面向云应用的存储缓存子系统

计算机科学 ›› 2014, Vol. 41 ›› Issue (4): 195-199.

面向云应用的存储缓存子系统

蔡涛,牛德姣,张永春,倪晓蓉,周东明

江苏大学计算机科学与通信工程学院镇江212013;江苏大学计算机科学与通信工程学院镇江212013;江苏大学计算机科学与通信工程学院镇江212013;江苏大学计算机科学与通信工程学院镇江212013;江苏大学计算机科学与通信工程学院镇江212013

出版日期:2018-11-14 发布日期:2018-11-14
基金资助:
本文受高等学校博士学科点专项科研基金(20093227110005),广东省自然科学基金(S2011010006118),江苏省高校自然科学基金(09KJB520001),浙江省自然科学基金(LY13F020012)资助

Storage Caching Sub-system for Cloud Application

CAI Tao,NIU De-jiao,ZHANG Yong-chun,NI Xiao-rong and ZHOU Dong-ming

Online:2018-11-14 Published:2018-11-14

摘要/Abstract

摘要： 性能是云应用的重要指标,云应用与现有应用在数据存储与管理方式和访问模式等方面存在较大差异,这使得传统的缓存管理算法难以适应云应用的要求。针对云应用的特性,设计了基于影响因子的缓存策略,亦即将元数据和数据缓存分开管理,使用影响因子综合管理影响缓存被再次访问的几率的多种因素,区别创建、打开、读取和修改等操作对缓存再次访问的几率的影响；设计了缓存关联管理策略,即利用元数据和数据之间的关联提高缓存管理的性能；设计了缓存主动调度策略,即通过主动淘汰较低影响因子的缓存项和动态调整元数据与数据缓存的大小来提高缓存子系统的适应能力和性能。最后实现了原型系统,并使用Filebench和Postmark进行了测试和分析,验证了面向云应用缓存子系统原型能提高1%～120%的I/O性能以及2%～87%的操作处理速度。

关键词: 缓存,云应用,文件系统

Abstract: Performance is very important for the cloud application．There are some differences between cloud and current application in management strategy and access pattern,which leads to the traditional cache management strategy can’t adapt to the requirement of the cloud application．According to the features of cloud application,the cache strategy based on impact factors was designed to implement the storage cache sub-system for cloud application．Metadata and data cache were used to manage separately,making a distinction about the impact of the creation,opening,reading,writing and modifying on the probability of the cache accessed again．Then the two cache correlation strategics were designed to improve the performance of cache management according to the relation between metadata and data．The actively sche-duling strategy was used to improve the adaptability based on eliminating the low impact factor cache entries actively and adjusting the space of metadata cache and data cache dynamically．At last,the prototype was released and was evaluated by Filenench and Postmark．The result shows the storage cache sub-system can improve 1%~120% I/O performance and 2%～87% operation processing speed.

Key words: Cache,Cloud application,File system

蔡涛,牛德姣,张永春,倪晓蓉,周东明. 面向云应用的存储缓存子系统[J]. 计算机科学, 2014, 41(4): 195-199. https://doi.org/

CAI Tao,NIU De-jiao,ZHANG Yong-chun,NI Xiao-rong and ZHOU Dong-ming. Storage Caching Sub-system for Cloud Application[J]. Computer Science, 2014, 41(4): 195-199. https://doi.org/

参考文献

[1] Zhou Yuan,Chen Z,Li K．Second-level buffer cache management[J]．Parallel and Distributed Systems,IEEE Transactions,2004,15(6):505-519
[2] Batsakis A,Burns R．NFS-CD:Write-Enabled cooperative cac-hing in NFS[J]．Parallel and Distributed Systems,IEEE Tran-sactions,2008,19(3):323-333
[3] Zhu Yi-feng,Hong Jiang．Race:a robust adaptive caching strategy for buffer cache[J]．Computers,IEEE Transactions,2008,57(1):25-40
[4] Saila F,Garcia Blas J,Carretro J,et al．Design and evaluation of Multiple-Level data staging for blue gene systems[J]．Parallel and Distributed Systems,IEEE Transactions,2011,22(6):946-959
[5] Prabhakar R,Srikantaiah S,Garg R,et al．Adaptive QOS decomposition and control for storage cache management in Multi-server environments[C]∥Cluster,Cloud and Grid Computing (CCGrid),201111th IEEE/ACM International Symposium．Newport Beach,CA,2011:402-413
[6] Liao Wei-keng,Ching A,Coloma K,et al．An implementationand evaluation of Client-Side file caching for MPI-IO[C]∥Parallel and Distributed Processing Symposium,2007．IPDPS 2007．IEEE International．Long Beach,CA,2007:1-10
[7] Eshel M,Haskin R L,Hildebrand D,et al．Panache:a parallelfile system cache for global file access[C]∥The 8th USENIX Conference on Files and Storage Technologies．2010:55-168
[8] Nukarapu D T,Bin Tang,Wang Li-qiang,et al．Data replication in data intensive scientific applications with performance guarantee[J]．Parallel and Distributed Systems,IEEE Transactions,2011,22(8):1299-1306
[9] Sung Hoon Baek K P,Sung Hoon B,Kyu Ho park．Prefetching with adaptive cache culling for striped disk[C]∥2008USENIX Annual Technical Conference．Boston,MA,USA,2008:363-376
[10] Jiang S,Zhang Xue-chen,Shuang Liang,et al．Improving networked file system performance using a Locality-Aware cooperative cache protocol[J]．Computers,IEEE Transactions,2010,59(11):1508-1519
[11] Wei Qi-ying,Qin Ting-ting,Fujita S．A Two-Level caching protocol for hierarchical Peer-to-Peer file sharing systems[C]∥Parallel and Distributed Processing with Applications (ISPA),2011IEEE 9th International Symposium．Busan,2011:195-200
[12] Anderson E,Hoover C,Li Xiao-zhou．New algorithms for file system cooperative caching[C]∥Modeling,Analysis Simulation of Computer and Telecommunication Systems (MASCOTS),2010IEEE International Symposium．Miami Beach,FL,2010:437-440
[13] Butt A R,Gniady C,Hu Y C．The performance impact of kernelprefetching on buffer cache replacement algorithms[J]．Compu-ters,IEEE Transactions,2007,56(7):889-908
[14] Spillane R P,Dixit S,Archak S,et al．Exporting kernel page caching for efficient user-level I/O[C]∥Mass Storage Systems and Technologies (MSST),2010IEEE 26th Symposium．Incline Village,NV,2010:1-13
[15] Yue Jian-hui,Zhu Yi-feng,Zhao Cai,et al．Energy efficient buffer cache replacement for data servers[C]∥Networking,Architecture and Storage (NAS),20116th IEEE International Conference．Dalian,Liaoning,2011:329-338
[16] Schoeberl M．A Time-Predictable object cache[C]∥Object/Component/Service-Oriented Real-Time Distributed Computing (ISORC),201114th IEEE International Symposium．2011:99-105
[17] Li Chong-min,Wang Hai-xia,Xue Yi-bo,et al．Fast hierarchical cache directory:a scalable cache organization for Large-Scale CMP[C]∥Networking,Architecture and Storage (NAS),2010IEEE Fifth International Conference．2010:367-376
[18] Park S O,Kim S J．An efficient array file system for multipleSmall-Capacity NAND flash memories[C]∥Network-Based Information Systems (NBiS),201114th International Conference．2011:569-572
[19] Yang Liu,Huang Jian-zhong,Xie Chang-sheng,et al．Raf:a random access first cache management to improve SSD-Based disk cache[C]∥Networking,Architecture and Storage (NAS),2010IEEE Fifth International Conference．2010:492-500

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed