Correct-by-Construction Network Programming for Stateful Data-Planes

Abstract

As switch hardware becomes faster, more stateful, and more programmable, functionality that was once confined to end hosts or the control plane is being pushed into the data plane. For example, recent work on adaptive congestion control and heavy hitter detection uses stateful switches to implement sophisticated functionality with only minor controller involvement. In applications where correctness depends on individual switches making coherent decisions, it is important that the switches have a consistent view of global state. However, such a consistency requirement makes it difficult to maintain efficiency (high throughput), due to the CAP theorem. Moreover, previous work on data-plane programming provides little to no built-in support for addressing this difficulty. We propose Callback State Machines(CSMs), a new high-level declarative network programming abstraction which allows operators to write correct data-plane programs against global state. CSMs offer programmers useful consistency guarantees without the need to manage how global state is replicated/updated at the individual switch level. To aid in the implementation of this high-level programming framework, we present a flexible new intermediate representation (IR) called TAPIR that natively supports stateful data plane functionality, as well as a compiler to generate device-specific code such as P4 from TAPIR code. Additionally, we demonstrate the power of TAPIR itself by using it to build a working implementation of the CONGA congestion control system.