The what , The from , and The to : The Migration Games in Deduplicated Systems

Abstract

Deduplication reduces the size of the data stored in large-scale storage systems by replacing duplicate data blocks with references to their unique copies. This creates dependencies between files that contain similar content, and complicates the management of data in the system. In this paper, we address the problem of data migration, where files are remapped between different volumes as a result of system expansion or maintenance. The challenge of determining which files and blocks to migrate has been studied extensively for systems without deduplication. In the context of deduplicated storage, however, only simplified migration scenarios have been considered. In this paper, we formulate the general migration problem for deduplicated systems as an optimization problem whose objective is to minimize the system’s size while ensuring that the storage load is evenly distributed between the system’s volumes, and that the network traffic required for the migration does not exceed its allocation. We then present three algorithms for generating effective migration plans, each based on a different approach and representing a different tradeoff between computation time and migration efficiency. Our greedy algorithm provides modest space savings, but is appealing thanks to its exceptionally short runtime. Its results can be improved by using larger system representations. Our theoretically optimal algorithm formulates the migration problem as an ILP (integer linear programming) instance. Its migration plans consistently result in smaller and more balanced systems than those of the greedy approach, although its runtime is long and, as a result, the theoretical optimum is not always found. Our clustering algorithm enjoys the best of both worlds: its migration plans are comparable to those generated by the ILP-based algorithm, but its runtime is shorter, sometimes by an order of magnitude. It can be further accelerated at a modest cost in the quality of its results.

When Deduplication Meets Migration: An Efficient and Adaptive Strategy in Distributed Storage Systems

“…These two policies are in reality the same as our emphasized space efficiency rationale. As for migration traffic [44], i.e., the amount of data that is moved across servers, we think it is implicitly consistent with our space efficiency rationale. The intrinsic reason is that replicating a block means transmitting this replica across the network, leading to more migration traffic.…”

“…It is because the generated data replicas in our work provide the opportunity for block copies to work when a block suffers from hardware failure or software crash. Load balance [7], [44], [45] is a major concentration in distributed storage systems, which often conflicts with our space efficiency rationale. To be specific, the system's space cost can be minimized by mapping all files to a single server, which enables detection and deletion of all duplicate blocks.…”

“…Diverse Migration Policies: We carry out in-depth literature collection and policy classification on the migration policies in deduplicated storage systems, including replication cost [7], [12], fault tolerance [42], [43], migration traffic [44], load balance [7], [44], [45], and energy consumption [43], [46].…”