ABSTRACT
Geographically distributed cloud data centers are increasingly favored by internet companies and governments, attributed to their superior storage and computational capabilities. However, while data redundancy techniques, especially erasure coding, play a foundational role in ensuring data reliability, the profound threat of natural disasters challenges the resilience of such centers. This paper investigates the placement of erasure-coded data considering disaster risks to enhance the reliability of cloud storage systems in distributed scenarios.
We introduce a disaster-aware approach for erasure-coded data placement, aiming to avoid areas susceptible to natural disasters, thus reducing potential damages. Central to this approach is a multi-objective optimization model specifically tailored for the nuances of erasure-coded data. It counters three main threats: the direct damage from disasters compromising data reliability; the potential decline in overall reliability due to uneven disaster exposures across a data set; and data unavailability from communication disruptions between erasure-coded blocks. To obtain an optimized solution, we employ the NSGA-II algorithm. The experimental results demonstrate that, under the circumstances of natural disasters, our method ensures greater reliability and availability of erasure-coded data compared to the random placement approach that disregards disaster risks and another method we introduced for comparison, which only optimizes for a single objective while considering disaster risks.
To our best knowledge, this study is the first exploration in ensuring the reliable placement of erasure-coded data within distributed cloud data centers under disaster risk.
- Zhang, Q., Cheng, G. H., & Boutaba, R. 2010. Cloud computing: state-of-the-art and research challenges. Journal of Internet Services and Applications, 1(1), 7-18. DOI: 10.1007/s13174-010-0007-6.Google ScholarCross Ref
- Buyya, R., Ranjan, R., & Calheiros, R. N. 2010. InterCloud: Utility-oriented federation of cloud computing environments for scaling of application services. In: Proceedings of the 10th international conference on algorithms and architectures for parallel processing. Springer, Berlin, Heidelberg, 13-31. DOI: 10.1007/978-3-642-13119-6_2.Google ScholarDigital Library
- Sabahi, F. 2012. Cloud Data Center Security. In: 2012 IEEE/ACIS 11th International Conference on Computer and Information Science. IEEE, 54-59.Google Scholar
- Ghemawat, S., Gobioff, H., & Leung, S. T. 2003. The Google File System. ACM SIGOPS Operating Systems Review, 37(5), 29-43.Google ScholarDigital Library
- Weatherspoon, H., & Kubiatowicz, J. D. 2002. Erasure Coding vs. Replication: A Quantitative Comparison. In: Proceedings of the First International Workshop on Peer-to-Peer Systems (IPTPS '02), 328-338.Google Scholar
- Li, X., Wang, H., Yi, S., Zhao, D., Huang, G., & Jiang, C. 2019. Disaster-and-evacuation-aware backup datacenter placement based on multi-objective optimization. IEEE Access, 7, 48196-48208.Google ScholarCross Ref
- Ma, L., Jang, X., Wu, B., Tang, J., Cui, Y., & Dong, S. 2016. Probabilistic region failure-aware data center network and content placement. Computer Networks, 103(C), 56-66.Google Scholar
- Mukherjee, B., Dikbiyik, F., Habib, N. F., Tornatore, M., & Mukherjee, B. 2015. Disaster-aware datacenter placement and dynamic content usage in cloud networks. IEEE/OSA Journal of Optical Communications and Networking, 7(7), 681-694.Google ScholarCross Ref
- Couto, R. S., Secci, S., Campista, M. E. M., Duarte, O. C., & Pujolle, G. 2015. Server placement with shared backups for resilient clouds. Computer Networks, 93, 423-434.Google ScholarDigital Library
- Zhou, A., Yi, B., & Luo, L. 2022. Tree-structured data placement scheme with cluster-aided top-down transmission in erasure-coded distributed storage systems. Computer Networks, 204, 108714.Google ScholarDigital Library
- Yu, B., & Pan, J. 2020. A Framework of Hypergraph-Based Data Placement Among Geo-Distributed Datacenters. IEEE Transactions on Services Computing, 13(3), 395-409.Google ScholarCross Ref
- Jin, H., Luo, R., He, Q., Wu, S., Zeng, Z., & Xia, X. 2023. Cost-Effective Data Placement in Edge Storage Systems with Erasure Code. IEEE Trans. Serv. Comput., 16(2), 1039-1050. doi: 10.1109/TSC.2022.3152849.Google ScholarCross Ref
- Wang, X., Jiang, X., & Pattavina, A. 2011. Assessing network vulnerability under probabilistic region failure model. In: IEEE International Conference on High Performance Switching & Routing. IEEE.Google ScholarCross Ref
- InternetMCI Network. 2011. http://www.topology-zoo.org/dataset.htmlGoogle Scholar
- National Geophysical Data Center / World Data Service (NGDC/WDS): NCEI/WDS Global Significant Earthquake Database. Accessed 2023. NOAA National Centers for Environmental Information. doi:10.7289/V5TD9V7K.Google ScholarCross Ref
- NOAA National Centers for Environmental Information. Storm Events Database. Accessed 2023. http://www.ncdc.noaa.gov/stormevents/.Google Scholar
Index Terms
- Reliable Placement of Erasure-coded Data in Distributed Cloud Data Centers under Disaster Risks: A Multi-Objective Optimization Approach
Recommendations
An evolutionary multi-objective optimization approach to disaster waste management
A framework to determine the locations of temporary storage sites is proposed.A sustainable solution methodology for disaster waste management is provided.The framework not only considers the cost but also the side effects on human beings.Utilization of ...
Cross-Rack-Aware Updates in Erasure-Coded Data Centers
ICPP '18: Proceedings of the 47th International Conference on Parallel ProcessingThe update performance in erasure-coded data centers is often bottlenecked by the constrained cross-rack bandwidth. We propose CAU, a cross-rack-aware update mechanism that aims to mitigate the cross-rack update traffic in erasure-coded data centers. ...
A "hitchhiker's" guide to fast and efficient data reconstruction in erasure-coded data centers
SIGCOMM'14Erasure codes such as Reed-Solomon (RS) codes are being extensively deployed in data centers since they offer significantly higher reliability than data replication methods at much lower storage overheads. These codes however mandate much higher ...
Comments