High-Precision End-to-End Latency Guarantees Using Packet Wash

Many future networking applications demand high-precision end-to-end latency guarantees. Past QoS schemes have focused on scheduling and prioritization schemes for packets in their entirety but further advances are needed. Packet Wash is a recently proposed technology that allow for selective prioritization of payload chunks within packets themselves. Its primary intention is to reduce the number of packet drops when network congestion occurs by providing an option to drop less important payload portions first. However, Packet Wash opens up other possibilities. This paper introduces the use of Packet Wash to facilitate high-precision end-to-end latency guarantees. It involves a novel QoS algorithm that allows to reduce packet dwell time in routers by applying Packet Wash operations when the packet is in danger of being late, trading off transmission of low-priority chunks against improved latency.