needed to view most of the images on this Web site - just a couple of clicks and you're "good to go."
Welcome
Photos of Larryblakeley
http://www.royblakeley.name/larry_blakeley/larryblakeley_photos_jpeg.htm
(Contact Info: larry at larryblakeley dot com)
Important Note: You will need to click this icon to download the free needed to view most of the images on this Web site - just a couple of clicks and you're "good to go."
I manage this Web site and the following Web sites: Leslie (Blakeley) Adkins - my oldest daughter
Lori Ann Blakeley (June 20, 1985 - May 4, 2005) - my middle daughter
Evan Blakeley- my youngest child
We propose a new scheme for content distribution of large files that is based on network coding. With network coding, each node of the distribution network is able to generate and transmit encoded blocks of information. The randomization introduced by the coding process eases the scheduling of block propagation, and, thus, makes the distribution more efficient. This is particularly important in large unstructured overlay networks, where the nodes need to make block forwarding decisions based on local information only. We compare network coding to other schemes that transmit unencoded information (i.e. blocks of the original file) and, also, to schemes in which only the source is allowed to generate and transmit encoded packets.
We study the performance of network coding in heterogeneous networks with dynamic node arrival and departure patterns, clustered topologies, and when incentive mechanisms to discourage free-riding are in place. We demonstrate through simulations of scenarios of practical interest that the expected file download time improves by more than 20-30% with network coding compared to coding at the server only and, by more than 2-3 times compared to sending unencoded information. Moreover, we show that network coding improves the robustness of the system and is able to smoothly handle extreme situations where the server and nodes leave the system.
An important problem of current collaborative content distribution networks is free-riding; many users take advantage of the resources offered to the network by other users without contributing their own resources. Free-riding can seriously degrade the performance of the content distribution [28], and, as a result, many collaborative networks have introduced mechanisms to discourage free-riding.
In our system we use two mechanisms to discourage free riding. The first is that we give priority to exchanges over free uploading to other nodes. In other words, when there is contention for the upload capacity of a user, the user will preferentially upload blocks of information to neighbors from which it is also downloading blocks. Thus, the nodes allocate their capacity preferentially to mutual exchanges and then use the remaining upload capacity for free downloads.
The second incentive mechanism that we use is inspired by the tit-for-tat approach used in the BitTorrent network [2]. A user does not upload content to another user unless it has also received enough content from that user; more specifically, the absolute difference of uploading minus downloading from one user to another is bounded.
In summary, with network coding all nodes are able to finish downloading even in extremal scenarios; on the other hand, only a small percentage of nodes finishes downloading when source coding or no coding is used. Moreover, even in such extremal scenarios, the performance perceived by the users will not be affected significantly with network coding. Thus, network coding adds robustness to the system and it is interesting to observe that the extra robustness is mostly introduced by the peers. This is opposed to source coding systems, which are dependent on the server to introduce extra encoded blocks and, hence, robustness.
The performance benefits provided by network coding in terms of throughput are more than 20-30% compared to coding at the server, and can be more than 2-3 times better compared to transmitting unencoded blocks. Moreover, we have observed that with network coding the system is much more robust to server and node departures.
Despite the rich literature in network coding, we are not aware of any operational content distribution network that uses network coding. Based on the system presented in this paper, we have implemented Avalanche, a real system using network coding. Through Avalanche, we are currently investigating the benefits of using network coding to distribute very large files to a large number of users in realistic settings.
- "Network Coding for Large Scale Content Distribution", IEEE/INFOCOM 2005 http://www.ieee-infocom.org/2005/, Miami. March 2005. (Also as Microsoft Research Technical Report (MSR-TR-2004-80). Christos Gkantsidis http://www.cc.gatech.edu/people/home/gantsich/ and Pablo Rodriguez Rodriguez http://www.rodriguezrodriguez.com/, http://www.research.microsoft.com/~pablo/papers/nc_contentdist.pdf
Directory: http://www.larryblakeley.com/Articles/p2p/
File Name: christos_gkantsidis_pablo_rodriguez200503.pdf
Post Date: March 18, 2005 at 11:55 AM CST; 1755 GMT