Encyclopedia of Database Systems

Living Edition
| Editors: Ling Liu, M. Tamer Özsu

Active Storage

  • Kazuo GodaEmail author
Living reference work entry
DOI: https://doi.org/10.1007/978-1-4899-7993-3_1326-2

Keywords

Disk Drive Data Traffic Host Computer Significant Performance Improvement Active Storage 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Synonyms

Definition

Active Storage is a computer system architecture which utilizes processing power in disk drives to execute application code. Active Storage was introduced in separate academic papers [1, 2, 3] in 1998. The term Active Storage is sometimes identified merely with the computer systems proposed in these papers. Two synonyms, Active Disk and Intelligent Disk, are also used to refer to Active Storage. The basic idea behind Active Storage is to offload computation and data traffic from host computers to the disk drives themselves such that the system can achieve significant performance improvements for data intensive applications such as decision support systems and multimedia applications.

Key Points

A research group at Carnegie Mellon University proposed, in [3], a storage device called Active Disk, which has the capability of downloading application-level code and running it on a processor embedded on the device. Active Disk has a performance advantage for I/O bound scans, since processor-per-disk processing can potentially reduce data traffic on interconnects to host computers and yield great parallelism of scans. E. Riedel et al. carefully studied the potential benefits of using Active Disks for four types of data intensive applications, and introduced analytical performance models for comparing traditional server systems and Active Disks. They also prototyped ten Active Disks, each having a DEC Alpha processor and two Seagate disk drives, and demonstrated almost linear scalability in the experiments.

A research group at University of California at Berkeley discussed a vision of Intelligent Disks (IDISKs) in [2]. The approach of Intelligent Disk is similar to that of Active Disk. Keeton et al. carefully studied the weaknesses of shared-nothing clusters of workstations and then explored the possibility of replacing the cluster nodes with Intelligent Disks for large-scale decision support applications. Intelligent Disks assumed higher complexity of applications and hardware resources in comparison with CMU’s Active Disks.

Another Active Disk was presented by a research group at the University of California at Santa Barbara and University of Maryland in [1]. Acharya et al. carefully studied programming models to exploit disk-embedded processors efficiently and safely and proposed algorithms for typical data intensive operations such as selection and external sorting, which were validated by simulation experiments.

These three works are often recognized as opening the gate for new researches of Intelligent Storage Systems in the post-“database machines” era.

Cross-References

Recommended Reading

  1. 1.
    Acharya A, Mustafa U, Saltz JH. Active disks: programming model, algorithms and evaluation. In: Proceedings of the 8th International Conference Architectural Support for Programming Languages and Operating System; 1998. p. 81–91.Google Scholar
  2. 2.
    Keeton K, Patterson DA, Hellerstein JM. A case for intelligent disks (IDISKs). SIGMOD Rec. 1998;27(3):42–52.CrossRefGoogle Scholar
  3. 3.
    Riedel E, Gibson GA, Faloutsos C. Active storage for large-scale data mining and multimedia. In: Proceedings of the 24th International Conference on Very Large Data Bases; 1998. p. 62–73.Google Scholar

Copyright information

© Springer Science+Business Media LLC 2016

Authors and Affiliations

  1. 1.The University of TokyoTokyoJapan

Section editors and affiliations

  • Masaru Kitsuregawa
    • 1
  1. 1.Inst. of Industrial ScienceUniv. of TokyoTokyoJapan