedia content consumption over the Internet, which has recently been introduced to information-centric networking in the form of DASH over CCN. In this paper, we enhance the performance of such design by taking advantage of congestion feedback available in ICN networks. By means of utility fairness optimization framework, we improve the adaptation logic in terms of fairness and stability of the multimedia bitrate delivered to content consumers. Interestingly, we find that such fairness and stability have a very positive impact on caching, making streaming adaptation highly friendly to the ubiquitous in-network caches of the ICN architectures.

Minimal network coding redundancy for peer-to-peer content distribution

Network coding, with the redundant information it generates, has been shown to achieve optimal multicast throughput. Such redundancy plays a key role in maximizing throughput over multiple content delivery paths between network nodes. In this paper, we study the use of network coding to speed up content distribution in peer-to-peer networks where only a constraint number of selected peers can encode, i.e. become network coders. Our goal is to optimize the redundancy network coding generates, i.e. we want to answer the question how much a given network coder should encode. This serves as a step towards understanding how network coding improves content distribution in order to optimize its deployment. Given the network topology, we analytically figure the optimal redundancy ratio at each network coder to achieve shortest distribution time. Simulation results confirm the effectiveness of the proposed coding redundancy in shortening content distribution time while noticeably reducing the amount of encoding.

Rarest-first and coding are not enough

Network coding has been applied successfully in peer-to-peer (P2P) systems to shorten the distribution time. Pieces of data, i.e. blocks, are combined, i.e. encoded, by the sending peers before forwarding to other peers. Even though requiring all peers to encode might achieve shortest distribution time, it is not necessarily optimal in terms of computational resource consumption. Short finish time, in many cases, can be achieved with just a subset of carefully chosen peers. P2P systems, in addition, tend to be heterogeneous in which some peers, such as hand-held devices, would not have the required capacity to encode. We therefore envision a P2P system where some peers encode to improve distribution time and other peers, due to limited computational capacity or due to some system-wide optimization, do not encode. Such a system gives rise to a block-selection problem which has never happened in both pure non-coding and full network coding-enabled P2P systems. We identify the problem and fix the current block-selection algorithm to address it. Simulation evaluation confirms the effectiveness of our proposed algorithm without which the system performance degrades considerably.

Demo: Dynamic adaptive streaming over NDN using explicit congestion feedback

Many video streaming services employ the dynamic adaptive streaming over HTTP (DASH) technique to achieve better users QoE. Named-data Networking (NDN) can bring many benefits to the video distribution with in-network cache and stateful forwarding. However, simply applying DASH technique to NDN causes problems. First, in-network cache may affect download time which makes bitrate adaptation a difficult job. Second, various representations of the same video segments can consume cache storage and make caching less efficient. To overcome these problems, cache-friendly bitrate adaptation was proposed in [1] which adopts the bitrate based on the congestion feedback from intermediate routers. In this demonstration, we show the congestion feedback-aware DASH video application. Our contributions are twofold: (1) we customize the NFD to input congestion feedback into a data packet, and (2) we implemented the NDN-plugin for VLC player whose adaptation logic is based on the congestion feedback, and thus it can follow the current network status.

Cache the Queues: Caching and Forwarding in ICN from a Congestion Control Perspective

Caching and multipath forwarding are essential ingredients of the Information-Centric Networking (ICN) architecture resulting from ubiquitous content storage and content-based node-by-node forwarding nature of ICN. Much is yet to see how they jointly act towards improving users quality of experience. To this end, we formulate in this paper a unified problem of caching and multipath forwarding as a network optimization problem to maximize user satisfaction which is expressed by their utility function. The formulation allows us to see caching and multipath forwarding in ICN from a congestion control perspective and to reinforce the advantage of a class of congestion-aware caching where in-network caches are used to absorb network congestion and to enhance users satisfaction. In that context, multipath forwarding plays the role of directing content delivery to less congested paths where network capacity is abundant or where requested content has been stored by in-network caches. We evaluate such a congestion control-based, coupled caching and multipath forwarding approach in simulations. The result confirms the advantage of our approach compared with existing ones.

CoNAT: A network coding-based interest aggregation in content centric networks

Interest aggregation is one effective technique used in content-centric networking to speed up content delivery and reduce load on the network and the content sources. Interests, i.e. requests, for one content chunk are grouped together so that only the first Interest is forwarded toward the content source, and consequently, only one reply, i.e. Data, is responded for the whole Interest group. Ordinary Interest aggregation, however, is limited to Interests for the same content chunk which means Interests for a content should closely synchronize in time to be able to aggregate. In this paper, using network coding, we remove such limitation: An Interest for an encoded chunk of a content can be aggregated with a previous Interest for virtually any other encoded chunk of that content which contains new information. The crucial point is how to quickly verify the novelty of information contained in an encoded chunk. Our proposed network coding-based Interest aggregation is lightweight and effectively increases the number of Interests which can be aggregated. As a result, it reduces content delivery time as well as the load on network and content sources compared with the ordinary Interest aggregation without network coding. Simulation results confirm the effectiveness of our proposed method.