This communication algorithm combines elements of fountain and network coding to achieve low delay, high throughput connections while allowing multiple uncoordinated source transmission of the same file to the same destination. When communicating data over lossy networks (such as Wi-Fi, the internet, and most peer to peer connections), a major problem is that packets of information will be lost, and the receiving device will not be able to reconstruct the data sent. Rather than slowing connections by having the receiving device communicate which packages are missing, network engineers solve the lossy connection problem by encoding messages with redundancy that allows devices to reconstruct messages with missing data. Researchers at the University of Florida have developed an algorithm that combines the best features of fountain coding and network coding (two redundancy coding techniques) to solve loss of information over connections, increasing transmission throughput. Due to its great efficiency, the combination is uniquely well-suited for peer to peer content delivery networks, file transfer, distributed file storage, social networks and mobile TV.
Communication algorithm for higher efficiency data transmission
When communicating data over lossy networks, a major problem is that packets of information will be lost, and the receiving computer will not be able to reconstruct the data sent. The simplest solution is for the receiving computer to notify the original computer when each packet arrives so the packets can be resent. In high traffic situations, or over lossy connections, this slows communication speeds and is far less than ideal. Forward Error Correction (FEC) codes are a solution to this problem that has the original computer combine packets to be decoded by the receiving computer. This way, if the receiver loses information, it can reconstruct the original data from the combined received packets. Fountain coding, a particularly efficient form of FEC, sends random combinations of packets together; the receiving node can decode this data as soon as the size of the data received exceeds the file size even slightly. Fountain coding is akin to spraying water from a wide nozzle hose at a bucket far away, no matter how much water misses the bucket, eventually it will be full -- no matter how many packets are lost over the network, with fountain coding the message will be reconstructed once the size of the "bucket" (file size) is reached. Network coding is another solution to the lossy connection problem that distributes packets and random combinations of packets to nodes on the network, that can then forward or combine and forward packets to the receiving node, to be reconstructed. This algorithm combines the most efficient portions of fountain and network coding to achieve extremely efficient data transmission.