1 / 20

A Digital Fountain Approach to Reliable Distribution of Bulk Data

A Digital Fountain Approach to Reliable Distribution of Bulk Data. John Byers, ICSI Michael Luby, ICSI Michael Mitzenmacher, Compaq SRC Ashu Rege, ICSI. Application: Software Distribution. New release of widely used software. Hundreds of thousands of clients or more.

chacha
Download Presentation

A Digital Fountain Approach to Reliable Distribution of Bulk Data

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A Digital Fountain ApproachtoReliable Distribution of Bulk Data John Byers, ICSI Michael Luby, ICSI Michael Mitzenmacher, Compaq SRC Ashu Rege, ICSI

  2. Application: Software Distribution • New release of widely used software. • Hundreds of thousands of clients or more. • Bulk data: tens or hundreds of MB • Heterogeneous clients: • Modem users: hours • Well-connected users: minutes

  3. Primary Objectives • Scale to vast numbers of clients • No ARQs or NACKs • Minimize use of network bandwidth • Minimize overhead at receivers: • Computation time • Useless packets • Compatibility • Networks: Internet, satellite, wireless • Scheduling policies, i.e. congestion control

  4. Impediments • Packet loss • wired networks: congestion • satellite networks, mobile receivers • Receiver heterogeneity • packet loss rates • end-to-end throughput • Receiver access patterns • asynchronous arrivals and departures • overlapping access intervals

  5. Digital Fountain k Source Instantaneous Encoding Stream Transmission k Received Instantaneous Can recover file from any set of k encoding packets. k Message

  6. 0 hours 1 hour 2 hours 3 hours 4 hours 5 hours Digital Fountain Solution Transmission File User 1 User 2

  7. Is FEC Inherently Bad? • Faulty Reasoning • FEC adds redundancy • Redundancy increases congestion and losses • More losses necessitate more transmissions • FEC consumes more overall bandwidth • But… • Each and every packet can be useful to all clients • Each client consumes minimum bandwidth possible • FEC consumes less overall bandwidth by compressing bandwidth across clients

  8. DF Solution Features • Users can initiate the download at their discretion. • Users can continue download seamlessly after temporary interruption. • Tolerates moderate packet loss. • Low server load - simple protocol. • Does scale well. • Low network load.

  9. Approximating a Digital Fountain k Source Encoding Time Encoding Stream Received (1 + c) k Decoding Time k Message

  10. Approximating a DF: Performance Measures • Time Overhead: • Time to decode (or encode) as a function of k. • Decoding Inefficiency: packets needed to decode k

  11. Work on Erasure Codes • Standard Reed-Solomon Codes • Dense systems of linear equations. • Poor time overhead (quadratic in k) • Optimal decoding inefficiency of 1 • Tornado Codes [LMSSS ‘97] • Sparse systems of equations. • Fast encoding and decoding (linear in k) • Suboptimal decoding inefficiency

  12. = source data = redundancy Tornado Z: Encoding Structure Irregular bipartite graph Irregular bipartite graph k stretch factor = 2 k = 16,000 nodes

  13. Encoding/Decoding Process

  14. Encoding time, 1K packets Decoding time, 1K packets Size Reed-Solomon Tornado Z Size Reed-Solomon Tornado Z 4.6 sec. 0.11 sec. 2.06 sec. 0.18 sec. 250 K 250 K 500 K 19 sec. 0.18 sec. 500 K 8.4 sec. 0.24 sec. 93 sec. 40.5 sec. 1 MB 0.29 sec. 1 MB 0.31 sec. 2 MB 442 sec. 0.57 sec. 2 MB 199 sec. 0.44 sec. 4 MB 30 min. 1.01 sec. 4 MB 13 min. 0.74 sec. 8 MB 2 hrs. 1.99 sec. 8 MB 1 hr. 1.28 sec. 16 MB 8 hrs. 3.93 sec. 16 MB 4 hrs. 2.27 sec. Timing Comparison Tornado Z: Average inefficiency = 1.055 Both codes: Stretch factor = 2

  15. Cyclic Interleaving Transmission Encoded Blocks Interleaved Encoding Blocks Encoding Copy 1 File Encoding Copy 2 Tornado Encoding

  16. T B blocks Cyclic Interleaving: Drawbacks • The Coupon Collector’s Problem • Waiting for packets from the last blocks: • More blocks: faster decoding, larger inefficiency

  17. Scalability over File Size

  18. Scalability over Receivers

  19. Digital Fountain Prototype • Built on top of IP Multicast. • Tolerating heterogeneity: • Layered multicast • Congestion control [VRC ‘98] • Experimental results over MBONE.

  20. Research Directions • Other applications for digital fountains • Dispersity routing • Accessing data from multiple mirror sites in parallel • Improving the codes • Implementation and deployment • Scale to large number of clients • Network interactions

More Related