This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: A Digital Fountain Approach to Reliable Distribution of Bulk Data John W. Byers Michael Luby Michael Mitzenmacher Ashutosh Rege Abstract The proliferation of applications that must reliably dis- tribute bulk data to a large number of autonomous clients motivates the design of new multicast and broad- cast protocols. We describe an ideal, fully scalable pro- tocol for these applications that we call a digital foun- tain. A digital fountain allows any number of hetero- geneous clients to acquire bulk data with optimal e- ciency at times of their choosing. Moreover, no feedback channels are needed to ensure reliable delivery, even in the face of high loss rates. We develop a protocol that closely approximates a digital fountain using a new class of erasure codes that for large block sizes are orders of magnitude faster than standard erasure codes. We provide performance mea- surements that demonstrate the feasibility of our ap- proach and discuss the design, implementation and per- formance of an experimental system. 1 Introduction A natural solution for software companies that plan to eciently disseminate new software over the Internet to millions of users simultaneously is multicast or broad- cast transmission . These transmissions must be UC Berkeley and International Computer Science Institute, Berkeley, California. Research supported in part by Na- tional Science Foundation operating grant NCR-9416101. Email: email@example.com International Computer Science Institute, Berkeley, California. Research supported in part by National Science Foundation operating grant NCR-9416101. Email: firstname.lastname@example.org Digital Systems Research Center, Palo Alto, California. Email: email@example.com International Computer Science Institute, Berkeley, California. Research supported in part by National Science Foundation operating grant NCR-9416101. Email: firstname.lastname@example.org fully reliable, have low network overhead, and support vast numbers of receivers with heterogeneous character- istics. Other activities that have similar requirements include distribution of popular images, database repli- cation and popular web site access. These applications require more than just a reliable multicast protocol, since users wish to access the data at times of their choosing and these access times will overlap with those of other users. While unicast protocols successfully use receiver ini- tiated requests for retransmission of lost data to provide reliability, it is widely known that the multicast analog of this solution is unscalable. For example, consider a server distributing a new software release to thou- sands of clients. As clients lose packets, their requests for retransmission can quickly overwhelm the server in a process known as feedback implosion. Even in the event that the server can handle the requests, the re- transmitted packets are often of use only to a small subset of the clients. More sophisticated solutions thatsubset of the clients....
View Full Document
This note was uploaded on 10/26/2011 for the course CS 7260 taught by Professor Staff during the Spring '08 term at Georgia Institute of Technology.
- Spring '08