An Analysis on Technical Scalability and Manageability of Storage Systems |
This paper presents a deterministic placement approachfor distributing orthogonal redundancy on distributed heterogeneous diskarrays, which is able to adapt on-line the storage system to thecapacity/performance demands by only moving a fraction of data layout. The evaluation reveals that our proposal achieve datalayouts delivering an improved performance and increased capacity while keepingthe effectiveness of the redundancy scheme even after several migrations.Finally, it keeps the complexity of the data management at an acceptable level.This paper describes the performance and manageability of scalable storagesystems based on Object Storage Devices (OSD). Object-based storage wasinvented to provide scalable performance as the storage cluster scales in size.For example, in our large file tests a 10-OSD system provided 325 MB/sec readbandwidth to 5 clients (from disk), and a 299-OSD system provided 10,334 MB/secread bandwidth to 151 clients. This shows linear scaling of 30x speedup with30x more client demand and 30x more storage resources. However, the system mustnot become more difficult to manage as it grows. This architecture has several advantages. A file’s physical location is decoupled fromits location in the namespace.This decoupling enables a powerful and flexible mechanism for the placement offile system objects. For example, different types of files, e.g., text orvideo, may reside anywhere in the namespace while being hosted by servers bestsuited to handling their content type. DiFFSalso provides lightweight protocols for online dynamic reconfiguration (volume reassignment and object migration) to addressfluctuating demand and potentially mobile file system entities. A DiFFS prototype has been implementedin Linux. Performance results indicate that the architecture achieves itsflexibility and scalability goals without sacrificing performance.