In wireless cloud storage systems, the recovery failure probability dependson not only wireless channel conditions but also storage size of eachdistributed storage node. For an efficient utilization of limited storagecapacity and the performance characterization of allocation strategies, weasymptotically analyze the recovery failure probability of a wireless cloudstorage system with a sum storage capacity constraint for both high SNR regimeand low SNR regime. Then, we find the optimal storage allocation strategyacross distributed storage nodes in terms of the asymptotic recovery failureprobability. Our analysis reveals that the maximal symmetric allocation isoptimal for high SNR regime and the minimal allocation (with $\lfloor T\rfloor$complete storage nodes and an incomplete storage node) is optimal for low SNRregime, where $T$ is the sum storage capacity. Based on the numericalinvestigation, we also show that in intermediate SNR regime, a balanceallocation between the minimal allocation and the maximal symmetric allocationwould not be required if we select one between them according to SNR.
展开▼
机译:在无线云存储系统中,恢复失败的概率不仅取决于无线通道状况,还取决于每个分布式存储节点的存储大小。为了有效利用有限的存储容量和分配策略的性能特征,轻松地分析了具有高SNR方案和低SNR方案的总存储容量约束的无线云存储系统的恢复失败概率。然后,根据渐近恢复故障概率,找到了跨分布式存储节点的最优存储分配策略。我们的分析表明,对于高SNR体制,最大对称分配是最优的,而对于最低SNR体制,最小分配(具有$ \ lfloor T \ rfloor $完整的存储节点和不完整的存储节点)是最佳的,其中$ T $是总存储容量。基于数值研究,我们还表明,在中等SNR体制下,如果根据SNR在最小分配和最大对称分配之间进行选择,则无需在最小分配和最大对称分配之间进行平衡分配。
展开▼