南京大学学报(自然科学版) ›› 2019, Vol. 55 ›› Issue (4): 667–677.doi: 10.13232/j.cnki.jnju.2019.04.017

• • 上一篇    下一篇

RSboFMC:提高数据可用性和负载均衡性的碎片矩阵缓存策略

齐小刚1,3,强敏1(),刘立芳2,3   

  1. 1. 西安电子科技大学数学与统计学院，西安，710071
    2. 西安电子科技大学计算机科学与技术学院，西安，710071
    3. 西安电子科技大学宁波信息技术研究院，宁波，315200
  • 收稿日期:2019-02-03 出版日期:2019-07-30 发布日期:2019-07-23
  • 通讯作者: 强敏 E-mail:mqiang@stu.xidian.edu.cn
  • 基金资助:
    教育部?中国移动联合基金(MCM20170103);西安市科技创新项目(201805029YD7CG13?6)

RSboFMC：Replication strategy based on fragment matrix and cachefor improving data availability and load balance

Xiaogang Qi1,3,Min Qiang1(),Lifang Liu2,3   

  1. 1. School of Mathematics and Statistics，Xidian University，Xi’an，710071，China
    2. School of Computer Science and Technology，Xidian University，Xi’an，710071，China
    3. Xidian?Ningbo Information Technology Institute，Ningbo，315200，China
  • Received:2019-02-03 Online:2019-07-30 Published:2019-07-23
  • Contact: Min Qiang E-mail:mqiang@stu.xidian.edu.cn

RichHTML

3

PDF (PC)

960

摘要：

保证动荡环境下数据可被访问概率对数据存储网络十分重要，其可行方法之一是设计合理的存储策略，提高网络的数据可用性.将存储策略分为复制策略和放置策略进行设计，提出了基于碎片矩阵和缓存的存储策略RSboFMC(Replication Strategy based on Fragment Matrix and Cache)，提高动荡环境下的数据可用性.其以重建效率和存储开销为目标，设计缓存机制和基于碎片矩阵的数据分块机制优化复制策略；以负载均衡为目标，设计基于分区和顺逆序的分发机制优化放置策略.仿真结果表明，RSboFMC在数据可用性和负载均衡性方面均优于其他策略，且具有良好的扩展性.

关键词: 数据分块机制, 数据可用性, 缓存机制, 数据分发机制

Abstract:

It is critical for data storage network to provide expected possibility of data accessed under churn. One of possible method is to design proper storage strategy for increasing data availability of networks. RSboFMC(Replication Strategy based on Fragment Matrix and Cache) was proposed based on the fragment matrix and the cache to raise data availability under churn. It divided the storage strategy into replication and placement strategies，then the cache mechanism and the data partitioning mechanism which was based on fragment matrix were designed to optimize the replication strategy in consideration of reconstruct effectiveness and storage cost，as well as the distributing mechanism based on zone partition and opposite sequences was designed to optimize the placement strategy，considering the loadbalance. Simulation results show RSboFMC outperforms other strategies in terms of the data availability and the load balance. Besides，it has good scalability.

Key words: data partitioning mechanism, data availability, cache mechanism, data distributing mechanism

中图分类号: 

  • TP302

图1

开销与缓存值的关系"

图2

算法流程图"

图3

分发机制示意图"

图4

不同冗余因子下的数据可用性"

图5

不同攻击强度下的数据可用性"

图6

不同网络规模下的数据可用性"

图7

基于分区且考虑顺逆序和区域最大副本数的分发机制下节点负载情况"

图8

基于随机放置的分发机制下节点负载分配情况"

图9

基于顺序放置的分发机制下节点负载分配情况"

图10

不同策略实际负载和理想负载偏差值分布图"

1 ZhangX X,XuF. Survey of research on big data storage∥Proceedings of the 2013 12th Interna?tional Symposium on Distributed Computing and Applications to Business,Engineering & Science.. Kingston upon Thames,United Kingdom：IEEE,2013：76-80.
2 YoonJ W,HongT Y,ChoiJ W,et al. Evaluation of P2P and cloud computing as platform for exhaustive key search on block ciphers. Peer?to?Peer Networking and Applications,2018,11(6)：1206-1216.
3 KarafiloskiE,MishevA. Blockchain solutions for big data challenges：a literature review∥IEEE EUROCON 2017-17th International Conference on Smart Technologies. Ohrid,Macedonia：IEEE,2017：763-768.
4 ChenK,ShenH Y,SapraK,et al. A social network based reputation system for cooperative p2p file sharing. IEEE Transactions on Parallel and Distributed Systems,2015,26(8),2140-2153.
5 WaluyoA B,TaniarD,RahayuW,et al. Trustworthy data delivery in mobile p2p network. Journal of Computer and System Sciences,2017,86：33-48.
6 ParkS H,ParkJ K. IoT industry and security technology trends. International Journal of Advanced Smart Convergence,2016,5(3)：27-31.
7 GhemawatS,GobioffH,LeyngS T. The Google file system∥Proceedings of the 19th ACM Symposium on Operating Systems Principles. Bolton Landing,NY,USA：ACM,2003：29-43.
8 艾潇. 面向数据中心的容错存储和数据访问优化问题的研究. 硕士学位论文. 哈尔滨：哈尔滨工业大学,2015.
Ai X.Research on fault tolerant storage and data access optimization in data center networks. Master Dissertation. Harbin：Harbin Institute of Technology,2015.)
9 KhanO,BurnsR,PlankJ,et al. Rethinking erasure codes for cloud file systems：minimizing I/O for recovery and degraded reads∥Proceedings of the 10th USENIX Conference on File & Storage Technologies. Berkeley,CA,USA：USENIX Association,2012：14-17.
10 QinY,AiX,ChenL J,et al. Data placement strategy in data center distributed storage systems∥2016 IEEE International Conference on Communication Systems (ICCS). Shenzhen,China：IEEE,2017：1-6.
11 ZhangL W,DengY H,ZhuW H,et al. Skewly replicating hot data to construct a powerefficient storage cluster. Journal of Network and Computer Applications,2015,50：168-179.
12 ZhuB,LiH,ShumK W,et al. HFR code：a flexible replication scheme for cloud storage systems. IET Communications,2015,9(17)：2095-2100.
13 InacioE C,DantasM A R. A coarse?grained page cache aware multivariate analytical model for the storage performance of a parallel file system. Concurrency and Computation,Practice and Experience,2018,30(8)：e4389.
14 LakshmanA,MalikP. Cassandra：a decen?tralized structured storage system. ACM SIGOPS Operating Systems Review,2010,44(2)：35-40.
15 SilvaT,KamienskiC,FernandesS,et al. A flexible DHT?based directory service for informa?tion management. Peer?to?Peer Networking and Applications,2015,8(3)：512-531.
16 SitE,HaeberlenA,DabekF,et al. Proactive replication for data durability∥Proceedings of IPTPS. Santa Barbara,CA,USA：IPTPS,2006：1-6.
17 AziziS,HashemiN,KhonsariA. HHS：an efficient network topology for large?scale data centers. The Journal of Supercomputing,2016,72(3)：874-899.
18 LiZ H,GuoZ Y,YangY Y. BCCC：an expandable network for data centers. IEEE/ACM Transactions on Networking,2016,24(6)：3740-3755.
19 LeongD,DimakisA G,HoT. Distributed storage allocation problems∥Proceedings of the 2009 Workshop on Network Coding,Theory,and Applications. Lausanne,Switzerland：IEEE,2009：86-91.
20 BhattacharyaH,ChattopadhyayS,Chattopad?hyayM,et al. A novel intelligent modeling of storage and bandwidth constraints in distributed storage allocation∥Mandal J,Dutta P,Mukho?padhyay S,et al. Computational Intelligence,Communications and Business Analytics. Springer Singapore,2017：336-346.
[1] 李君科,郭兵,李明江,郭勇,周锦程,李德光. BATS:异构系统环境下的任务调度节能方法[J]. 南京大学学报(自然科学版), 2020, 56(2): 206-215.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 王　倩，聂秀山，耿蕾蕾，尹义龙. D2D通信中基于Q学习的联合资源分配与功率控制算法[J]. 南京大学学报(自然科学版), 2018, 54(6): 1183 -1192 .
[2] 孔　颉, 孙权森, 纪则轩, 刘亚洲. 基于仿射不变离散哈希的遥感图像快速目标检测新方法[J]. 南京大学学报(自然科学版), 2019, 55(1): 49 -60 .
[3] 洪佳明,黄云,刘少鹏,印鉴. 具有结果多样性的近似子图查询算法[J]. 南京大学学报(自然科学版), 2019, 55(6): 960 -972 .
[4] 郑文萍,刘韶倩,穆俊芳. 一种基于相对熵的随机游走相似性度量模型[J]. 南京大学学报(自然科学版), 2019, 55(6): 984 -999 .
[5] 王卫星,刘兆伟,石敬华. 基于时间敏感滑动窗口的CP⁃nets结构学习[J]. 南京大学学报(自然科学版), 2020, 56(2): 175 -185 .
[6] 陈俊芬,赵佳成,韩洁,翟俊海. 基于深度特征表示的Softmax聚类算法[J]. 南京大学学报(自然科学版), 2020, 56(4): 533 -540 .

版权所有 © 2021 《南京大学学报(自然科学)》

地址：江苏省南京市鼓楼区汉口路22号，南京大学学报编辑部，210093

电话：025-83592704 , 传真：025-83592704        技术支持：北京玛格泰克科技发展有限公司