Erik Riedel, Garth Gibson, and Christos Faloutsos. Active storage for large-scale data mining and multimedia. In Very Large Databases (VLDB), New York, NY, August 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....by adjusting read ahead to file size, whereas HDC reduces the number of disk accesses by increasing the cache hit rate at the disk controller. Programmable disks have also been studied. However, the proposed approaches sought to reduce the amount of data transferred between disks and main memory [1, 22], to implement flexible RPC on disks [29] and or to offload file system functionality from the host processor to the disks [9] All of these works assumed significant processing capacity at the disks. Unlike these proposals, our focus in on improving disk throughput through more sophisticated ....

E. Riedel, G. A. Gibson, and C. Faloutsos. Active Storage for Large-Scale Data Mining and Multimedia. In Proceedings of the 24th International Conference on Very Large Data Bases, VLDB, pages 62--73, August 1998.

....[2] 3] 4] These studies have motivated the utility of devices with more intelligence and function. Direct network attachment of disks is being proposed to enable data transfers from devices directly to clients rather than through servers [5] 6] 7] 8] Earlier work on active disks [9] [10], 11] proposed to migrate parts of applications to storage devices resulting in filtered data. There are two main advantages of active disks. One is the parallelism available among disks. There could be more aggregate CPUpower at disks than at servers. The other is the ability to dramatically ....

....more sophisticated scheduling algorithms on storage devices to optimize the performance of different applications. 6RELATED WORK In response to the increasing storage and computational demand for applications such as decision support database, multimedia, the Active Disk, and IDisk models [9] [10], 11] have been proposed. These models propose to take advantage of the processing power on individual disks to run application level code. Analytical models and prototype simulators of active storage have been developed [22] An evaluation of the active disk model for decision support database ....

E. Riedel, G. Gibson, and C. Faloustos, "Active Storage for LargeScale Data Mining and Multimedia," Proc. 24th Very Large Databases Conf., pp. 62-73, Aug. 1998.

....resources operates on streams of data in pipelined fashion with negligible activation and post processing costs. The basic idea of the model is then to estimate the performance of the system based on the performance of its bottleneck resource. Our model is similar in flavor to that described in [10] for active disks. Table I lists the model parameters. For each application, N is defined as a di#erent function of N out , and out ; N h is a di#erent function of N , and ; Nn is a di#erent function of N out ; and finally, N d is a di#erent function of N out . In the ....

Erik Riedel, Garth A. Gibson, and Christos Faloutsos, "Active storage for large-scale data mining and multimedia," in Proceedings of the 24th International Conference on Very Large Data Bases, VLDB, August 1998, pp. 62--73.

....and disks in NASD support writes and attribute read requests. This is the price that SPIRAL pays in order to achieve client side transparency and to minimize changes to the NAD OS. Some other approaches proposed to offload computation to smart storage devices include active disks and IDISK [1, 21, 14]. The derived virtual device (DVD) model [18] proposed in the Netstation project provides a mechanism for safe shared device access in an untrusted environment by creating DVDs and managing them through a network virtual device manager. The proposed third party transfer scheme using DVDs is ....

E. Riedel, G. Gibson, and C. Faloustos. Active storage for large-scale data mining and multimedia. In Proceedings of the 24th VLDB Conference, August 1998.

....flexibility. For example, the River system [2] exploits this flexibility to provide robust stream based access to large data sets. For individual accesses, dynamic set interfaces [28] can provide similar flexibility. The approaches described here may fit under such interfaces. Active disk systems [1, 15, 24] enhance storage systems with the ability to perform some application level functions. Data filtering is one of the more valuable uses of active disk capabilities, and the project database function is a column based access pattern. Thus, such a use of active disk interfaces would allow the ....

Erik Riedel, Garth Gibson, and Christos Faloutsos. Active storage for large-scale data mining and multimedia applications. International Conference on Very Large Databases (New York, NY, 24--27 August, 1998.

....grows, designers are forced to develop techniques to leverage the negative effects of this gap. One attractive solution for this problem is to distribute intelligence to storage structures such as memories and secondary storage devices (e.g. Active Memories [4, 19, 28, 32] and Active Disks [2, 17, 23, 30]) The common goal of these techniques is to offioad parts of the processing to the off chip structures such that the amount of communication between the storage and processing elements is minimized. As the VLSI technology advances, the gap between the performance of the execution cores and the ....

Riedel, E., G. Gibson, and C. Faloutsos. Active storage for large-scale data mining and multimedia. In Proceedings of, 62-73, 1998.

.... To know that we know what we know, and that we do not know what we do not know, that is true knowledge. Confucius As microprocessors and memory chips become smaller, faster, and cheaper, embedding processing and memory in peripheral devices has become an increasingly attractive proposition [1, 19, 32, 40]. Placing processing power and memory capacity within a smart disk system allows functionality to be migrated from the file system into the disk (or RAID) thus providing a number of potential advantages over a traditional system. For example, when computation takes place near data, one can ....

....awareness. New programming environments: In contrast to integration underneath a traditional file system, other work has focused on incorporating active storage into entirely new parallel programming environments. Recent work on active disks includes that by Acharya et al. 1] Riedel et al. [32], and Amiri et al. 2] Much of this research focuses on how to partition applications across host and disk CPUs to minimize data transferred across system busses. 3 Inferring On Disk Structures: Fingerprinting the File System For a semantically smart disk to implement interesting ....

E. Riedel, G. Gibson, and C. Faloutsos. Active Storage For Large-Scale Data Mining and Multimedia. In VLDB, New York, NY, August 1998.

....the prototype and discuss the various issues in optimizing the performance of such a system in mixed work loads. Keywords Active storage, virtual block address, mixed workloads I. INTRODUCTION CTIVE storage has been proposed to migrate application specific processing to storage de vices [1] [2], 3] With the proliferation of Server Area Networks (SANs) and the growing use of network attached disks (NADs) employing active storage sys tems with NADs provides an attractive architecture that scales with the increasing size of datasets. Offloading computation closer to the data source ....

....well as the interconnect network, and better scalability. However, existing approaches either requires porting file systems onto the devices (which is difficult for NADs based a block interface, such as iSCSI) or extensive modification to the OS (particularly the file system) Previous work [1] [2], 3] 4] has provided sig nificant evidence that single applications such as Xiaonan Ma and A. L. Narasimha Reddy are wih he Department of Electrical Engineering, Texas A M UniversiW, College Station, TX 77843 3128. E maih xiaonan, reddy ee.amu.edu database decision support can benefit from ....

[Article contains additional citation context not shown here]

E. Riedel, G. Gibson, and C. Faloustos, "Active storage for large-scale data mining and multimedia," Proc. of 2Jth VLDB Conf., 1998.

....copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and or a fee. ASPLOS 02 San Jose, California Copyright 2002 ACM 1 58113 574 2 . 5.00. found in active disks [1, 18, 37], smart network interfaces [7, 16] and even intelligent memories [29, 36] Of particular importance is active storage, which has the potential to greatly improve distributed storage services, in terms of performance, functionality, and even simplicity of design. Although early research on active ....

....36] Of particular importance is active storage, which has the potential to greatly improve distributed storage services, in terms of performance, functionality, and even simplicity of design. Although early research on active storage has thoroughly demonstrated its potential performance benefits [1, 2, 18, 37], previous systems require entirely new programming environments, and some target only a limited class of applications (e.g. parallel applications and database primitives) Existing distributed file services have no clear migration path into the coming active world. In this paper, we present the ....

[Article contains additional citation context not shown here]

E. Riedel, G. A. Gibson, and C. Faloutsos. Active Storage For Large-Scale Data Mining and Multimedia. In Proc. of the 24th International Conference on Very large Databases (VLDB '98), August 1998.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos. Active storage for large-scale data mining and multimedia applications. International Conference on Very Large Databases (New York, NY, 24-27 August, 1998.

....to perform application level processing. Instead of processing low level storage protocols, these disk drives are able to execute application code, making significantly bet ter use of the storage s aggregate processing power and the interconnection network between storage devices and servers [Acharya98, Keeton98, Riedel98]. Applications can take advantage of the parallelism in large storage systems and the reduction in bandwidth possible if data is operated on at the edges of the network, before it is placed onto the expensive, shared storage interconnect. The motivation for Active Disks is illustrated in Table 1, ....

....very large SMP systems and very high bandwidth I O channels. A system with Active Disks can achieve much higher application level throughput than a system with traditional disks by offloading a significant portion of the host s work to the disks and reducing interconnect traffic. Previous work [Acharya98, Keeton98, Riedel98] has shown that similar systems can achieve performance increases across a range of applications such as data mining, image processing, and some database functions. In this paper, we focus on how Active Disks can support a real database system. We begin by explaining the modifications to a ....

[Article contains additional citation context not shown here]

Riedel, E., Gibson, G. and Faloutsos, C. "Active Storage For Large-Scale Data Mining and Multimedia" VLDB, August 1998.

No context found.

Erik Riedel, Garth Gibson, and Christos Faloutsos. Active storage for large-scale data mining and multimedia. In Very Large Databases (VLDB), New York, NY, August 1998.

No context found.

Riedel, E., Gibson, G. A., and Faloutsos, C. (1998). Active Storage For Large-Scale Data Mining and Multimedia. In Proc. of the 24th International Conference on Very Large Data Bases.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos. Active storage for large-scale data mining and multimedia. In Proceedings of VLDB, August 1998. 10

No context found.

E. Riedel, G. A. Gibson, and C. Faloutsos, "Active storage for large-scale data mining and multimedia," in Proc. 24th Intl Conf. Very Large Data Bases, pp. 62--73, Aug. 1998.

No context found.

Erik Riedel, Garth Gibson, and Christos Faloutsos. Active storage for large-scale data mining and multimedia. In Very Large Databases (VLDB), New York, NY, August 1998.

No context found.

Erik Riedel, Garth Gibson, and Christos Faloutsos. Active storage for large-scale data mining and multimedia applications. International Conference on Very Large Databases (New York, NY, 24--27 August, 1998). Published as Proceedings VLDB, pages 62--73. Morgan Kaufmann Publishers Inc., 1998.

No context found.

Erik Riedel, Garth Gibson, and Christos Faloutsos, "Active Storage For Large-Scale Data Mining and Multimedia, " in Proceedings of the 24th International Conference on Very Large Databases, pp. 62-73, August 1998.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos. Active Storage For LargeScale Data Mining and Multimedia. In Proceedings of the 24th International Conference on Very Large Databases (VLDB 24), New York, New York, August 1998.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos, "Active Storage for Large-Scale Data Mining and Multimedia Applications, " Int'l. Conf. Very Large DBs, New York, NY, Aug. 24--27, 1998, pp. 62--73.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos. Active Storage For LargeScale Data Mining and Multimedia. In Proceedings of the 24th International Conference on Very Large Databases (VLDB 24), New York, New York, August 1998.

No context found.

RIEDEL, E., GIBSON, G. A., AND FALOUTSOS, C. Active Storage For Large-Scale Data Mining and Multimedia. In Proc. of International Conference on Very Large Data Bases (1998).

No context found.

E. Riedel, C. Faloutsos, and G. Gibson. Active Storage for Large-Scale Data Mining and Multimedia Applications. In Proceedings of VLDB'98, 1998.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos. Active Storage For Large-Scale Data Mining and Multimedia. In Proceedings of the Conference on Very Large DataBases, August 1998.

No context found.

E. Riedel, G. Gibson, and C. Faloutsos. Active storage for large-scale data mining and multimedia. In VLDB, 1998.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC