Abstract
Big data processing and analysis increasingly rely on workflow technologies for knowledge discovery and scientific innovation. The execution of big data workflows is now commonly supported on reliable and scalable data storage and computing platforms such as Hadoop. There are a variety of factors affecting workflow performance across multiple layers of big data systems, including the inherent properties (such as scale and topology) of the workflow, the parallel computing engine it runs on, the resource manager that orchestrates distributed resources, the file system that stores data, as well as the parameter setting of each layer. Optimizing workflow performance is challenging because the compound effects of the aforementioned layers are complex and opaque to end users. Generally, tuning their parameters requires an in-depth understanding of big data systems, and the default settings do not always yield optimal performance. We propose a profiling-based cross-layer coupled design framework to determine the best parameter setting for each layer in the entire technology stack to optimize workflow performance. To tackle the large parameter space, we reduce the number of experiments needed for profiling with two approaches: i) identify a subset of critical parameters with the most significant influence through feature selection; and ii) minimize the search process within the value range of each critical parameter using stochastic approximation. Experimental results show that the proposed optimization framework provides the most suitable parameter settings for a given workflow to achieve the best performance. This profiling-based method could be used by end users and service providers to configure and execute large-scale workflows in complex big data systems.
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
Notes
- 1.
A module is a processing unit, executed in serial or parallel, in a workflow, and is also referred to as a job or subtask in some context.
References
Zaharia, P.W.M., Xin, R.S., et al.: Apache spark: a unified engine for big data processing. Commun. ACM 59(11), 56–65 (2016)
Oinn, J.F.T., Addis, M., et al.: Taverna: a tool for the composition and enactment of bioinformatics workflows. Bioinformatics 20, 3045–3054 (2004)
Ludascher, I.A.C.B.B., et al.: Scientific workflow management and the Kepler system. Spec. Issue Workflow Grid Syst. 18, 1039–1065 (2005)
Deelman, E., Blythe, J., et al.: Pegasus: mapping scientific workflows onto the grid. In: Dikaiakos, M.D. (ed.) AxGrids 2004. LNCS, vol. 3165, pp. 11–20. Springer, Heidelberg (2004). https://doi.org/10.1007/978-3-540-28642-4_2
Kumar, G.M.V.S., Sadayappan, P., et al.: An integrated framework for performance-based optimization of scientific workflows. In: Proceedings of the 18th ACM International Symposium on High Performance Distributed Computing, Garching, Germany, pp. 177–186 (2009)
Chiu, G.A.D., Deshpande, S., et al.: Cost and accuracy sensitive dynamic workflow composition over grid environments. In: Proceedings of the 2008 9th IEEE/ACM International Conference on Grid Computing, Washington, DC, USA, pp. 9–16 (2008)
Holl, M.P.S., Zimmermann, O., et al.: A new optimization phase for scientific workflow management systems. Future Gener. Comput. Sci. 36, 352–362 (2014)
Counaris, A., Torres, J.: A methodology for spark parameter tuning. Big Data Res. 11, 22–32 (2018)
Wang, B.H.G., Xu, J., et al.: A novel method for tuning configuration parameters of spark based on machine learning. In: IEEE 18th International Conference on High Performance Computing and Communications, Sydney, NSW, Austrilia (2016)
Liao, G., Datta, K., Willke, T.L.: Gunther: search-based auto-tuning of mapreduce. In: Wolf, F., Mohr, B., an Mey, D. (eds.) Euro-Par 2013. LNCS, vol. 8097, pp. 406–419. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-40047-6_42
Wu, A.G.D., et al.: A self-tuning system based on application profiling and performance analysis for optimizing hadoop mapreduce cluster configuration. In: 20th Annual International Conference on High Performance Computing, Bangalore, India (2014)
Li, S.M., Zeng, L., et al.: MRONLINE: MapReduce online performance tuning. In: Proceedings of the 23rd International Symposium on High-Performance Parallel and Distributed Computing, New York, NY, USA, pp. 165–176 (2014)
Shu, T., Wu, C.: Performance optimization of \(\mathit{H}\)adoop workflows in public clouds through adaptive task partitioning. In: Proceedings of the IEEE INFOCOM, Atlanta, GA, USA, 1–4 May 2017
Wu, C., Lin, X., Yu, D., Xu, W., Li, L.: End-to-end delay minimization for scientific workflows in clouds under budget constraint. IEEE Trans. Cloud Comp. 3(2), 169–181 (2015)
Yun, D., Wu, C., Gu, Y.: An integrated approach to workflow mapping and task scheduling for delay minimization in distributed environments. JPDC 84, 51–64 (2015)
Ye, Q., Wu, C.Q., Cao, H., et al.: Storage-aware task scheduling for performance optimization of big data workflows. In: The 8th IEEE International Conference on Big Data and Cloud Computing, Melbourne, Australia, 11–13 December 2018
Wang, B.H.E.G. , Xu, J.: A novel method for tuning configuration parameters of spark based on machine learning. In: 2016 IEEE 18th International Conference on HPC and Communications, Sydney, NSW, Australia, 12–14 December 2016
Petridis, P., Gounaris, A., Torres, J.: Spark parameter tuning via trial-and-error. In: Angelov, P., Manolopoulos, Y., Iliadis, L., Roy, A., Vellasco, M. (eds.) INNS 2016. AISC, vol. 529, pp. 226–237. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-47898-2_24
Gounaris, A., Torres, J.: A methodology for spark parameter tuning. Big Data Res. 11, 22–32 (2018)
Jia, G.C.E.Z., Xue, C.: Auto-tuning spark big data workloads on POWER8: prediction-based dynamic SMT threading. In: 2016 International Conference on Parallel Architecture and Compilation Techniques (PACT), Haifa, Israel, 11–15 September 2016
Holmes, A.: Hadoop in Practice. Manning Publications Co., Greenwich (2012)
Li, S.M.E.M., Zeng, L.: MRONLINE: MapReduce online performance tuning. In: Proceedings of the 23rd International Symposium on High-Performance Parallel and Distributed Computing, Vancouver, BC, Canada, 23–27 June 2014
Ding, D.Q.E.X., Liu, Y.: Jellyfish: online performance tuning with adaptive configuration and elastic container in hadoop yarn. In: 2015 IEEE 21st International Conference on Parallel and Distributed Systems (ICPADS), Melbourne, Australia, 14–17 December 2015
Flight Data. http://stat-computing.org/dataexpo/2009/the-data.html
Library Checkout Data. https://data.seattle.gov/Community/Checkouts-by-Title/tmmm-ytt6
Parking Violation Data. https://data.cityofnewyork.us/City-Government/Open-Parking-and-Camera-Violations/nc67-uf89
Service Request Data. https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9
Spall, J.C.: Multivariate stochastic approximation using a simultaneous perturbation gradient approximation. IEEE Trans. Autom. Control 37, 332–341 (1992)
Kiefer, J.W.J.: Stochastic estimation of the maximum of a regression function. Ann. Math. Stat. 23(3), 462–466 (1952)
Spall, J.C.: Introduction to Stochastic Search and Optimization: Estimation, Simulation, and Control. Wiley, Hoboken (2005)
Ross, B.: Mutual information between discrete and continuous data sets. PLOS ONE 9(2), 1–5 (2014)
Abramowitz, M., Stegun, I.: Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables. Dover Publishing Inc., New York (1972)
Spall, J.C.: Implementation of the simultaneous perturbation algorithm for stochastic optimization. IEEE Trans. Aerosp. Electron. Syst. 34, 817–823 (1998)
Heger, D.: Hadoop performance tuning-a pragmatic & iterative approach. CMG J. 4, 97–113 (2013)
White, T.: Hadoop: The Definitive Guide. O’Reilly Media Inc., Sebastopol (2012)
Lawler, G., Limic, V.: Random Walk: A Modern Introduction. Cambridge University Press, Cambridge (2010)
Glover, F.: Tabu search: a tutorial. Informs J. Appl. Anal. 20(4), 1–185 (1990)
Montgomery, E.A.P.D.C., Vining, G.: Introduction To Linear Regression Analysis, vol. 821. Wiley, Hoboken (2012)
Nocedal, J., Wright, S.: Numerical Optimization. Springer, Heidelberg (2006)
Apache, Hadoop (2016). http://hadoop.apache.org
Spark (2016). http://spark.apache.org
Oozie (2016). https://oozie.apache.org
Acknowledgments
This research is sponsored by U.S. National Science Foundation under Grant No. CNS-1828123 with New Jersey Institute of Technology.
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
Ye, Q., Wu, C.Q., Liu, W., Hou, A., Shen, W. (2020). Profiling-Based Big Data Workflow Optimization in a Cross-layer Coupled Design Framework. In: Qiu, M. (eds) Algorithms and Architectures for Parallel Processing. ICA3PP 2020. Lecture Notes in Computer Science(), vol 12454. Springer, Cham. https://doi.org/10.1007/978-3-030-60248-2_14
Download citation
DOI: https://doi.org/10.1007/978-3-030-60248-2_14
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-60247-5
Online ISBN: 978-3-030-60248-2
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)