A scalability hierarchical fault tolerance strategy: Community Fault Tolerance
Jianping Chen, Yao Lu, Ioan Sorin Comsa, Pierre Kuonen
Abstract
Jianping Chen, Yao Lu, Ioan Sorin Comsa, Pierre Kuonen
Abstract
Most of hierarchical fault tolerance strategies did not pay much attention to scalability of fault tolerance. In distributed system, scalability is a very important feature. To tolerant failures when the scale of the system changing is a normal and important scenario. Especially in nowadays, almost all the cloud computing companies provide their computing services elastically. To add extra devices or remove devices in order to provide different services happens all the time. In such a scenario, it is very important that the fault tolerance strategy is scalable. In this paper, we introduce dynamic programming thoughts to build hierarchical regions as communities for fault tolerance strategy and apply different strategies based on communities instead of a single process. We call this fault tolerance strategy as Community Fault Tolerance. It cannot only reduce the memory overload by eliminating the number of records of messages inside the community region, but also provides a good characteristic of scalability. The scalability property of our strategy makes it handle with the scenario of adding devices or removing devices in the distributed system easily.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Most of hierarchical fault tolerance strategies did not pay much attention to scalability of fault tolerance. In distributed system, scalability is a very important feature. To tolerant failures when the scale of the system changing is a normal and important scenario. Especially in nowadays, almost all the cloud computing companies provide their computing services elastically. To add extra devices or remove devices in order to provide different services happens all the time. In such a scenario, it is very important that the fault tolerance strategy is scalable. In this paper, we introduce dynamic programming thoughts to build hierarchical regions as communities for fault tolerance strategy and apply different strategies based on communities instead of a single process. We call this fault tolerance strategy as Community Fault Tolerance. It cannot only reduce the memory overload by eliminating the number of records of messages inside the community region, but also provides a good characteristic of scalability. The scalability property of our strategy makes it handle with the scenario of adding devices or removing devices in the distributed system easily.
Key concepts: Scalability, Fault tolerance, Computer science, Distributed computing, Software fault tolerance, Cloud computing, Computer network, Database