On the average file unavailability for specific storage disk arrangements in Cloud systems

• Autorzy: Besson M. , Tanguy Ch. , Grall A.
• Języki publikacji: EN

Abstrakty

EN

Cloud computing is a growing field since data storage is becoming ever more decentralized. Providers of Cloud solutions want to insure the safety and availability of their customers’ data. In order to increase these performance indices, several storage policies have been implemented: replication, erasure codes, etc. A few of them rely on randomized procedures. In this paper, we focus on the influence of a specific storage policy on the availability of a given file. Taking only disk failures into account, we provide a general formula for the average file unavailability U ̅, which is a generalization of the well-known k-out-of-n problem, to which it reduces when disks are identical. We then calculate U ̅ for several configurations when disks have different reliabilities, and show that the disk arrangement has a major impact on the result. We also provide an approximation which could be helpful for more complex arrangements.

Słowa kluczowe

EN

cloud; storage system design; k-out-of-n systems

• Wydawca: Polskie Towarzystwo Bezpieczeństwa i Niezawodności
• Czasopismo: Journal of Polish Safety and Reliability Association
• Rocznik: 2015
• Tom: Vol. 6, No. 1
• Strony: 23--30
• Opis fizyczny: Bibliogr. 24 poz., rys., wykr.

Twórcy

autor

Besson M.

• Orange Labs, Issy-les-Moulineaux, France

autor

Tanguy Ch.

• Orange Labs, Issy-les-Moulineaux, France

autor

Grall A.

• LM2S, Troyes, France

Bibliografia

• [1] Bhagwat, D., Pollack, K., Long, D.D.E., Schwarz, T., Miller, E. L. & Pâris, J.F. (2006). Providing High Reliability in a Minimum Redundancy Archival Storage System. Proc. of the 14th IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS 2006), 413-421.
• [2] Dai, Y.S., Yang, B., Dongarra, J. & Zhang, G. (2009). Cloud Service Reliability: Modeling and Analysis. Proc. of the 15th IEEE Pacific Rim International Symposium on Dependable Computing.
• [3] Dhillon, B.S. (2005). Reliability, Quality, and Safety for Engineers. CRC Press, 2000 N.W. Corporate Blvd., Boca Raton, Florida.
• [4] Druault-Vicard, A. & Tanguy, C. (2008). Exact Failure Frequency Calculations for Extended Systems. http://arxiv.org/abs/cs/0612141. (unpublished).
• [5] Elerath, J.G. & Pecht, M. (2007). Enhanced Reliability Modeling of RAID Storage Systems. Proc. of the 37th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN'07) pp. 175-184.
• [6] Greenan, K., Plank, J.S. & Wylie, J.J. (2010). Mean time to meaningless: MTTDL, Markov models, and storage system reliability. In Proc. of the Second Workshop on Hot Topics in Storage and File Systems.
• [7] Hafner, J.L. & Rao, K.K. (2006). Notes on Reliability Models for Non-MDS Erasure Codes. IBM Research Division, Technical Report RJ10391 (A0610-035), October 2006.
• [8] Huang C., Simitci H., Xu Y., Ogus A., Calder B., Gopalan P., Li J & Yekhanin S. (2012). Erasure Coding in Windows Azure Storage. Proc. of the 2012 USENIX Annual Technical Conference (ATC'12).
• [9] Iliadis, I. & Venkatesan, V. (2014). Expected Annual Fraction of Data Loss as a Metric for Data Storage Reliability. Proc. of the 22nd International Symposium on Modelling, Analysis Simulation of Computer and Telecommunication Systems (MASCOTS 2014) pp. 375-384.
• [10] Jiang, W., Hu, C., Zhou, Y. & Kanevsky, A. (2008). Are Disks the Dominant Contributor for Storage Failures? A Comprehensive Study of Storage Subsystem Failure Characteristics. ACM Transactions on Storage 4, 3, Article 7.
• [11] Kuo, W. & Zuo, M. J. (2003). Optimal Reliability Modeling. John Wiley & Sons, Inc., Hoboken, New Jersey.
• [12] Leslie, M., Davies, J. & Huffman, T. (2006). A Comparison of Replication Strategies for Reliable Decentralised Storage. Journal of Networks, 1, 6, 36-44.
• [13] Lian, Q., Chen, W. & Zhang, Z. (2005). On the Impact of Replica Placement to the Reliability of Distributed Brick Storage Systems. Proc. of the 25th IEEE International Conference on Distributed Computing Systems (ICDCS 2005), 187-196.
• [14] Pinheiro, E., Weber, W.-D. & Barroso, L. A. (2007). Failure Trends in a Large Disk Drive Population. Proceedings of the USENIX Conference on File and Storage Technologies, 17-28.
• [15] Rao, K.K., Hafner, J.L. & Golding, R.A. (2011). Reliability for Networked Storage Nodes. IEEE Transactions on Dependable and Secure Computing, 8, 3, 404-418.
• [16] Rausand, M. & Høyland, A. (2004). System Reliability Theory: Models, Statistical Methods, and Applications. John Wiley & Sons, Inc., Hoboken, New Jersey.
• [17] Rosenthal, D.S.H. (2010). Bit Preservation: A Solved Problem? International Journal of Digital Curation 5(1), 134-148.
• [18] Schroeder, B. & Gibson, G.A. (2006). Understanding disk failure rates: What does an MTTF of 1,000,000 hours mean to you? ACM Transactions on Storage, 3, 3, Article 8.
• [19] Sun, D.W., Chang, G.R., Gao, S., Jin, L.Z. & Wang, X.W. (2012). Modeling a Dynamic Data Replication Strategy to Increase System Availability in Cloud Computing Environments. Journal of Computer Science and Technology 27, 2, 256-272.
• [20] Tanguy, C., Besson, M., Krishnaswamy, R. & Grall, A. (2015). On data unavailability and file loss in coded data storage systems for the Cloud, submitted to ESREL 2015.
• [21] Thomasian, A. & Blaum, M. (2012). Mirrored Disk Organization Reliability Analysis. IEEE Transactions on Computers 55, 12, 1640-1644.
• [22] Venkatesan, V. & Iliadis, I. (2012). A General Reliability Model for Data Storage Systems. Proc. of the Ninth International Conference on Quantitative Evaluation of Systems (QEST 2012), 209-219.
• [23] Venkatesan, V. & Iliadis, I. (2013). Effect of Codeword Placement on the Reliability of Erasure Coded Data Storage Systems. In Quantitative Evaluation of Systems, Lecture Notes in Computer Science, vol. 8054, editors K. Joshi, M. Siegle, M. Stoelinga & P. R. D’Argenio, Springer: 241-257.
• [24] Venkatesan, V., Iliadis, I., Fragouli C. & Urbanke, R. (2011). Reliability of Clustered vs. Declustered Replica Placement in Data Storage Systems. Proc. of the 19th Annual IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems, 307-317.