Data trading races with Avalanche algorithms

时间:2019-03-03 08:19:02166网络整理admin

By Will Knight Mammoth software patches, movie files and TV-on-demand could be delivered swiftly to internet users using a smart network transfer technology developed by researchers at Microsoft. Its creators claim the networking scheme, known as Avalanche, improves upon the highly successful download algorithms employed by BitTorrent, a software tool that optimises the transfer of large amounts of data, such as software or video files. BitTorrent speeds up downloading, even when a file is in great demand, by slicing data into chunks and enabling users to share them with one another. Such a “peer-to-peer” download mechanism vastly increases available network bandwidth, even when many users start scrambling for a file. BitTorrent does not, however, ensure that these chunks of data reach a recipient in the most efficient manner, and this can mean a frustrating wait for that elusive last chunk. Avalanche – developed by Christos Gkantsidis and Pablo Rodriguez at Microsoft Research Cambridge, UK – aims to boost the efficiency of such a process. It encodes individual blocks of data using linear equations – in such a way that they slightly overlap – and passes on the encoded pieces instead. This overlapping process means that the original data can be reconstructed even if the user does not receive all the blocks of data. The researchers say computer simulations show the scheme could be two or three times more efficient than schemes that do not use encoding. “Peers do not need to find specific pieces in the system to complete, any encoded piece will suffice,” Gkantsidis and Rodriguez write on a project page that explains the research. “This makes the system very robust as peers disconnect. Also, no peer becomes a bottleneck, since no block is more important than another. Finally, network bandwidth is considerably reduced since the same information does not travel multiple times over bottleneck links.” Despite being used legitimately, mainly to trade software code, BitTorrent has also proven controversial because it is a popular way to share movie files, many of which are copyrighted. Neither Gkantsidis nor Rodriguez were available to comment on Avalanche, but a Microsoft representative told New Scientist that Avalanche, unlike BitTorrent, has been designed with copy prevention in mind. “It includes strong security to ensure content providers are uniquely identifiable, and to prevent unauthorised parties from offering content for download,” the spokesperson says. “Avalanche also ensures content downloaded to each client machine is exactly the same as the content shared by the content provider.” Adam Langley, a peer-to-peer networking expert in London, UK, says the system has its merits but the simulations used to test it could be flawed. “For one, I believe that their simulator assumes unlimited processing power and hard drive transfer speeds at the user nodes,” he told New Scientist. “I would suggest that a real-world network would take quite a hit from this.” “It’s a bad idea to give much weight to simulations, especially of something as hairy as real-world internet behaviour,” agrees Bram Cohen, the US programmer behind BitTorrent. “There isn’t even a fleshed out network protocol.” In a post to his online weblog on 20 June,