Flexible, wide-area storage for distributed systems using semantic cues

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2009.This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Cataloged from student submitted PDF version of thesis.Includes bibliographical references (p. 81-87).There is a growing set of Internet-based services that are too big, or too important, to run at a single site. Examples include Web services for e-mail, video and image hosting, and social networking. Splitting such services over multiple sites can increase capacity, improve fault tolerance, and reduce network delays to clients. These services often need storage infrastructure to share data among the sites. This dissertation explores the use of a new file system (WheelFS) specifically designed to be the storage infrastructure for wide-area distributed services. WheelFS allows applications to adjust the semantics of their data via semantic cues, which provide application control over consistency, failure handling, and file and replica placement. This dissertation describes a particular set of semantic cues that reflect the specific challenges that storing data over the wide-area network entails: high-latency and low-bandwidth links, coupled with increased node and link failures, when compared to local-area networks. By augmenting a familiar POSIX interface with support for semantic cues, WheelFS provides a wide-area distributed storage system intended to help multi-site applications share data and gain fault tolerance, in the form of a distributed file system. Its design allows applications to adjust the tradeoff between prompt visibility of updates from other sites and the ability for sites to operate independently despite failures and long delays. WheelFS is implemented as a user-level file system and is deployed on PlanetLab and Emu-lab.(cont.) Six applications (an all-pairs-pings script, a distributed Web cache, an email service, large file distribution, distributed compilation, and protein sequence alignment software) demonstrate that WheelFS's file system interface simplifies construction of distributed applications by allowing reuse of existing software. These applications would perform poorly with the strict semantics implied by a traditional file system interface, but by providing cues to WheelFS they are able to achieve good performance. Measurements show that applications built on WheelFS deliver comparable performance to services such as CoralCDN and BitTorrent that use specialized wide-area storage systems.by Jeremy Andrew Stribling.Ph.D