A distributed storage system optimization method based on a partition processing consensus algorithm
A distributed storage and partition processing technology, applied in the direction of input/output to record carrier, etc., can solve the problem of performance degradation of distributed storage system, and achieve the effect of optimal delay and throughput performance
Active Publication Date: 2019-06-28
广州市凯蒙科技有限公司
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] Aiming at the above problems, the method of the present invention aims to optimize the performance of the distributed storage system, and proposes a distributed storage system optimization method based on the partition processing consensus algorithm to solve the problem that the performance of the distributed storage system decreases with the increase of client command conflicts It can improve the delay and throughput performance of the system when the client command conflict affects the performance of the distributed storage system, and can better adapt to different client environments and meet the needs of practical applications.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0036] 1. Introduction to basic theory
[0037] 1. EPaxos algorithm
[0038] EPaxos is designed with high throughput in the cluster and low latency in the wide area network environment. In a small number of copies (F, Among them, N is the total number of replicas) In the case of a failure, the reliability of the overall system can still be guaranteed.
[0039] figure 2 Indicates the specific processing flow of the EPaxos algorithm: the client generally submits commands to the nearest replica RL (to reduce delay). The replica RL receives the client's command, and Replicas for a round of message exchange (including RL itself, figure 2 Middle FastPath stage). like replicas disagree about some information in the client command (that is, the client command conflicts), and the replica RL will conduct another round of message exchange with F+1 replicas ( figure 2 in the SlowPath stage). Afterwards, the RL can reply to the client and notify the rest of the replicas that ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a distributed storage system optimization method based on a partition processing consensus algorithm, belongs to the field of distributed system performance optimization, and mainly solves the problem that the performance of an existing distributed storage system is reduced along with the increase of client command conflicts. The optimization method monitors the processingcondition of a client command in real time, and judges whether partition processing is needed or not according to the processing condition of the client command. Under the condition that partition processing is not needed, each copy in the distributed storage system processes a client command by adopting an EPaxos consensus algorithm; The method comprises the following steps: firstly generating apartitioning scheme for a condition needing partitioning processing, and then coordinating a processing flow of each copy in the distributed storage system according to a specific partitioning schemeto perform partitioning processing. Under the condition that the performance of the distributed storage system is influenced by the client command conflict, the delay and throughput performance of thesystem can be improved, meanwhile, different client environments can be better adapted, and the requirements of practical application are met.
Description
technical field [0001] The invention belongs to the field of distributed system performance optimization, and relates to a distributed storage system performance optimization method, specifically a distributed storage system optimization method based on a partition processing consensus algorithm, which can be used for distributed system performance optimization. Background technique [0002] With the rapid development of the network, the scale of applications is gradually expanding, and the amount of data continues to grow...People gradually realize the indispensable role of distributed systems in engineering practice. As a kind of distributed system, distributed storage system is the basis of large-scale applications. Companies such as Google, Amazon, Yahoo, and Alibaba all have their own distributed storage systems. [0003] Unlike traditional storage systems that store data centrally, distributed storage systems store data scattered on different physical devices (copies)...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & AuthorityApplications(China)
IPC IPC(8): G06F3/06
Inventor葛洪伟赵守月杨金龙江明
Owner广州市凯蒙科技有限公司