Benchmarking Eventually Consistent Distributed Storage Systems
Type of Event:
Cloud storage services and NoSQL systems, which have recently found widespread adoption, typically offer only "Eventual Consistency", a rather weak guarantee covering a broad range of consistency behavior. The degree of actual (in-)consistency as a service quality, however, is unknown. As a consequence, the responsibilities imposed on application developers using an eventually consistent storage system are un-known. Furthermore, comparing the diverse cloud storage offerings and NoSQL systems against each other is difficult.
We argue for "Consistency Assessment" via benchmarking and simulation to address this problem. In this talk, we present a novel approach using consistency benchmarking, modeling and simulation as a means to predict the consistency guarantees of a distributed storage system. We also show, using a set of concrete examples, how the results can be used to efficiently handle inconsistencies within applications or in
an additional middleware layer. Using our approach, consistency as a service quality is made transparent and becomes measurable, thereby providing important insight for any application that uses an eventually consistent storage system.
Start: 2013-12-18 at 5:30 pm
End: 2013-12-18 at 7:00 pm
Building: 11.40, Room: 226
iCal event: (iCal)
Host: Research group Economics and Technology of eOrganizations