Erasure coding is increasingly deployed in modern clustered storage systems to provide low-cost reliable storage. In particular, Locally Repairable Codes (LRCs) are a popular family of repair-efficient erasure codes that receive wide deployment in practice. In this paper, we analyze the storage process formulated as a data partitioning phase plus a node selection phase for LRCs in clustered storage systems. We show that the conventional flat partitioning and random partitioning incur significant cross-cluster repair traffic, while the random node selection causes storage and network imbalance. To this end, we design a new storage scheme composed of an optimal partitioning strategy and an enhanced node selection strategy for LRCs. Our partitioning strategy minimizes the cross-cluster repair traffic by dividing each group of blocks into the minimum number of clusters and further compactly placing the blocks. Our node selection strategy improves load balance by choosing less-loaded clusters and nodes to store blocks with potential higher access frequency at higher priority. We implement our storage scheme on a key-value store prototype atop Memcached. Evaluation on a LAN testbed shows that our scheme greatly improves the repair performance and load balance ratio compared to the baseline.

Unlike general routing strategies, network coding (NC) can combine encoding functions with multi-path propagation over a network. This allows network capacity to be achieved to support complex security solutions. Moreover, NC has intrinsic security advantages against passive attacks over traditional routing techniques. However, due to the transmission of the global encoding kernels, the system is fragile to eavesdropping attacks with multiple probes. This paper proposes a generic unicast secure transmission scheme based on random linear network coding (RLNC). Specifically, the intended receiver generates a random matrix upon receiving the request from the source node, and then transmits each row vector of this matrix over a link reversely to the source node. Each intermediate node rearranges all received vectors to form a matrix by row, and then post-multiplies its local encoding kernel by this matrix to obtain a new matrix. Similarly, each row vector of the new matrix is reversely transmitted over a link to the source node. This procedure is performed until we have the source node, where the generalized inverse of the received matrix (or part of it) can be used as its local encoding kernel. Hence, the intended receiver can use the generated matrix (or the corresponding part) to decode the received data packets directly. We also analyze the security to demonstrate that the proposed scheme is at least as secure as other methods against wiretapping attacks. We also evaluate the performance of the proposed scheme to demonstrate its utility.