Replication Placement on Disks

Abstract

When extra storage space is available on the striping disks being described in the last chapter, the storage system may keep extra copies of the stored objects to enhance the performance of the storage system. If any one of the copy or the original copy is corrupted, the corrupted copy can possibly be recovered by comparison with its replicas. The replication strategy thus increases reliability of the storage system by applying redundancy on the stored objects. Extra copies of objects may be created and stored on the storage system to increase the storage system performance. The presence of replicas on light loading disks may be able to reduce the period of time that an object is inaccessible. Thus, the replication strategy increases the availability of the stored objects. The replication strategy can have several advantages. First, the replica on idle disks can increase the availability of data on corrupted and busy disks. Second, the replica on local server can reduce the network load to access objects from remote servers. Third, the replica on local server can also reduce the need to wait for the filling of initial buffer prior to consumption. Fourth, replica can avoid disk multitasking by avoiding the need to serve multiple streams from the same disk head. We will describe the streaming redundant array of inexpensive disks (RAID) method that increases availability and fault tolerance in the next section. After that, we present the Lancaster storage server to reduce network load. Then, we show two data replication methods to reduce start-up latency. Afterwards, we explain how the data replication method can avoid disk multitasking. Before we conclude this chapter, we describe the replication method that balances the space and workload of storage devices.