Font Size: a A A

Research On Optimal Design Of Distributed File System In Cabin Environment

Posted on:2020-06-16Degree:MasterType:Thesis
Country:ChinaCandidate:Z H LiFull Text:PDF
GTID:2392330596975405Subject:Systems Engineering
Abstract/Summary:PDF Full Text Request
For the C919 large passenger aircraft,the traditional on-board entertainment system uses a single point of storage,which uses a single server to store files.However,in the face of the storage capacity of storage files such as high-definition video and the scale of use of nearly two hundred users,the traditional storage structure is obviously not competent.Therefore,based on the topology of the cabin link,this paper studies the optimization design of the file system of the airborne entertainment system.In this paper,the optimization design of storage system is carried out mainly from two aspects.Firstly,this paper innovatively proposes a distributed file system construction scheme based on link structure;secondly,on the basis of distributed file system.This paper studies the decentralized distribution in depth,and propose an improved load balancing algorithm based on hot zone judgment,MTTL-K algorithm.The main research work of this paper is as follows.(1)Build the distributed file system.This paper uses distributed file system instead of traditional single point storage.GlusterFS is adopted as the distributed file system in this paper after testing the performance of the file system on the market.For distributed file system with link topology,the placement of replica nodes the placement strategy of the replica nodes will affect the system throughput rate.This paper analyzes and compares three different replica node placement strategies,and uses NS3 to simulate and select a scheme with the best performance.(2)Study the load balancing algorithm suitable for cabin.This paper proposes an MTTL-K load balancing algorithm based on hot zone judgment and DHT data migration.The algorithm combines consistent hashing and distributed hash table to configure nodes.The performance of the algorithm is optimized from the perspectives of load rate judgment,migration target node selection and migration object selection.At the same time,the algorithm considers load transfer node blocking and jitter.This paper use storage space utilization and CPU utilization to describe node load utilization.We prefer to trasnsfer high-load files to low-load target nodes.It is the core strategy of the algorithm.We need to determine the load rate of the target node after receiving the load to prevent jitter.(3)Simulation and test verify the superiority.This paper uses OPNET simulation software to simulate the MTTL-K algorithm,and compares the convergence time,system load balance and load transfer times with the original algorithm to verify the effectiveness of the MTTL-K algorithm.Build a test platform and compare the improved file system with the traditional single point storage system.Through the analysis and discussion of simulation results and test results,the availability and efficiency of the file system in this paper are verified.
Keywords/Search Tags:The cabin environment, Distributed file system, Load balancing algorithm, MTTL-K algorithm
PDF Full Text Request
Related items