A Comparison of Redundancy Techniques for Private and Hybrid Cloud Storage

 

E.M. Hernandez-Ramirez*¹, V.J. Sosa-Sosa², I. Lopez-Arevalo³

 

^1,2,3 Laboratorio de Tecnologías de Información (Campus Tamaulipas) del Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV) Cd. Victoria, Tamaulipas, México. *emhr1983@gmail.com.

 

ABSTRACT

File redundancy techniques have been very useful mechanisms for offering fault tolerance and data availability in any kind of storage. Cloud storage is not the exception. This paper presents an evaluation of classical file redundancy techniques implemented in two cloud-storage deployment models, private and hybrid. A small prototype of a private and hybrid cloud storage was implemented for this evaluation. The performance impact when file redundancy is only applied in a private cloud versus when redundancy is also distributed in a public cloud (the hybrid model) is analyzed. Additional to classical file redundancy techniques, an innovative method was evaluated for file redundancy based on an information dispersal algorithm (IDA). The usage of IDA represents a good option for managing sensitive data in hybrid cloud storage. In this technique, only parts of a file need to be sent to the public cloud, avoiding the complete file to be read from outside of the private zone. In this context, there is a trade-off between performance (for reconstructing the original file, it is first necessary to obtain all of its fragments) and the security level that could determine the viability of using IDA.

Keywords: cloud computing, data management, virtualization.

 

RESUMEN

Las técnicas de redundancia de archivos han sido mecanismos muy útiles para ofrecer tolerancia a fallos y disponibilidad de datos en cualquier tipo de almacenamiento. El almacenamiento en la nube no es la excepción. Este trabajo presenta una evaluación del comportamiento de técnicas clásicas de redundancia de archivos, implementadas en una nube de almacenamiento utilizando dos modelos de despliegue, privado e híbrido. Para esta evaluación se desarrolló un prototipo de un sistema de almacenamiento en la nube que sigue los modelos de despliegue antes mencionados. Se compara el impacto en el rendimiento cuando se aplica redundancia sólo en una nube privada versus cuando la redundancia también se distribuye en una nube pública (modelo híbrido). Adicionalmente, se evaluó un método innovador para la redundancia en archivos basado en el Algoritmo de Dispersión de Información (IDA). El uso del IDA surge como una buena opción para la administración de datos sensibles en una nube de almacenamiento híbrida. Con esta técnica de replicación, sólo fragmentos de un archivo serán enviados a la nube pública (infraestructura de terceros), evitando que el archivo completo pueda ser leído desde afuera de la zona privada. En este contexto, existe un compromiso entre el desempeño (para reconstruir el archivo original es necesario obtener primero sus fragmentos) y el nivel de seguridad que determinará la viabilidad de usar el IDA.

 

DESCARGAR ARTÍCULO EN FORMATO PDF

 

References

[1] John F. Gantz et al, "The Expanding Digital Universe: A Forecast of Worldwide Information Growth Through 2010", An IDC White Paper -sponsored by EMC (online). Available from: http://www.emc.com/collateral/analystreports/expanding-digital-idc-white-paper.pdf        [ Links ]

[2] M. R. Palankar et al, "Amazon S3 for science grids: a viable solution?" DADC08, Proceedings of the 2008 international workshop on Data-aware distributed computing, pp. 55-64. June 2008.         [ Links ]

[3] Rackspace Cloud Files (online). Available from: http://www.rackspace.com/cloud/cloudhostingproducts/files        [ Links ]

[4] Jeffrey Dean, "Evolution and future directions of large-scale storage and computation systems at Google", SoCC10, Proceedings of the 1st ACM symposium on Cloud computing, pp. 1-1 2010.         [ Links ]

[5] Irfan Habib, "Virtualization with KVM", Linux Journal. Volume 2008 Issue 166, Article No. 8. Feb. 2008.         [ Links ]

[6] P. Barham et al, "Xen and the art of virtualization", SOSP 2003 Proceedings of the 19th ACM symposium on Operating systems principles, pp. 164-177. 2003.         [ Links ]

[7] B. Sotomayor et al, "Virtual Infrastructure Management in Private and Hybrid Clouds", IEEE Internet Computing, vol. 13, no. 5, pp. 14-22, Sep./Oct. 2009.         [ Links ]

[8] Bill Childers, "Build your own cloud with Eucalyptus", Linux Journal, volume 2010 issue 195, Article No. 1. July 2010.         [ Links ]

[9] Openstack homepage (online). Available from: http://www.openstack.org/        [ Links ]

[10] Comparative of XEN and KVM (online). Available from: http://virt.kernelnewbies.org/XenVsKVM        [ Links ]

[11] M. O. Rabin, "Efficient dispersal of information for security, load balancing, and fault tolerance", J. ACM 36, 2 (April 1989), pp 335-348.         [ Links ]

[12] Dropbox (online). Available from: http://www.dropbox.com/features        [ Links ]

[13] J. L. González and R. Marcelín Jiménez. "Phoenix: Fault Tolerant Distributed Web Storage". FTRA, Journal of Convergence of the Future Technology Research Association, vol 2, no.1. 2011 pp. 79-86.         [ Links ]

[14] Autobench (online). Available from: http://www.xenoclast.org/autobench        [ Links ]

[15] Abu-Libdeh, H et al., RACS: a case for cloud storage diversity, Proceedings of the 1st ACM Symposium on Cloud Computing. Pp. 229-240. June 2010.         [ Links ]

[16] Bowers K. D. et al., "HAIL: a high-availability and integrity layer for cloud storage", Proceedings of the 16th ACM Conference on Computer and Communications Security. pp. 187-198. November 2009.         [ Links ]

[17] Wala, Secure online storage (online). Available from: http://www.wuala.com        [ Links ]

[18] ADrive, Web storage (online). Available from: http://www.adrive.com        [ Links ]

[19] M. Quezada Naquid et al, "Fault Tolerance and Load Balance Tradeoff in a Distributed Storage System", Computación y Sistemas, vol. 14, no. 2, October-December 2010.         [ Links ]

[20] S. Siva Sathya et al, "Replication Strategies for Data Grids", ADCOM 2006, Proceedings on Advanced Computing and Communications, pp. 123-128. Dec. 2006.         [ Links ]