Panda is a flow scheduling in data centers based on PIAS. Panda can bound low latency for delay-sensitive applications and optimize the flow complete times for throughput-intensive applications on the premise that flow information is not known a priori. We observe that most packets generated by delay-sensitive applications are small, while by throughput-intensive applications are large. Panda takes advantage of the distinct flow size distributions to differentiate the two kinds of applications.
At its heart, Panda derives an optimal threshold to divide packets into two categories: large and small, ensuring that small packets dominate traffic from delay-sensitive applications and large ones dominate traffic from throughput-intensive applications. In addition, Panda allocates each flow a counter which is initiated with zero. Large packets increase the counter while small packets decrease it. Then Panda assigns priorities to flows according to their counters.
Panda has two components: a flow generator and PIAS kernel.
Panda is my master senior project and the thesis (in Chinese) is in docs/.