A Survey on Multithreading Alternatives for Soft Error Fault Tolerance

Abstract

Smaller transistor sizes and reduction in voltage levels in modern microprocessors induce higher soft error rates. This trend makes reliability a primary design constraint for computer systems. Redundant multithreading (RMT) makes use of parallelism in modern systems by employing thread-level time redundancy for fault detection and recovery. RMT can detect faults by running identical copies of the program as separate threads in parallel execution units with identical inputs and comparing their outputs. In this article, we present a survey of RMT implementations at different architectural levels with several design considerations. We explain the implementations in seminal papers and their extensions and discuss the design choices employed by the techniques. We review both hardware and software approaches by presenting the main characteristics and analyze the studies with different design choices regarding their strengths and weaknesses. We also present a classification to help potential users find a suitable method for their requirement and to guide researchers planning to work on this area by providing insights into the future trend.

Multi-core Devices for Safety-critical Systems

“…But, the addition of these diagnosis techniques in the hardware, can also have an impact on silicon die and lead to cost increase, power overheads and reduced performance. So, the combination of hardware and software techniques combined at different abstraction levels are usually required to detect complex hardware faults such as MBU and transient faults, as supported hardware techniques alone might not be sufficient [34,39,56,48,36]. However, cross-requirements research contributions that aim at reconciling reliability, DC and time predictability are still scarce (e.g., [57,58]).…”

“…Therefore, devices targeting industrial markets with high reliability requirements are prone to include such features as oppose to high volume markets that have lower reliability requirements [38]. Because of this, application specific or generic software fault-tolerance techniques can be used in order to improve hardware reliability, e.g., multithreading for soft error fault-tolerance [56], algorithmic based fault tolerance [46], software redundant execution (e.g., software application, replicated instructions and check inserted by the compiler) [46], combination of software and hardware techniques with WCET guarantees [138], software level thermal co-management [36,137], application specific techniques (e.g., [4,6]).…”

Dynamic Distributed Based on PLC Software Redundancy

“…With its own advantages, the DDS has gradually developed into the preferred storage system for various types of data storage today [2]. When the dynamic distributed storage system faces the loss of disk data, it is necessary to introduce a redundant fault-tolerant mechanism to ensure that the system can work normally [3]. In order to provide users with reliable file storage services, dynamic distributed storage systems need to adopt fault-tolerant technology to improve file availability [4][5].…”

Multi-Threaded Mitigation of Radiation-Induced Soft Errors in Bare-Metal Embedded Systems

“…The extra process-ing units of these systems can be used to accelerate software computation strategies aimed at mitigating or detecting soft errors, nevertheless, the majority of the existing software protection techniques are not designed to take advantage of these resources. Recent works, such as [8], point to distribute the replica computation across different processing units as a way to gain efficiency without losing reliability. Simultaneous Muti-Threading (SMT) is a long ago defined concept that has been used along with the concept of Sphere-of-Replication (SoR) to achieve reliability [9].…”