Full Description

Scope

This standard specifies the functions needed to implement a fault tolerant Futurebus+TM backplane. It does not specify how to implement a fault tolerant system or how to implement a fault tolerant backplane. The functions specified are required for any system using Futurebus+ with a fault tolerant requirement. It is expected that this standard will be used in its entirety and will be cited by Futurebus+ application environment profiles requiring fault tolerance.This standard is composed of eight clauses and an annex. References required by this standard are defined in clause 2. Key words used in this standard are defined in clause 3.. Detection mechanisms (not covered by ISO/IEC 10857 :1994 and IEEE Std 896.2-1991)1 that support single-bit and single-point fault coverage are discussed in clause 4. A detailed description of error handling requirements is described in clause 5. A description of the error reports generated from the detection of an error are characterized in clause 6. Basic recovery functions (i.e., retry) are covered in clause 7. CSR definitions not covered by clause 3 of IEEE Std 896.2-1991 are specified in clause 8. The annex contains a Failure Modes and Effects Analysis (FMEA) of the signals protected by this standard.This standard is an extension of the concepts developed in ISO/IEC 10857: 1994 and IEEE Std 896.2-1991. Knowledge of those standards is necessary for a complete understanding of this standard. Because the cache coherency and message passing protocols described in ISO/IEC 10857 :1994 are built on top of the basic signaling conventions, they are beyond the scope of this standard and are not addressed by this document. However, they may be addressed by a future extension to this standard.

Purpose

The high-level goal of this standard is to ensure that Futurebus+ systems requiring fault tolerance can be implemented with minimal impact on performance and costs. Specific goals supported by this standard include the following:a) No bad data is transferred across the backplane undetected; b) At a minimum, single-bit errors are detected; c) Errors are logged and reported; d) Backplane supports system halt capability after a fault is detected. It is unrealistic to consider all possible errors for fault detection. It is, however, reasonable to consider a set of errors that covers a significant portion of the possible errors. The set containing single-bit and single-point failures is considered such a set. All single-bit and single-point failures on all signal lines should be detected and reported to all modules on the bus. Without 100% detection and reporting, it is possible for an error to corrupt state information in a module on the bus without the knowledge of the module or system. While this is not ideal, it is equally unrealistic to require the detection of 100% of all single-bit and single-point failures, just as it is unrealistic to consider detection of all possible errors. It is always possible to conceive and contrive a set of circumstances that create a window of opportunity to defeat detection mechanisms. The success of detection mechanisms relies on the fact that the probability that an error may not be apparent is acceptably small. For the purposes of this standard, this probability is acceptably small if it exceeds the probability of a double-bit error. For systems requiring better fault coverage, measures beyond the scope of this standard have to be taken. The aim of the extensions specified in this standard is to establish the low-level framework that brings the fault detection, recovery, and error reporting capabilities of Futurebus+ up to the level demanded by any system that claims fault tolerance. These extensions do not establish a complete definition of a fault tolerant environment. Higher level issues, like systems with dual-redundant buses, systems using networked back-planes to achieve fault isolation, and many other configurations, are left to the system designer or to profiles. NOTE -- These goals have been formulated cooperatively and are consistent with those required by the following Futurebus+ working groups: Military, Telecommunications, and Spaceborne.

Abstract

New IEEE Standard - Inactive-Withdrawn.This standard is one in a family of Futurebus+TM standards. The Futurebus+ standards provide a set of tools with which to implement a bus architecture with performance and cost scalability over time for multiple generations of single- and multiple-bus multiprocessor systems. This standard provides fault tolerant extensions to Futurebus+ standards. As such, this standard provides the logical layer requirements for the transmission of data in a fault tolerant environment. When used in conjunction with other IEEE standards, the details to develop modular, open-architecture-based systems fulfilling user needs across a wide computing spectrum are available.

Product Details

Published: 05/25/1995 ISBN(s): 9781559375160 Number of Pages: 98File Size: 1 file , 950 KB Product Code(s): STDWD94278