Tenant Level Checkpointing of Meta-data for Multi-tenancy SaaS

Traditional checkpointing techniques are facing a grave challenge when applied to multi-tenancy software-as-a-service (SaaS) systems due to the huge scale of the system state and the diversity of users´ requirements on the quality of services. This paper proposes the notion of tenant level checkpointing and an algorithm that exploits Big Data techniques to checkpoint tenant´s meta-data, which are widely used in configuring SaaS for tenant-specific features. The paper presents a prototype implementation of the proposed technique using NoSQL database Couchbase and reports the experiments that compare it with traditional implementation of checkpointing using file systems. Experiments show that the Big Data approach has a significantly lower latency in comparison with the traditional approach.