On November 25 my student, Jonas Buys, is going to defend his doctoral dissertation. You are cordially invited to his on-line doctoral defense:
Voting-based approximation of dependability attributes and its application to redundancy schemata in distributed computing environments
Business- and mission-critical distributed applications are increasingly expected to exhibit highly dependable characteristics, particularly in the areas of availability and QoS-related factors such as timeliness. For this type of applications, a complete cessation or a subnormal performance of the service they provide, as well as late or invalid results, are likely to result in significant monetary penalties, environmental disaster or human injury. However, software components deployed within distributed computing systems may inherently suffer from several types of impairments, such as long response times or temporary unavailability.
Adopting classic redundancy-based fault-tolerant design patterns, such as NVP, in highly dynamic distributed computing systems does not necessarily result in the anticipated improvement in dependability. This primarily stems from the statically predefined redundancy configurations hardwired within such dependability strategies, i.e. a fixed degree of redundancy and, accordingly, an immutable selection of functionally-equivalent software components, which may negatively impact the schemes’ overall effectiveness, at least from the following two angles. Firstly, a static, context-agnostic redundancy configuration may in time lead to a more rapid exhaustion of the available redundancy and, therefore, fail to properly counterbalance any disturbances possibly affecting the operational status (context) of any of the components integrated within the dependability scheme. Secondly, the amount of redundancy, in conjunction with the voting algorithm, determines how many simultaneously failing versions the NVP composite can tolerate. A predetermined degree of redundancy is, however, cost ineffective in that it inhibits to economise on resource consumption in case the actual number of disturbances could be successfully overcome by a lesser amount of redundancy.
In this thesis, a novel dependability strategy is introduced encompassing advanced redundancy management, aiming to autonomously tune its internal redundancy configuration in function of the observed disturbances. Designed to sustain high availability and reliability, this adaptive fault-tolerant strategy may dynamically alter the amount of redundancy and the selection of functionally-equivalent resources employed within the redundancy scheme. In doing so, the algorithm relies on a number of measures designed for approximating the operational status of the redundancy configuration in terms of availability, and of individual resources in terms of reliability. Discrete-event simulation is used to analyse the effectiveness and performance of the algorithm, and to illustrate how it addresses the shortcomings commonly observed in conventional NVP approach.
#adaptive #faulttolerance #dependability #redundancy #redundancysupport #antifragile #antifragility #computationalantifragility #antifragileengineering