Many service applications use actors as a programming model for the middle tier, to simplify synchronization, fault-tolerance, and scalability. However, efficient operation of such actors in multiple, geographically distant datacenters is challenging, due to the very high communication latency. Caching and replication are essential to hide latency and exploit locality; but it is not a priori clear how to combine these techniques with the actor programming model.
We present Geo, an open-source geo-distributed actor system that improves performance by caching actor states in one or more datacenters, yet guarantees the existence of a single latest version by virtue of a distributed cache coherence protocol. Geo's programming model supports both volatile and persistent actors, and supports updates with a choice of linearizable and eventual consistency. Our evaluation on several workloads shows substantial performance benefits, and confirms the advantage of supporting both replicated and single-instance coherence protocols as configuration choices. For example, replication can provide fast, always-available reads and updates globally, while batching of linearizable storage accesses at a single location can boost the throughput of an order processing workload by 7x.
Fri 27 Oct Times are displayed in time zone: (GMT-07:00) Tijuana, Baja California change
|13:30 - 13:52|
Philip A. BernsteinMicrosoft Research, Sebastian BurckhardtMicrosoft Research, Sergey BykovMicrosoft, n.n., Natacha CrooksUniversity of Texas at Austin, USA, Jose FaleiroYale University, USA, Gabriel KliotGoogle, n.n., Alok KumbhareMicrosoft Research, n.n., Muntasir Raihan RahmanMicrosoft, Vivek ShahUniversity of Copenhagen, Denmark, Adriana SzekeresUniversity of Washington, USA, Jorgen ThelinMicrosoft Research, RedmondDOI
|13:52 - 14:15|
Oded PadonTel Aviv University, Giuliano LosaUniversity of California at Los Angeles, USA, Mooly SagivTel Aviv University, Sharon ShohamTel Aviv universityDOI
|14:15 - 14:37|
Victor B. F. GomesUniversity of Cambridge, UK, Martin KleppmannUniversity of Cambridge, Dominic P. MulliganUniversity of Cambridge, Alastair R. BeresfordUniversity of Cambridge, UKDOI
|14:37 - 15:00|
Alexander Bakst, Klaus v. GleissenthallUniversity of California at San Diego, USA, Ranjit JhalaUniversity of California at San Diego, USA, Rami Gökhan KıcıUniversity of California at San Diego, USADOI