Reducing Synchronization Cost in Distributed Multi-resource Allocation Problem

Abstract

Generalized distributed mutual exclusion algorithms allow processes to concurrently access a set of shared resources. However, they must ensure an exclusive access to each resource. In order to avoid deadlocks, many of them are based on the strong assumption of a prior knowledge about conflicts between processes' requests. Some other approaches, which do not require such a knowledge, exploit broadcast mechanisms or a global lock, degrading message complexity and synchronization cost. We propose in this paper a new solution for shared resources allocation which reduces the communication between non-conflicting processes without a prior knowledge of processes conflicts. Performance evaluation results show that our solution improves resource use rate by a factor up to 20 compared to a global lock based algorithm.

A Resource Usage Efficient Distributed Allocation Algorithm for 5G Service Function Chains

“…To the best of our knowledge this problem has not been specifically addressed. Some papers address the drinking philosophers problems and mentioned possible extension of their algorithms to systems with N instances [2,15] but did not consider the selection of the instances as a specific constraint.…”

“…Several variants of the problem have been defined considering a single resource [14,27] or multiple instances of a single type of resource [25,26]. Very few solutions have been proposed for the problem of Mutual Exclusion for systems with multiple instances of multiple resources [2,15].…”

“…We show that for systems with multiple instances of resources the selection of the instances to allocate has an influence on performances. Our algorithm extends the LASS algorithm [15] for systems with N instances of M resources. We introduce a subroutine to select the instance of the resource based on the orders of the requests as the LASS algorithm does not address this constraint.…”

Towards multi-SDN services: Dangers of concurrent resource allocation from multiple providers

“…Towards a solution to this problem, we will study in more depth in which situations these deadlocks can occur in SDN and NFV environment. In a short term perspective, we plan to apply the algorithm proposed by Lejeune et al in 2015 [? ] while taking into account the challenges that may arise such as the latency caused by the algorithm, the QoS to be fulfilled to the users as well as the adaptation of the algorithm to SDN and NFV environment.…”

A Fault-Tolerant Algorithm For Distributed Resource Allocation

“…Existe una diferencia entre ambos y es que los algoritmos de exclusión mutua requieren la especificación de la identidad del recurso a ser asignado, mientras que en la asignación distribuida cualquier instancia de un recurso puede asignarse sin especificar su identidad. Al problema de asignación de múltiples instancia de un recurso se lo conoce como exclusión mutua h-de-k [10,11,12,13,14,15] en donde existen múltiples instancias k de un mismo recurso y a cada nodo se le puede asignar hasta un número h de esas instancias (1≤h≤k). Algunos de estos algoritmos no prevén la consideración de aspectos que son fundamentales para implementaciones prácticas.…”