Data dashboard: Aspera’s user interface lets a user control data-transfer rates and shows transfer times and real-time network information.
Unlike TCP, FASP does not wait for confirmation of receipt, but simply assumes that all packets have arrived, says Simon Hudson of Cloud2, a provider of cloud-computing services in East Yorkshire, U.K., and an early adopter of FASP. Under this protocol, only packets that are confirmed to have been dropped are re-sent. “And instead of sending lots of small packets, it sends fewer large packets,” Hudson says. The result is that the available bandwidth is used more efficiently–more data gets through, and it gets there faster.
Another issue is traffic monitoring, says Anna Liu, a cloud-computing researcher at the University of New South Wales in Sydney, Australia. “In cloud computing, the challenge is the unpredictable nature of the public network,” she says. “You can’t control what else is happening on the network due to other people’s activities.”
FASP handles this unpredictability by monitoring all network traffic and altering the size of packets and the rate and order in which they are sent, according to available bandwidth and other traffic issues. This way, the data flow can be regulated, ensuring that FASP data gets through without saturating the network. This also means it becomes possible to guarantee file-transfer times, says Munson. When transferring data over a 100 Mbps connection, Munson says, “FASP will achieve about 95 Mbps or better.”
Since Amazon is such a big player in cloud computing, its adoption of FASP could broaden the appeal of the technology, says Trigg. “It’s the 800-pound gorilla in the market,” he says. “If you improve the network connection, you lower the hurdle and allow more people to use it.”