Đang chuẩn bị liên kết để tải về tài liệu:
A Shared Global Event Propagation System to Enable Next Generation Distributed Services

Building upon the FeedTree distribution system, we foresee a potential for entirely new services based on RSS which cannot be accomplished today. By using single-writer logs [18] in combination with a distributed storage mechanism such as a DHT [22, 15, 9], we can record permanently every RSS item published, allowing a distributed archival store of micronews across the In- ternet. Clients of such a system would easily be able to find out what they “missed” if they had been offline for so long that old RSS items are no longer available in any conventional, static RSS feed. Another area for future work is anonymous RSS feeds involving an anonymizing peer-to-peer routing system,. | A Shared Global Event Propagation System to Enable Next Generation Distributed Services Paul Brett Rob Knauerhase Mic Bowman Robert Adams Aroon Nataraj Jeff Sedayao Michael Spindel Intel Labs Intel Corporation Abstract The construction of highly reliable planetary-scale distributed services in the unreliable Internet environment entails significant challenges. Our research focuses on the use of loose binding among service components as a means to deploy distributed services at scale. An event-based publish subscribe messaging infrastructure is the principal means through which we implement loose binding. A unique property of the messaging infrastructure is that it is built on a collection of off-the-shelf instant messaging servers running on PlanetLab. Using this infrastructure we have successfully constructed long-running services such as a PlanetLab node status service with more than 2000 components. 1. Introduction As many can attest building a planetary-scale distributed service differs substantially from building a traditional distributed service within a data center. A planetary-scale service must be architected for reliability even though it is built on highly unreliable components. It must adapt to a rapidly changing compute and communication environment. It must provide appropriate quality of service and performance for a globally distributed client community while also scaling to accommodate highly variable workloads. Our research focuses on building planetary-scale services as a composition of many small highly replicated component services. The components of a service are loosely bound together using an eventbased publish subscribe messaging infrastructure called Planetary-Scale Event Propagation and Routing PsEPR pronounced pepper . Loose binding inverts the convention that a client initiates a connection to a service and waits for a response. Instead in the spirit of Pandemonium Selfridge59 a service publishes events to a communication channel with no