Abstract
Many problems in science and engineering are usually emulated as a set of mutually interacting models, resulting in a coupled or multiphysics application. These component models show challenges originating from their interdisciplinary nature and from their computational and algorithmic complexities. In general, these models are independently developed and maintained, so that they commonly employ the global file system for exchanging their data in the coupled application.
To effectively use the local file cache on the compute node for exchanging the data among the processes of such applications, and consequently boosting I/O performance, this article presents a novel mechanism to migrate a process from one compute node to another node on the basis of block I/O dependency. In this newly proposed mechanism, the block I/O dependency between two involved processes running on the different nodes is profiled as block access similarity by taking advantage of the Cohen’s kappa statistic. Then, the process is supposed to be dynamically migrated from its source node to the destination node, on which there is another process having heavy block I/O dependency. As a result, both processes can exchange their data by utilizing the local file cache instead of the global file system to reduce I/O time. The experimental results demonstrate that the I/O performance can be significantly improved, and the time required for executing the application can be resultantly decreased, as expected.
- H. Abdi. 2007. The Kendall rank correlation coefficient. In Encyclopedia of Measurement and Statistics. Sage, Thousand Oaks, CA. 508--510.Google Scholar
- R. Ahmad, A. Gani, and S. Hamid. 2015. Virtual machine migration in cloud data centers: a review, taxonomy, and open research issues. J Supercomput. 71, 7 (2015), 2473--2515. Google Scholar
Digital Library
- Y. Amir, B. Awerbuch, and A. Barak et al. 2000. An opportunity cost approach for job assignment in a scalable computing cluster. IEEE Trans. Parallel Distrib. Syst. 11, 7 (2000), 760--768. Google Scholar
Digital Library
- V. Anthony and J. Garrett. 2005. Understanding interobserver agreement: the kappa statistic. Fam. Med. 37, 5 (2005), 360--363.Google Scholar
- K. Barker, A. Chernikov, N. Chrisochoides et al. 2004. A load balancing framework for adaptive and asynchronous applications. IEEE Trans. Parallel Distrib. Syst. 15, 2 (2004), 183--192. Google Scholar
Digital Library
- J. Benesty, J. Chen, and Y. Huang et al. 2009. Pearson correlation coefficient. In Noise Reduction in Speech Processing. Springer, Berlin, 1--4.Google Scholar
- BTIO Benchmark. 2011. Retrieved from http://www.nas.nasa.gov/.Google Scholar
- A. Choudhary. 2015. Active Storage with Analytics Capabilities and I/O Runtime System for Petascale Systems. Report No. DOE-NWU-25848. Northwestern University, Evanston, IL.Google Scholar
- I. Cores, G. Rodriguez, and P. Gonzalez et al. 2014. Failure avoidance in MPI applications using an application-level approach. Comput. J. 57, 1 (2014), 100--114.Google Scholar
Cross Ref
- X. Cui, P. Zhu, and X. Yang et al. 2014. Optimized big data K-means clustering using MapReduce. J. Supercomput. 70, 3 (2014), 1249--1259. Google Scholar
Digital Library
- M. DeGroot and M. Schervish. 2011. Probability and Statistics, 4th ed. Pearson Education Limited, London.Google Scholar
- X. Ding, S. Jiang, F. Chen, and X. Zhang et al. 2007. DiskSeen: Exploiting Disk Layout and Access History to Enhance I/O Prefetch. In Proceedings of the USENIX Annual Technical Conference (ATC’07). USENIX Association. Google Scholar
Digital Library
- J. Dongarra and P. Beckman et al. 2011. The International Exascale Software Roadmap. Int. J. High Perf. Comput. Appl. 25, 1 (2011), 3--60. Google Scholar
Digital Library
- J. Duell. 2000. The design and implementation of berkeley labs linux checkpoint/restart. Technique Report, Lawrence Berkeley National Laboratory.Google Scholar
- FUSE: Filesystem in Userspace. Retrieved from http://fuse.sourceforge.net/.Google Scholar
- B. Hunt, E. Kostelich, and I. Szunyogh. 2007. Efficient data assimilation for spatiotemporal chaos: A local ensemble transform Kalman filter. Physica D 230, 1 (2007), 112--126.Google Scholar
- K. Ibrahim, S. Hofmeyr, C. Iancu, and E. Roman. 2011. Optimized pre-copy live migration for memory intensive applications. In Proceedings of the International Conference on High Performance Computing, Network, and Storage Analysis (SC’2011). Google Scholar
Digital Library
- Iozone Filesystem Benchmark. Retrieved from http://www.iozone.org/.Google Scholar
- S. Jiang, X. Ding, Y. Xu, and K. Davis. 2013. A prefetching scheme exploiting both data layout and access history on disk. ACM Trans. Stor. 9, 3 (2013), Article 10. Google Scholar
Digital Library
- E. Jeannot, G. Mercier, and F. Tessier. 2014. Process placement in multicore clusters: Algorithmic issues and practical techniques. IEEE Trans. Parallel Distrib. Syst. 25, 4 (2014), 993--1002. Google Scholar
Digital Library
- F. Joseph, J. Cohen, and B. Everitt. 1969. Large sample standard errors of kappa and weighted kappa. Psychol. Bull. 72, 5 (1969), 323--327.Google Scholar
Cross Ref
- KDD Cup 1999 Data. Retrieved from http://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html.Google Scholar
- J. Larson, R. Jacob, and E. Ong. 2005. The model coupling toolkit: A new Fortran90 toolkit for building multiphysics parallel coupled models. Int. J. High Perform. C 19, 3 (2005), 277--292. Google Scholar
Digital Library
- Z. Li, Z. Chen, S. Srinivasan, and Y. Zhou. 2004. C-Miner: Mining Block Correlations in Storage Systems. In Proceedings of the 3rd USENIX Conference on File and Storage Technologies (ATC’04). USENIX. Google Scholar
Digital Library
- J. Liao, Y. Ishikawa. 2012a. Partial replication of metadata to achieve high metadata availability in parallel file systems. In Proceedings of 41st International Conference on Parallel Processing (ICPP’12). 168--177. Google Scholar
Digital Library
- J. Liao. 2012b. A new concurrent checkpoint mechanism for embeded multi-core systems. Comput. Inform. 31, 3 (2012), 693--709.Google Scholar
- J. Liao, F. Trahay, B. Gerofi, Y. Ishikawa. 2016. Prefetching on storage servers through mining access patterns on blocks. IEEE Trans. Parallel Distrib. Syst. 27, 9 (Sep. 2016), 2698--2710. Google Scholar
Digital Library
- J. Liao, B. Gerofi, G. Lien , S. Nishizawa, T. Miyoshi, H. Tomita, W. Liao, A. Choudhary, and Y. Ishikawa. 2017. A flexible I/O arbitration framework for netCDF based big data processing workflows on high-end supercomputers. Concurr. Comput. Pract. Exper. 29, 15 (Aug. 2017), 12 pages.Google Scholar
Cross Ref
- A. Mashtizadeh, M. Cai, and G. Tarasuk-Levin et al. 2014. XvMotion: Unified virtual machine migration over long distance. In Proceedings of the 2014 USENIX Annual Technical Conference (USENIX ATC’14). Google Scholar
Digital Library
- V. Medina and J. Garcia. 2014. A survey of migration mechanisms of virtual machines. ACM Comput. Surv. 46, 3 (2014), Article 30. Google Scholar
Digital Library
- V. Melnykov, W. Chen, and R. Maitra. 2012. Mixsim: An R package for simulating data to study performance of clustering algorithms. J. Stat. Softw. 51, 12 (2012), 1--25.Google Scholar
Cross Ref
- F. Milojicic, F. Douglis, Y. Paindaveine, and S. Zhou et al. 2000. Process migration. ACM Comput. Surv. 32, 3 (2000), 241--299. Google Scholar
Digital Library
- T. Miyoshi, G. Lien, S. Satoh, and Y. Ishikawa et al. 2016. “Big data assimilation” toward post-peta-scale severe weather prediction: An overview and progress. Proc. IEEE 104, 11 (Nov. 2016), 2155--2179.Google Scholar
Cross Ref
- F. Molteni. 2003. Atmospheric simulations using a GCM with simplified physical parametrizations. I: Model climatology and variability in multi-decadal experiments. In Climate Dynamics, Vol. 20, 175--191.Google Scholar
Cross Ref
- L. Myers and J. Sirois. 2006. Spearman correlation coefficients, differences between. Wiley StatsRef: Statistics Reference Online.Google Scholar
- X. Ouyang, S. Marcarelli, R. Rajachandrasekar, and D. Panda. 2010. RDMA-based job migration framework for MPI over infiniband. In Proceedings of the IEEE International Conference on Cluster Computing (Cluster’10). 116--125. Google Scholar
Digital Library
- X. Ouyang, R. Rajachandrasekar, X. Besseron, and D. Panda. 2011a. High performance pipelined process migration with RDMA. In Proceedings of the 11th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID’11). IEEE Computer Society, 314--323. Google Scholar
Digital Library
- X. Ouyang, R. Rajachandrasekar, X. Besseron, and D. Panda et al. 2011b. CRFS: A lightweight user-level filesystem for generic checkpoint/restart. In Proceedings of 2011 International Conference on Parallel Processing (ICPP’11). 375--384. Google Scholar
Digital Library
- S. Petri and H. Langendorfer. 1995. Load balancing and fault tolerance in workstation clusters migrating groups of communicating processes. ACM SIGOPS Operat. Syst. Rev. 29, 4 (1995), 25--36. Google Scholar
Digital Library
- E. Riedel, G. Gibson, and C. Faloutsos. 1998. Active storage for large-scale data mining and multimedia applications. In Proceedings of 24th Conference on Very Large Databases (VLDB’98). 62--73. Google Scholar
Digital Library
- S. Valcke, R. Budich, and M. Carter et al. 2006. The PRISM software framework and the OASIS coupler. In Proceedings of the 18th Annual BMRC Modelling Workshop.Google Scholar
- C. Vecchiola, S. Pandey, and R. Buyya. 2009. High-performance cloud computing: A view of scientific applications. In Proceedings of the 10th International Symposium on Pervasive Systems, Algorithms, and Networks (ISPAN’09). 4--16. Google Scholar
Digital Library
- R. Vyas, H. Maheta, and V. Dabhi et al. 2014. Load balancing using process migration for linux based distributed system. In Proceedings of International Conference on Issues and Challenges in Intelligent Computing Techniques (ICICT’14). 248--252.Google Scholar
- C. Wang, F. Mueller, C. Engelmann, and S. Scott. 2008. Proactive process-level live migration in HPC environments. In Proceedings of the International Conference on High Performance Computing, Networks, and Storage Analysis (SC’08). 1--12. Google Scholar
Digital Library
- J. Wang and X. Liang. 2005. Qualitative Data Analysis. East China Normal University Press, 92--93. {in Chinese}Google Scholar
- D. Williams, H. Jamjoom, and H. Weatherspoon. 2012. The Xen-Blanket: Virtualize once, run everywhere. In Proceedings of the 7th ACM European Conference on Computer Systems (EuroSys’12). ACM, New York, NY, 113--126. Google Scholar
Digital Library
- J. Wong. 2018. C-MapReduce. Retrieved January 2018 from https://github.com/jeffrey-garcia/C-MapReduce.Google Scholar
- Y. Xie, D. Feng, Y. Li, and D. Long. 2016. Oasis: An active storage framework for object storage platform. Future Generation Computer Systems 56 (2016), 746--758. Google Scholar
Digital Library
- F. Xu, F. Liu, L. Liu, and H. Jin et al. 2014. iaware: Making live migration of virtual machines interference-aware in the cloud. IEEE Trans. Comput. 63, 12 (2014), 3012--3025. Google Scholar
Digital Library
- F. Xu, F. Liu, L. Liu, and H. Jin et al. 2014b. Managing performance overhead of virtual machines in cloud computing: a survey, state of the art, and future directions. Proc. IEEE 102, 1 (2014), 11--31.Google Scholar
Cross Ref
- X. Zhang, K. Davis, and S. Jiang. 2011. Qos support for end users of i/o-intensive applications using shared storage systems. In Proceedings of the 2011 ACM/IEEE Conference on Supercomputing (SC’11). ACM, New York, NY. Google Scholar
Digital Library
- F. Zhang, C. Docan, M. Parashar et al. 2012. Enabling in-situ execution of coupled scientific workflow on multi-core platform. In Proceedings of IEEE 26th International Parallel 8 Distributed Processing Symposium (IPDPS’12). 1352--1363. Google Scholar
Digital Library
- F. Zheng, H. Zou, and G. Eisenhauer et al. 2013. Flexio: I/O middleware for location-flexible scientific data analytics. In Proceedings of IEEE 27th International Symposium on Parallel and Distributed Processing (IPDPS’13). 320--331. Google Scholar
Digital Library
Index Terms
Adaptive Process Migrations in Coupled Applications for Exchanging Data in Local File Cache
Recommendations
Optimizing Local File Accesses for FUSE-Based Distributed Storage
SCC '12: Proceedings of the 2012 SC Companion: High Performance Computing, Networking Storage and AnalysisModern distributed file systems can store huge amounts of information while retaining the benefits of high reliability and performance. Many of these systems are prototyped with FUSE, a popular framework for implementing user-level file systems. ...
Dynamic Process Migration Based on Block Access Patterns Occurring in Storage Servers
An emerging trend in developing large and complex applications on today’s high-performance computers is to couple independent components into a comprehensive application. The components may employ the global file system to exchange their data when ...
Application performance on the Direct Access File System
The Direct Access File System (DAFS) is a distributed file system built on top of direct-access transports (DAT). Direct-access transports are characterized by using remote direct memory access (RDMA) for data transfer and user-level networking. The ...






Comments