Fujitsu protocol

Fujitsu introduced a transport protocol for IP networks in 2013 that is an alternative to the connectionless UDP protocol and the connection-oriented TCP protocol and has a data throughput that is many times higher. The Fujitsu protocol is particularly suitable for error-prone and congested transmission links with high packet losses.

In order to reduce packet retransmissions with the TCP protocol as far as possible, the mechanism developed by Fujitsu distinguishes whether a data packet is lost or merely arrives late. As a result, only the data packets that have not arrived are retransmitted. This procedure reduces latency considerably, while at the same time increasing data throughput. Primarily, of course, on long and poor connections. These can be line connections or radio connections.

The figures released by Fujitsu for data throughput and latency are impressive. However, they refer to national and intercontinental connections, show a 30-fold higher data throughput and a latency reduced to one sixth. Other procedures implemented in the Fujitsu protocol concern real-timetraffic shaping, which ensures optimal utilization of transmission bandwidth. This mechanism takes into account the TCP traffic on connections and only occupies the remaining available transmission bandwidth. This prevents the Fujitsu protocol from occupying excessive bandwidth.

The Fujitsu protocol can be used without changing the TCP clients. It is recommended for connections that are susceptible to interference, but also for virtual desktops, file transfers oflarge files, mobile computing and for surfing the web.