att_abstract={{Disaster Recovery (DR) is a desirable feature for all enterprises, and a crucial one for many. However, adoption of DR remains limited due to the stark tradeoffs it imposes. To be able to recover an application to the point of crash, one is limited by financial considerations, substantial application overhead, or minimal geographical separation between the primary and recovery sites. In this paper, we argue for cloud-based DR and pipelined synchronous replication as an antidote to these problems. Cloud hosting promises economies of scale and on-demand provisioning that are a perfect fit for the infrequent yet urgent needs of DR. However the WAN latency between a cloud site and an enterprise can become a major performance bottleneck when synchronously replicating an application's data into the cloud. Pipelined synchrony addresses this problem by tracking the causal consequences of the disk modifications that are persisted to a recovery site, while allowing the application to make forward progress in its handling of client requests. In this manner, we efficiently overlap replication delay with application processing for multi-tier distributed servers, while retaining full consistency guarantees for application state in the event of a disaster. PipeCloud, our prototype, is able to sustain these guarantees for multi-node servers composed of black-box VMs, with no need of application modification, resulting in a perfect fit for the arbitrary nature of VM-based cloud hosting. Our extensive evaluation shows that PipeCloud achieves significant performance improvements over existing replication strategies, and demonstrates proper disaster failover to the Amazon EC2 platform. PipeCloud can increase throughput by an order of magnitude and reduces response times by more than half compared to synchronous replication,  all while providing the same zero data loss consistency guarantees.}},
	att_authors={hl143p, jv2782, kr2812},
	att_categories={C_NSS.3, C_NSS.4, C_NSS.11},
	att_copyright_notice={{(c) ACM, 2011. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in [ACM Symposium on Cloud Computing {{, 2011-10-27}}.
	att_tags={Disaster Recover,  Cloud Computing,  Replication,  Virtualization},
	author={Horacio Lagar-cavilla and Jacobus Van der merwe and Kadangode Ramakrishnan and Tim Wood and Prashant Shenoy},
	institution={{ACM Symposium on Cloud Computing}},
	title={{PipeCloud: Using Causality to Overcome Speed-of-Light Delays in Cloud-Based Disaster Recovery}},