The IRMA Community
Newsletters
Research IRM
Click a keyword to search titles using our InfoSci-OnDemand powered search:
|
A General Framework of Algorithm-Based Fault Tolerance Technique for Computing Systems
Abstract
The Algorithm-Based Fault Tolerance (ABFT) approach transforms a system that does not tolerate a specific type of faults, called the fault-intolerant system, to a system that provides a specific level of fault tolerance, namely recovery. The ABFT philosophy leads directly to a model from which error correction can be developed. By employing an ABFT scheme with effective convolutional code, the design allows high throughput as well as high fault coverage. The ABFT techniques that detect errors rely on the comparison of parity values computed in two ways. The parallel processing of input parity values produce output parity values comparable with parity values regenerated from the original processed outputs and can apply convolutional codes for the redundancy. This method is a new approach to concurrent error correction in fault-tolerant computing systems. This chapter proposes a novel computing paradigm to provide fault tolerance for numerical algorithms. The authors also present, implement, and evaluate early detection in ABFT.
Related Content
|
Chaymaâ Boutahiri, Ayoub Nouaiti, Aziz Bouazi, Abdallah Marhraoui Hsaini.
© 2024.
14 pages.
|
|
Imane Cheikh, Khaoula Oulidi Omali, Mohammed Nabil Kabbaj, Mohammed Benbrahim.
© 2024.
30 pages.
|
|
Tahiri Omar, Herrou Brahim, Sekkat Souhail, Khadiri Hassan.
© 2024.
19 pages.
|
|
Sekkat Souhail, Ibtissam El Hassani, Anass Cherrafi.
© 2024.
14 pages.
|
|
Meryeme Bououchma, Brahim Herrou.
© 2024.
14 pages.
|
|
Touria Jdid, Idriss Chana, Aziz Bouazi, Mohammed Nabil Kabbaj, Mohammed Benbrahim.
© 2024.
16 pages.
|
|
Houda Bentarki, Abdelkader Makhoute, Tőkési Karoly.
© 2024.
10 pages.
|
|
|