This paper presents the design and implementation of the adaptable virtual storage system (AVSS) and introduces the capacity virtualization and performance virtualization for storage systems. AVSS has the following characteristics: 1) adoption of extended YFQ algorithm to control the allocation of bandwidth resources, realization of the performance isolation and guarantees of virtual disks; 2) adoption of hierarchy structure and dynamic mapping mechanism to manage heterogeneous storage resources flexibly and effectively, which lays a foundation for allocating storage resources on demand; 3) application of data-access frequency statistics and dynamic behavior analysis to supervise storage layout reorganization. The experimental results proved the correctness of our design. AVSS can isolate different applications and avoid performance interference. It can adjust storage layout according to the behavior of applications and improve the utilization of storage resources while improving the performance of the storage system.

It is a challenging task that enables performance isolation while providing performance guarantees among applications that share storage infrastructure. This paper presents admission-control-based proportional allocation (APA), a novel I/O request scheduler that combines the admission control of worst-case resources reserved with the proportional resources allocation to manage the bandwidth resources for performance isolation. APA has three characteristics: 1) admission control for I/O requests of applications; 2) fairness allocation for idle bandwidth resources; 3) performance monitoring and feedback-based weight adjustment. We implemented APA in a prototype performance virtual storage system (PVSS). Through experiments on a real system, we show that APA has ability to provide the performance isolation and predictable performance guarantees, and to avoid the performance interference among virtual disks.

On demand of various applications for storage virtualization, this paper presents a stackable block-level virtual snapshot device (VSD) based on the virtual storage pool. With the strategy of stackable block device, VSD employs a read-only snapshot model with serial connection, and a tree model mixed with serial and parallel connection for writable snapshot. Moreover, two data updating schemes, Copy on write and realloc on write, are integrated into VSD. Thus, VSD supports various applications such as data backup and recovery, data version management and volume clone and so on. Finally, the storage requirements and I/O performance of VSD are evaluated.

To achieve theoretical optimum of spatial locality based replacement algorithms, this paper proposes a Block Correlations Directed cache replacement policy (BCD) which uses both of history and runtime access information to predict spatial locality, designs a replacement policy to use prediction results to improve the utilization of the cache and reduces the penalty incurred by incorrect predications. For most of real system workloads, BCD can reduce the cache miss ratio by 11%~38% compared with LRU and the miss ratio is only high than the theoretical optimum by 0%-7.9%.

The paper presents a novel network RAID storage system based on the out-of-band virtualization architecture and the backend centralized redundant management. The application servers can inquire the mapping information of virtual disk from the out-of-band virtualization server and directly access the storage nodes. The read request can fetch the data from special storage node, while the write request is not only stored into the storage node, and also mirrored into the redundant server by the storage node. The redundant server can cache updated data in local disks with log-structured mode, and calculate the parity of RAID5 in the background process when the system is idle. It relieves the bottle problem of I/O performance and low reliability danger of single controller in the front-end centralized management system. And the layout of RAID1/RAID5 on the data block has acquired the trade-off among performance, reliability and cost.

In scientific computing environments, large amounts of floating-point data often need to be transferred between computers as well as to and from storage devices. Compression can reduce the number of bits that need to be transferred and stored. However, the run-time overhead due to compression may be undesirable in high-performance settings where short communication latencies and high bandwidths are essential. This paper describes and evaluates a new compression algorithm that is tailored to such environments. It typically compresses numeric floating-point values better and faster than other algorithms do. On our data sets, it achieves compression ratios between 1.2 and 4.2 as well as compression and decompression throughputs between 2.8 and 5.9 million 64-bit double-precision numbers per second on a 3 GHz Pentium 4 machine.

Execution traces, such as are used to study and analyze program behavior, are often so large that they need to be stored in compressed form. This paper describes the design and implementation of four value prediction-based compression (VPC) algorithms for traces that record the PC as well as other information about executed instructions. VPC1 directly compresses traces using value predictors, VPC2 adds a second compression stage, and VPC3 utilizes value predictors to convert traces into streams that can be compressed better and more quickly than the original traces. VPC4 introduces further algorithmic enhancements and is automatically synthesized. Of the 55 SPECcpu2000 traces we evaluate, VPC4 compresses 36 better, decompresses 26 faster, and compresses 53 faster than BZIP2, MACHE, PDATS II, SBC, and SEQUITUR. It delivers the highest geometric-mean compression rate, decompression speed, and compression speed because of the predictors' simplicity and their ability to exploit local value locality. Most other compression algorithms can only exploit global value locality.

Communication-intensive parallel applications spend a significant amount of their total execution time exchanging data between processes, which leads to poor performance in many cases. In this paper, we investigate message compression in the context of large-scale parallel message-passing systems to reduce the communication time of individual messages and to improve the bandwidth of the overall system. We implement and evaluate the cMPI message-passing library, which quickly compresses messages on-the-fly with a low enough overhead that a net execution time reduction is obtained. Our results on six large-scale benchmark applications show that their execution speed improves by up to 98% when message compression is enabled.