Optimization of checkpointing/recovery strategy in cloud computing with adaptive storage management

Abstract

Summary Cloud Computing is a type of distributed system that is usually based on the services offered to the user based on SLA contract. In this case, the implementation of a fault‐tolerant system that ensures the reliability and the services continuity becomes a major requirement. In this paper, we propose a fault tolerance strategy based on checkpointing and replication. Our approach uses a smart checkpoint infrastructure for cloud computing tasks. The checkpoints are stored in alternative already paid VMs. This allows resuming a task execution faster and cheaper after a node crash. Since checkpoints are distributed and replicated, our approach increases also the system reliability. The experimental results show the effectiveness of the proposed strategy in term of energy consumption, SLA (System Level Aggregation) violation, and reliability.

Intelligent Fault-Tolerant Mechanism for Data Centers of Cloud Infrastructure

“…e integer linear programing method is used here to handle VM replica placement. In [10], to increase the reliability of the system, a checkpointing/restart mechanism was proposed along with a replication scheme.…”

Distributed Application Checkpointing for Replicated State Machines

“…Another area of literature, regarding checkpointing in the cloud, consists recovery processes of distributed running tasks. A recent study proposes a system in this aspects and evaluates over energy consumption, service level agreement violations, recovery time of tasks and failure rates [17]. A very recent study also reports storage checkpoint recovery times for bag-of-tasks jobs over Amazon Web Services [18].…”

Avaliação dos Serviços de Armazenamento da Amazon Web Services para Gravação e Recuperação de Checkpoints

“…Neste trabalho estamos interessados na avaliação de desempenho da gravação e recuperação de checkpoints no contexto dos serviços de armazenamentos oferecidos pela AWS. Porém, a maioria dos trabalhos relacionados não apresentam resultados práticos e avaliações destes serviços [Nicolae and Cappello 2011, Zhou et al 2017, Meroufel and Belalem 2018, Yaothanee and Chanchio 2019. Além disso, esses trabalhos se concentram na proposta de abordagens de armazenamento de checkpoints e não na avaliação dos serviços.…”

“…Assim como em [Nicolae and Cappello 2011], os autores assumem um nível de controle disponível apenas para o provedor. Já em [Meroufel and Belalem 2018], uma solução de armazenamento acessível ao usuárioé proposta. Os autores apresentam uma estratégia para workflows, no qual os checkpoints são replicados em outras MVs.…”