Abstract
With big data analysis and cloud storage improving at an astonishing rate in recent years, more and more parallel computing and data analysis techniques are invented and applied to accessing data from a distributed storage system, which has noticeable differences with accessing data from a separate dedicated storage system. The first and the last processes of a data analysis task in a shared filesystem may have a remarkable gap in latency. The latency of parallel file requests from these processes, however, lacks analytical model especially when erasure code is employed as storage strategy. The main contribution of this work is development of a tight bound of parallel file request latency to give a theoretical upper bound of lantency. Additionally, an exact batch file chunk request analytical model in distributed systems adopting erasure code is studied by the technique of two embedded Markov chains, performance measures such as queue length distribution, waiting time distribution, etc. are derived in closed form. Experimental evaluation verifies accuracy of the theoretical upper bound on latency.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Aggarwal, V., Fan, J., Lan, T.: Taming tail latency for erasure-coded, distributee storage systems. In: IEEE INFOCOM 2017-IEEE Conference on Computer Communications. pp. 1–9. IEEE (2017)
Aggarwal, V., Lan, T.: Tail index for a distributed storage system with pareto file size distribution. arXiv preprint arXiv:1607.06044 (2016)
Bertsimas, D., Natarajan, K., Teo, C.P.: Tight bounds on expected order statistics. Probab. Eng. Inf. Sci. 20(4), 667–686 (2006)
Cannataro, M., Talia, D., Srimani, P.K.: Parallel data intensive computing in scientific and commercial applications. Parallel Comput. 28(5), 673–704 (2002)
Chao, L., Li, C., Liang, F., Lu, X., Xu, Z.: Accelerating apache hive with MPI for data warehouse systems. In: 2015 IEEE 35th International Conference on Distributed Computing Systems, pp. 664–673. IEEE (2015)
Chen, S., et al.: When queueing meets coding: optimal-latency data retrieving scheme in storage clouds. In: IEEE INFOCOM 2014-IEEE Conference on Computer Communications, pp. 1042–1050. IEEE (2014)
Dao, T.C., Chiba, S.: HPC-reuse: efficient process creation for running MPI and hadoop MapReduce on supercomputers. In: 2016 16th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid), pp. 342–345. IEEE (2016)
Ghemawat, S., Gobioff, H., Leung, S.T.: The google file system (2003)
hadoop.apache.org: https://hadoop.apache.org/docs/r3.0.3/
https://stats.stackexchange.com/users/10479/yves, Y.: Variance of arrival process with shifted exponential distribution. Cross Validated https://stats.stackexchange.com/q/87287. Accessed 29 Dec 2014
Huang, C., et al.: Erasure coding in windows azure storage. In: Presented as Part of the 2012 Annual Technical Conference, pp. 15–26 (2012)
Huang, D., et al.: Achieving load balance for parallel data access on distributed file systems. IEEE Trans. Comput. 67(3), 388–402 (2018)
Huang, L., Pawar, S., Zhang, H., Ramchandran, K.: Codes can reduce queueing delay in data centers. In: 2012 IEEE International Symposium on Information Theory Proceedings, pp. 2766–2770. IEEE (2012)
Joshi, G., Liu, Y., Soljanin, E.: On the delay-storage trade-off in content download from coded distributed storage systems. IEEE J. Sel. Areas Commun. 32(5), 989–997 (2014)
Kumar, V., Grama, A., Gupta, A., Karypis, G.: Introduction to Parallel Computing: Design and Analysis of Algorithms Benjamin. Cummings, Redwood City (1994)
tahoe lafs.org: Tahoe-lafs docs, January 2019. https://tahoe-lafs.readthedocs.io/en/tahoe-lafs-1.12.1
Li, W.: An investigation into batch renewal process and batch markovian arrival process and batch markovian arrival process and their performance impact on queueing models. Ph.D. thesis, University of Bradford, UK (2007)
Schurman, E., Brutlag, J.: The user and business impact of server delays, additional bytes, and http chunking in web search. In: Velocity Web Performance and Operations Conference (2009)
Su, Y., Feng, D., Hua, Y., Shi, Z.: Predicting response latency percentiles for cloud object storage systems. In: 2017 46th International Conference On Parallel Processing (ICPP), pp. 241–250. Proceedings of the International Conference on Parallel Processing (2017)
Weatherspoon, H., Kubiatowicz, J.D.: Erasure coding vs. replication: a quantitative comparison. In: Druschel, P., Kaashoek, F., Rowstron, A. (eds.) IPTPS 2002. LNCS, vol. 2429, pp. 328–337. Springer, Heidelberg (2002). https://doi.org/10.1007/3-540-45748-8_31
Xiang, Y., Lan, T., Aggarwal, V., Chen, Y.F.R.: Multi-tenant latency optimization in erasure-coded storage with differentiated services. In: 2015 IEEE 35th International Conference on Distributed Computing Systems, pp. 790–791. IEEE (2015)
Xiang, Y., et al.: Joint latency and cost optimization for erasure-coded data center storage. IEEE/ACM Trans. Netw. (TON) 24(4), 2443–2457 (2016)
Yu, X., Li, W.: Performance modelling and analysis of MapReduce/Hadoop workloads. In: The 21st IEEE International Workshop on Local and Metropolitan Area Networks, pp. 1–6. IEEE (2015)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Zou, X., Li, W. (2020). Tight Bound of Parallel Request Latency for Erasure-Coded Distributed Storage System. In: Qiu, M. (eds) Algorithms and Architectures for Parallel Processing. ICA3PP 2020. Lecture Notes in Computer Science(), vol 12452. Springer, Cham. https://doi.org/10.1007/978-3-030-60245-1_25
Download citation
DOI: https://doi.org/10.1007/978-3-030-60245-1_25
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-60244-4
Online ISBN: 978-3-030-60245-1
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)