Avalanche : Long Way To Go, But!!

Kevin Schofield writing on Avalanche highlights that one of the potential problems with P2P distribution systems is the "last byte" problem. If your scheme is to break up the original file into smaller pieces and distribute them around, that means that recipients need to do the reverse: identify at least one copy of each of the N pieces, potentially spread across N machines, download them all, and stictch them back together. If N-1 of those machines are fast and responsive, but one is under heavy load or has a bad network connection, then you're not done until that one slow machine finishes sending its piece of the file. Further, if one of the pieces suffers from data corruption, then the end result will be corrupted. The researchers on the Avalanche project are working on ways to avoid the "last byte" problem, and to provide some error-checking ability. Somewhat independently, but equally important, they are looking at how to support the notion of "authorized" content, which could be an official, verifiable patch to a piece of software, or a way to protect the intellectual property rights of content creators. Last week the researchers described these broad parameters of their research, and drilled down more on the network protocol work. They also released a research paper that went into more detail on their algorithm and in order to evaluate the efficiency of their algorithm,provided benchmark comparisons they had done to the most established P2P distribution system out there today: BitTorrent.
Content owners should have the ability to protect their content in appropriate ways (they may choose not to use that ability), and if we can create technologies that combine the best network efficiencies of P2P networks with ways to appropriately protect content, great things could happen. Bram Cohen Avalanche is vaporware. It isn't a product which you can use or test with, it's a bunch of proposed algorithms. There isn't even a fleshed out network protocol. The 'experiments' they've done are simulations. It's a bad idea to give much weight to simulations, especially of something so hairy as real-world internet behavior. BitTorrent in a way which is useful. One thing badly missing from this paper is back-of-the-envelope calculations about all of the work necessary to implement error correction. Potential problems are on the wire overhead, CPU usage, memory usage, and disk access time. Particularly worrisome for their proposed scheme is disk access. If the size of the file being transferred is greater than the size of memory, their entire system could easily get bogged down doing disk seeks and reads, since it needs to do constant recombinations of the entire file to build the pieces to be sent over the wire Looking at history Microsoft always starts late - comes with an inferior product/idea, slowly enters market - by all means increase market share and get to dominate - it owuld not be per se bad for the P2P industry should Microsoft begin to focus there.
Category :P2P