The Experts below are selected from a list of 15 Experts worldwide ranked by ideXlab platform
Jun Wang - One of the best experts on this subject based on the ideXlab platform.
-
Exploiting redundancy to construct energy-efficient, high-performance RAIDs
2005Co-Authors: Hailong Cai, Xiaoyu Yao, Jun WangAbstract:Recent studies show that Disk-based I/O subsystems account for a non-trivial portion of energy consumption in data-intensive environment such as storage servers and data centers. Previous powerefficient I/O solutions for a Single Disk Drive or mobile computers cannot be applied to data-intensive environment where the I/O load is much more intensive. Current solutions seek help from multi-speed Disks but this type of Disk will not be on the market soon. Solutions have to be sought on extensions on conventional storage system techniques such as RAIDs. This paper seeks a novel way to build Energy-Efficient, high-performance RAID system called EERAID by exploiting inherent redundant information such as mirroring and parity at the RAID controller level. As representative mirroring and parity redundant RAID systems, RAID 1 and RAID 5 are particularly studied. For RAID 1, we develop a Windows Round Robin (WRR) request scheduling policy. For RAID 5, we introduce a Transformable Read policY(TRY) to transform Disk requests from spun-down Disks to active Disks, and develop a Power and Redundant Aware Destage policy (PRAD) to defer writes to work together with TRY. Comprehensive trace-Driven simulations show that EERAID can significantly conserve energy with a minor performance impairment, by 12-30 % energy saving for both small-scale and large scale RAID level 1, and 10-20% energy saving for small-scale RAID level 5
-
eeraid energy efficient redundant and inexpensive Disk array
ACM SIGOPS European Workshop, 2004Co-Authors: Jun WangAbstract:Recent research works have been presented on conserving energy for multi-Disk systems either at a Single Disk Drive level or at a storage system level and thereby having certain limitations. This paper studies several new redundancy-based, power-aware, I/O request scheduling and cache management policies at the RAID controller level to build energy-efficient RAID systems, by exploiting the redundant information and destage issues of the array for two popular RAID levels, RAID 1 and RAID 5. For RAID 1, we develop a Windowed Round Robin (WRR) request scheduling policy; for RAID 5, we introduce a N-chance Power Aware cache replacement algorithm (NPA) for writes and a Power-Directed, Transformable (PDT) request scheduling policy for reads. Trace-Driven simulation proves EERAID saves much more energy than legacy RAIDs and existing solutions.
An Zhu - One of the best experts on this subject based on the ideXlab platform.
-
algorithms for the database layout problem
International Conference on Database Theory, 2005Co-Authors: Gagan Aggarwal, Tomas Feder, Rajeev Motwani, Rina Panigrahy, An ZhuAbstract:We present a formal analysis of the database layout problem, i.e., the problem of determining how database objects such as tables and indexes are assigned to Disk Drives. Optimizing this layout has a direct impact on the I/O performance of the entire system. The traditional approach of striping each object across all available Disk Drives is aimed at optimizing I/O parallelism; however, it is suboptimal when queries co-access two or more database objects, e.g., during a merge join of two tables, due to the increase in random Disk seeks. We adopt an existing model, which takes into account both the benefit of I/O parallelism and the overhead due to random Disk accesses, in the context of a query workload which includes co-access of database objects. The resulting optimization problem is intractable in general and we employ techniques from approximation algorithms to present provable performance guarantees. We show that while optimally exploiting I/O parallelism alone suggests uniformly striping data objects (even for heterogeneous files and Disks), optimizing random Disk access alone would assign each data object to a Single Disk Drive. This confirms the intuition that the two effects are in tension with each other. We provide approximation algorithms in an attempt to optimize the trade-off between the two effects. We show that our algorithm achieves the best possible approximation ratio.
Gagan Aggarwal - One of the best experts on this subject based on the ideXlab platform.
-
algorithms for the database layout problem
International Conference on Database Theory, 2005Co-Authors: Gagan Aggarwal, Tomas Feder, Rajeev Motwani, Rina Panigrahy, An ZhuAbstract:We present a formal analysis of the database layout problem, i.e., the problem of determining how database objects such as tables and indexes are assigned to Disk Drives. Optimizing this layout has a direct impact on the I/O performance of the entire system. The traditional approach of striping each object across all available Disk Drives is aimed at optimizing I/O parallelism; however, it is suboptimal when queries co-access two or more database objects, e.g., during a merge join of two tables, due to the increase in random Disk seeks. We adopt an existing model, which takes into account both the benefit of I/O parallelism and the overhead due to random Disk accesses, in the context of a query workload which includes co-access of database objects. The resulting optimization problem is intractable in general and we employ techniques from approximation algorithms to present provable performance guarantees. We show that while optimally exploiting I/O parallelism alone suggests uniformly striping data objects (even for heterogeneous files and Disks), optimizing random Disk access alone would assign each data object to a Single Disk Drive. This confirms the intuition that the two effects are in tension with each other. We provide approximation algorithms in an attempt to optimize the trade-off between the two effects. We show that our algorithm achieves the best possible approximation ratio.
Hailong Cai - One of the best experts on this subject based on the ideXlab platform.
-
Exploiting redundancy to construct energy-efficient, high-performance RAIDs
2005Co-Authors: Hailong Cai, Xiaoyu Yao, Jun WangAbstract:Recent studies show that Disk-based I/O subsystems account for a non-trivial portion of energy consumption in data-intensive environment such as storage servers and data centers. Previous powerefficient I/O solutions for a Single Disk Drive or mobile computers cannot be applied to data-intensive environment where the I/O load is much more intensive. Current solutions seek help from multi-speed Disks but this type of Disk will not be on the market soon. Solutions have to be sought on extensions on conventional storage system techniques such as RAIDs. This paper seeks a novel way to build Energy-Efficient, high-performance RAID system called EERAID by exploiting inherent redundant information such as mirroring and parity at the RAID controller level. As representative mirroring and parity redundant RAID systems, RAID 1 and RAID 5 are particularly studied. For RAID 1, we develop a Windows Round Robin (WRR) request scheduling policy. For RAID 5, we introduce a Transformable Read policY(TRY) to transform Disk requests from spun-down Disks to active Disks, and develop a Power and Redundant Aware Destage policy (PRAD) to defer writes to work together with TRY. Comprehensive trace-Driven simulations show that EERAID can significantly conserve energy with a minor performance impairment, by 12-30 % energy saving for both small-scale and large scale RAID level 1, and 10-20% energy saving for small-scale RAID level 5
Tomas Feder - One of the best experts on this subject based on the ideXlab platform.
-
algorithms for the database layout problem
International Conference on Database Theory, 2005Co-Authors: Gagan Aggarwal, Tomas Feder, Rajeev Motwani, Rina Panigrahy, An ZhuAbstract:We present a formal analysis of the database layout problem, i.e., the problem of determining how database objects such as tables and indexes are assigned to Disk Drives. Optimizing this layout has a direct impact on the I/O performance of the entire system. The traditional approach of striping each object across all available Disk Drives is aimed at optimizing I/O parallelism; however, it is suboptimal when queries co-access two or more database objects, e.g., during a merge join of two tables, due to the increase in random Disk seeks. We adopt an existing model, which takes into account both the benefit of I/O parallelism and the overhead due to random Disk accesses, in the context of a query workload which includes co-access of database objects. The resulting optimization problem is intractable in general and we employ techniques from approximation algorithms to present provable performance guarantees. We show that while optimally exploiting I/O parallelism alone suggests uniformly striping data objects (even for heterogeneous files and Disks), optimizing random Disk access alone would assign each data object to a Single Disk Drive. This confirms the intuition that the two effects are in tension with each other. We provide approximation algorithms in an attempt to optimize the trade-off between the two effects. We show that our algorithm achieves the best possible approximation ratio.