Sheng Lin

Expander code: A scalable erasure-resilient code to keep up with data growth in distributed storage

To ensure high reliability and storage efficiency, erasure codes are preferred in storage systems. With the prevalent of distributed storage systems such as clouds storage, how to design a scalable and efficient erasure-resilient code is challenging. We propose a scalable binary linear code to keep up with data growth which has the following properties. Given the group size k and the code block length n, the proposed code corrects any two bit erasures among the n bits. The redundancy overhead of the code is 2/(k + 2), and each data bit affects exactly 2 parity bits. As results of these properties, if a data bit is changed or added, only two parity bits need to be updated; and the recovery of an erasured bit requires accessing at most k other bits and the recovery of two erasured bits requires at most 2k other bits. We give the construction algorithm by the order expansion of regular graphs; moreover, we optimize the failure resilience during the construction procedure. Compared with existing codes, our proposed code has notable benefits in storage scalability, redundancy overhead and I/O bandwidth. The deployment of the proposed code in distributed storage systems can be simple and practical.

A Distributed MAC Layer Congestion Control Method to Achieve High Network Performance for EAST Experiments

Many applications would require fast data transfer in Wireless Local Area Networks (WLANs) nowadays. A representative example is the EAST (Experimental Advanced Superconducting Tokamak) project where physics researchers need to transport massive experiment data using the TCP (Transmission Control Protocol). However, the high contention level and the high error rate in wireless networks have a great impact on the TCP performance. To alleviate this problem, this paper proposes a MAC layer congestion control method to deal with wireless packet loss due to errors (as opposed to congestion). Our mechanism is implemented at the end wireless nodes based on the IEEE 802.11 DCF mechanism but without any modification to the TCP layer. We first propose the concept of the MAC layer congestion window in which the MAC layer will send all the packets in a window when it obtains access to the wireless channel (unlike the traditional DCF mechanism that just sends only one packet). Then we allow our congestion control mechanism to adjust its MAC congestion window based on the contention degree and the packet loss rate at the MAC layer. By performing wireless congestion control at the MAC layer, our mechanism can mitigate the effect of wireless packet loss on TCP, and therefore improve the TCP performance. The simulation and experiment results show that our mechanism can achieve better performance than the traditional MAC layer mechanisms in WLANs.

Optimal Decisions for a Fuzzy Two-Echelon Supply Chain

This paper discusses the optimal decisions of pricing and selling effort for a two-echelon supply chain with uncertain consumer demands, manufacturing costs, and selling costs. In order to maximize the -optimistic value of the profits, based on different market structures, one centralized decision model and three decentralized decision models are developed, and the corresponding analytical equilibrium solutions are obtained using the game-theoretical approach. The results illustrate that no matter what decision case is, the optimal retail and wholesale prices in the case of considering selling effort are, respectively, larger than those of no selling effort; the optimal profits of the manufacturer, the retailer, and the whole supply chain system in the case of considering selling effort are, respectively, larger than those of no selling effort except for the profit of the retailer in the case that the manufacturer plays the leader’s role. Finally, one numerical example is presented, which illustrates the effectiveness of the proposed models.

Extremal graphic model in optimizing fractional repetition codes for efficient storage repair

Consider that a set of balls of n different colors are thrown into m bins with the assumption that the ball number of each color is constant and the number of balls in each bin is also constant. Our optimal goal is to find a feasible placement such that the distinct colors of remaining balls should be at least c after removing any k bins (k ≤ m) with the minimum number of balls. We present that the optimal colored bins in bins is equivalent to the optimization of Fractional Repetition (FR) codes in distributed storage systems. Here balls correspond to coded packets and bins correspond to storage nodes. This problem can be represented as biregualr graph and then deduced to the Zarankiewicz problem, which is a well-known extremal graph theoretic problem. We present the problem with the relation to combinatorial design theory, especially t-designs and propose the explicit construction algorithm for the optimization problem from t-designs. Some constructions of the optimized FR codes by 2-designs are analyzed to tolerate the desired k fault-tolerance with c = n - 1.

Performance Optimization and Evaluation of Space Management in Cloud Storage Systems

Cloud service providers and enterprises usually deploy high performance storage to manage the unrelenting growth of data. In this paper, we focus on performance optimization and evaluation by using optimal regeneration codes in such cloud storage systems. We present an efficient data maintenance management framework to reduce network repair traffic with the minimum data movement while keeping the desired fault-tolerance in storage systems. In the management framework, it has two phases including the traditional erasure coding process and the optimal placement process. We formally represent the optimal placement as a variant of the bin packing problem by bipartite graphs. Then, we model the placement transform by the interchange graph and propose an efficient heuristic algorithm to find the optimal solution. All feasible solutions are linked together by interchange operation and thus the search space can be taken as an interchange graph. Finally, we evaluate the performance of the optimal placement during data maintenance with different practical settings in our experiments. The experimental results show that the amount of network repair traffic can be reduced by about (10,%) than the initial placement and by about 2X than traditional erasure coding placement.

An Enhanced TCP for Optimizing Channel Utilization in Dynamic Spectrum Access Networks

Dynamic Spectrum Access DSA network grows rapidly in recent years. It is proposed to solve the problem of reasonable utilization of wireless spectrum resources. DSA is a new spectrum sharing paradigm which takes advantage of spectrum holes to ease the spectrum shortage problem and improve the spectrum utilization. However, due to the frequent spectrum switch, the performance of TCP will degrade greatly in the DSA network, and thus the channel is unutilized. To tackle with this problem and improve high throughput performance, this paper proposes a protocol called TCP WSP based on the assumption that there will be a Wide Stationary Process WSP between two spectrum switches. Then the authors propose a mechanism to predict the spectrum switch and adjust the sending rate of TCP based on the prediction to optimize channel utilization in DSA networks. They implement our mechanism on NS2 simulator, and the results show that their mechanism can achieve high throughput performance.

New pattern erasure codes

In this paper, we study binary pattern erasure codes, i.e., binary codes that are resiliant to erasures from a family P of possible erasures. We give an algorithmic proof of the existence of a binary linear code with codewords of length n that is resiliant to erasures from P when P satisfies the properties: every pattern p ϵ P has size m and every letter in the alphabet occurs in at most c patterns. The density of the parity matrix is plays a important role in storage applications, so we also introduce a new low density code basing on graph theory.

A MAC layer congestion control method to achieve high network performance for EAST experiment

Many applications would require fast data transfer in Wireless Local Area Networks (WLANs). A representative example is that EAST experiment data are retrieved by some physics researchers using the Transmission Control Protocol (TCP). However, due to the high contention degree and the high error rate in wireless networks, the packets may be loss for wireless reasons but not for congestion. This will greatly degrade the TCP performance. On one hand, the wireless packet loss is not congestion, but the traditional TCP assumes that every packet drop is congestion and thus decreases its congestion window, which will degrade its performance. On the other hand, due to the MAC layer retransmission policy employed by the IEEE 802.11 DCF mechanism, the lost packets at the MAC layer will be retransmitted for some times. Thus the waiting time of the packets in the MAC layer queue will be increased. So if we ignore all the packet loss for wireless reasons as the other improved mechanisms do, the network work congestion will be aggravated and its performance will be degraded. To alleviate the impact of the wireless packet loss to TCP in WLANs, this paper proposes a MAC layer congestion control method which is implemented at the end wireless nodes based on IEEE 802.11b DCF mechanism. At first, we propose a concept of MAC layer congestion window which means the MAC layer will send all the packets in a window when it gets access to the wireless channel, other than just sends only one packet as the traditional DCF mechanism does. Then our congestion control mechanism adjusts the MAC layer congestion window based on the contention degree and the MAC layer packet loss rate. If the MAC layer contention degree or packet error rate is high, we will increase the congestion window to improve the successful transmission rate, and we will decrease the congestion window when the packet loss rate is lower than the average wireless packet loss rate. We also use a threshold to control the increase of the congestion window. The threshold is set according to the number of wireless nodes. By performing wireless congestion control at the MAC layer, our mechanism can mitigate the effect of wireless loss to TCP, and therefore improve the TCP performance. The simulation and experiment results show that our mechanism can have better performance than traditional MAC layer mechanisms in WLANs.