Jiaqing Huang

Polynomial Time Construction Algorithm of BCNC for Network Coding in Cyclic Networks

Network coding in cyclic networks meets more problems than in acyclic networks. Recently, S.-Y.R.Li et al.proposed a framework of convolutional network coding (CNC) as well as its four properties for cyclic networks with theoretic fundamentals of discrete valuation ring (DVR). The four properties, Convolutional Multicast (CM), Convolutional Broadcast (CB), Convolutional Dispersion (CD) and Basic Convolutional NetworkCode (BCNC), are notions of increasing strength in this order with regard to linear independence among the global encoding kernels. The existence of a BCNC then implies the existence ofthe rest. That is, BCNC is the best convolutional network code in terms of linear independence. However, the code construction algorithm of BCNC was not presented explicitly. To the best of our knowledge, this is the first paper to propose a polynomial time construction algorithm of BCNC for network coding in cyclic networks, which can deal with different characteristics of cycles in terms of topology, including link cycles but flow acyclic, simpleflow cycles and knots. Finally, polynomial time complexity of the algorithm was proved as well as its effectiveness.

A General Co/Decoder of Network Coding in HDL

This paper presents a practical and general coder and decoder of network coding (NC) with HDL (Hardware Description Language) logic for wire-speed nodes in multisource multicast networks. The NC coders apply random linear network coding (RLNC) and the decoders recover the original packets by Cramers rule. All these mathematical operations are carried out in the Galois Field (256). The structures and algorithms of NC coder and decoder were designed in detail and implemented in HDL with NetFPGA boards provided by Stanford University. Comparing with traditional stored-and-forward mechanism, network emulations showed that networks with wire-speed NC coder and decoder nodes could achieve the capacity bound of max-flow min-cut theorem in case of bottlenecks, and the end-to-end delay was guaranteed on a small constant.

A simulation study on network coding parameters in P2P content distribution system

We consider a peer to peer (P2P) content distribution simulation system based on network coding. In such a system, the data transmitted among peers is an encoded version of the original content. With network coding, the system can achieve the maximum theoretical network flow bound. However, in practical applications, the operation of network coding causes additional computational consumption and storage overhead, and whether the P2P system can benefit from network coding remains a controversial issue. In this paper, we introduce some new network coding parameters to gain more insights to the properties of network coding and have an extensive analysis from simulation. Based on these parameters, we construct an implementation of the network coding-based content distribution scheme in NS-2 platform. According to our simulation results, in some cases, network coding can not obtain a better performance compared to the traditional P2P scheme. However, when the parameters of network coding are chosen properly, the network coding-based system will gain a better performance compared to the traditional P2P system by at most 20% to 30%, which means that the use of network coding can give considerable benefit to the P2P content distribution system.

Unified construction algorithm of network coding in cyclic networks

Network coding in cyclic networks may have better performance than network coding in acyclic networks with regard to the multi-unicast scenarios. Harvey et al. showed that network coding in cyclic networks can be strictly better than fractional routing in conservative networks which have widely practical scenarios such as P2P networks. Hence, we motivated investigating how to achieve that better performance of network coding in cyclic networks by a general construction algorithm. Li et al. presented there are four levels for network code in cyclic networks, including Basic Convolutional Network Code(BCNC), Convolutional Dispersion(CD), Convolutional Broadcast(CB) and Convolutional Multicast(CM). Subsequently, it is interesting to investigate how to construct all four levels of network coding in cyclic networks. Based on our previous work of construction algorithm of BCNC, we proposed a unified algorithm to construct network coding in cyclic networks using notion of flow set. Our contributions were as follows:(1)we showed insights of the essential difference between two classes(i.e. BCNC and CD/CB/CM) of network codes in cyclic networks. (2)we showed insights how to uniformly handle cycles for these two classes of network codes in cyclic networks by proposing the unified construction algorithm. Here, we used the cycles classifications defined by Barbero et al., including link cycles but flow-acyclic, simple flow cycles and flow knots(simply knots).

Constructing Network Coding in Cyclic Networks with the Best Linear Independence

Capacity of Network coding in cyclic networks can be strictly better than routing. Harvey et al. showed that it exists in the practical scenarios such as the widespread P2P networks. Thus, it is of application significance to investigate how to construct network code in cyclic networks. According to Li et al., there are four levels of network code in cyclic networks, including Convolutional Multicast(CM), Convolutional Broadcast(CB), Convolutional Dispersion(CD), and Basic Convolutional Network Code(BCNC). This paper focuses on the BCNC that is with the best linear independence, which means it can provide the most efficient transmission in that it has the minimum redundance. The contributions of this paper are (1)to improve our previous BCNC algorithm; (2)to show insights of the essential difference between BCNC and CM/CB/CD which is of great theoretical significance to understand how different kinds of cycles affect the network coding construction.

Applying Network Coding to Cyclic Networks

This work addresses the application of network coding on cyclic networks. Cycles and delay are indispensable factors when designing and implementing data networks. However, most attention in the literature of network coding has been on acyclic networks. Hence, in order to bridge the gap of network coding and the practical networks, Convolutional Network Coding (CNC) with delay as an essential part was proposed for network coding on cyclic networks.

An Improved Network Coding-Based Cooperative Content Distribution Scheme

Reducing the additional computational overheads caused by network coding operations in the a novel cooperative content distribution paradigm which is based on network coding is valuable for both theoretical interests and practical interests. Regretably, little work about this problem, especially from the perspective of practical interests, has been reported in the present literature. To issue this problem, in this paper, we propose an improved network coding scheme which is based on the new mechanisms of clustered network coding, random choice and independence checking. Simulation results show that the proposed scheme can reduce the additional computational overheads significantly and achieve better performance than current used network coding scheme.

Sensing at the nanoparticle trapped tip of funnel nanochannel

This paper explored the ion transportation in a funnel nanochannel with nanoparticles trapped at the tip and its applicability in nucleic acid detection. Simulation showed that H+ distribution in funnel nanochannel was extremely sensitive to the surface charge density of nanoparticle trapped at the tip, which contributed to the reported high sensitivity. A high sensitivity of 1fM for biotin was achieved in 100μM KCl solution. Sensing capability of 22 mer ssDNA and 22 bp dsDNA was also proved for the first time, which provided a potential application in highly sensitive microRNA detection on chip.

Quantitative analysis for computation delay induced by network coding

Quantitative complexity of network coding as well as its computation delay on a node is presented in this paper. First, quantitative complexity of four fundamental calculation operations were analyzed. Based on that, quantitative complexity of encoding, decoding as well as independence detection for network coding were given theoretically. At last, the experimental results showed our quantitative results of computation delay have the same order to the delay of practical network coding on a node.

A load balancing scheme for two-stage switches with minimum buffers for scalability

In this paper, we propose a novel load-balanced scheme-split aggregated flow (SAF) and design the buffering mechanism for two-stage self-routing switch. By grouping the signal lines with concentrators, the proposed scheme can obtain the statistical multiplex gain and reduce the complexity of computation. The mathematical analysis and simulations show that 100% throughput can be achieved for any admissible traffic pattern. Compared with other load-balancing schemes used crossbar as the basic fabric, this scheme has the distinctive advantages including the lowest complexity buffer of O(N), low queuing delay O(1), and free from the out-of-sequence problem. These properties make it practicably suitable for very large scale switching structures in Next Generation Network (NGN).