KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 4 of 4
  • Some of the metrics are blocked by your 
    Item type:Publication,
    An implementation of low-density MDS array codes for data protection in distributed network data storage system
    (2016-09-07)
    Puttarak, Nattakan
    ;
    Kaewprapha, Phisan
    We present an implementation of disk arrays(RAIDs) in distributed networked data storage environment by applying an error-correcting to accommodate fault tolerance feature. Any of the current state-of-the-art distributed file systems, such as MooseFS [1], can be used as an underlying data space where low density-MDS (Maximum distance separable-Low density parity check) array codes [3] is used as a logical protection layer implemented through FUSE interface (File System In User Space). The redundancy/parity scheme is based on graph structures leading to an MDS code, then can be simply implemented in a parity matrix (H matrix) of an LDPC code. The ability of error correction and data recovery is shown as the bit-error rate (BER) has been investigated. It achieves the same trend compared to the simulation results in [3]. The low-density MDS array code can mitigate hardware failure at higher disk space efficiency comparing to the repetition code.
  • Some of the metrics are blocked by your 
    Item type:Publication,
    Latency-reduced retransmission in multicast network using systematic network code
    (2016-09-07)
    Muangboonma, Nakhon
    ;
    Puttarak, Nattakan
    ;
    Kaewprapha, Phisan
    The reliability of multicast network where source delivers many frames of data to multiple receivers over a large scale network is degraded by delay of retransmissions after some data frame has been lost. This paper proposes a coding technique applying on network layer to combat this problem by using systematic network code with overheard retransmission technique. Latency of retransmission and throughput will be measured to represent the performance of this algorithm. The results show that this technique can improve reliability and performance of multicast network by decreasing a delay of retransmission for 40-70 % and yet increasing throughput for more than 8-19% compared with a conventional method.
  • Some of the metrics are blocked by your 
    Item type:Publication,
    Network localization using tree search algorithm
    (2016-09-06)
    Kaewprapha, Phisan
    ;
    Tansarn, Thaewa
    ;
    Puttarak, Nattakan
    we consider a network localization problem by modeling this as a unit disk graph where nodes are randomly placed with uniform distribution in an area where connectivity between nodes are defined when the distances fall within a unit range. Under a condition that a number of nodes know their locations (anchor nodes), this paper proposes a heuristic approach to find a realization for the rest of the network by applying a tree search algorithm in a depth-first-search manner utilizing graph properties to speed up the search, aka pruning the search tree by modeling an evaluation function from those properties. The evaluation function is used to select the order of the unknown nodes to iterate to. In [1,2], it is demonstrated that number of connections to previously localized nodes reduces the average feasible iterations within reasonable time. This paper also extends the idea further by accommodating various other properties of graph into the evaluation function. The results show that node degrees, node distances and shortest paths to anchor nodes drastically improve the number of iterations required for realizing feasible localization instance by 40%.
  • Some of the metrics are blocked by your 
    Item type:Publication,
    Low-density MDS array codes based on rigid graph structure
    (2014-10-15)
    Puttarak, Nattakan
    ;
    Kaewprapha, Phisan
    In data communications system including wired and wireless transmissions, the reliability, integrity, speed, and cost efficiency are the main issues that people desire. An error correcting code is one of the efficient and nearly-optimal methods to improve the system performance. Based on the idea of constructing MDS codes called CGR codes in [3] for disk arrays, this paper introduces a new perspective and construction of LDPC codes due to its sparsity and advantage complexity in implementation. The performance of this code in Additive White Gaussian Noise (AWGN) and fading channels gains about 1.0 dB at the 10<sup>-3</sup> of bit error rate (BER) compared with an EVENODD code.