A Generic Fault-Tolerant Architecture for Real-Time by David Powell

By David Powell

The layout of computers to be embedded in serious real-time purposes is a fancy activity. Such structures must never purely warrantly to satisfy tough real-time time cut-off dates imposed by means of their actual surroundings, they have to warrantly to take action dependably, regardless of either actual faults (in undefined) and layout faults (in or software). A fault-tolerance strategy is vital for those promises to be commensurate with the security and reliability necessities of many existence- and mission-critical functions. This booklet explains the motivations and the result of a collaborative project', whose aim was once to noticeably lessen the lifecycle expenses of such fault­ tolerant structures. The end-user businesses engaging during this venture already install fault-tolerant structures in serious railway, house and nuclear-propulsion purposes. besides the fact that, those are proprietary structures whose architectures were adapted to satisfy domain-specific specifications. This has ended in very high priced, rigid, and sometimes hardware-intensive suggestions that, by the point they're built, established and licensed to be used within the box, can already be out-of-date by way of their underlying and software program technology.

Show description

Read or Download A Generic Fault-Tolerant Architecture for Real-Time Dependable Systems PDF

Best quality control books


During the global, milk and milk items are integral parts of the nutrients chain. not just do person shoppers use liquid milk for drinks and cooking, yet meals brands use massive amounts of milk powder, centred milks, butter, and cream as uncooked fabrics for extra processing.

Improving Efficiency by Shrinkage (Statistics: A Series of Textbooks and Monographs)

Deals a remedy of other forms of James-Stein and ridge regression estimators from a frequentist and Bayesian standpoint. The e-book explains and compares estimators analytically in addition to numerically and contains Mathematica and Maple courses utilized in numerical comparability. ;College or college bookshops may possibly order 5 or extra copies at a different pupil cost, to be had on request.

Antifreeze Solutions in Home Fire Sprinkler Systems

Antifreeze strategies in domestic hearth Sprinkler structures examines the use of antifreeze options in residential sprinklers, and analyzes their effectiveness in controlling a fireplace and supporting in containment. The booklet additionally investigates the potential of a large-scale ignition taking place from suggestions of various combinations, and proposes the optimum ones for decreasing flammability.

Extra info for A Generic Fault-Tolerant Architecture for Real-Time Dependable Systems

Sample text

This implies in particular that they can tolerate a Byzantine faulty clock in a three-node system. In these algorithms [Srikanth & Toueg 1987, Dolev et al. 1995], each node periodically seeks to be the system synchroniser, and succeeds if it is the fastest non-faulty node. More precisely, each node broadcasts its resynchronisation message and collects during a given waiting period the resynchronisation messages broadcast by other nodes (like in convergence-averaging algorithms). As soon as it has received m + 1 resynchronisation messages (including possibly its own), it restarts its local clock for the next period and relays the m + 1 resynchronisation messages to all other nodes.

If ~ is negative, then the local node is late. l = T + R . 4 Initial Synchronisation Two solutions can be distinguished to perform initial synchronisation of a set of nodes: • Start all nodes independently. Initial synchronisation algorithms exist that can synchronise such a set of non-synchronised active nodes (even in the This status will be subsequently used to effectively identify the faulty node. We focus here just on the synchronisation algorithm, without taking into account the pool reconfiguration mechanism.

This implies in particular that they cannot tolerate a Byzantine faulty clock in a three-node system. The basic idea of these algorithms is as follows. The resynchronisation is performed periodically: each node broadcasts a resynchronisation message when its local clock time has counted R seconds since the last resynchronisation period. At the same time, the node collects during a given waiting period (which is far shorter than R) the resynchronisation messages broadcast by other nodes: it records the arrival time (according to its local clock time) of each received resynchronisation message.

Download PDF sample

Rated 4.10 of 5 – based on 11 votes