CNS Core: Small: Efficient Interoperability Testing of Heterogeneous Network Protocol Implementations

点击翻译按钮获取中文摘要

There has been an increasing interest in using hardware to implement network protocols, such as in ultra-high-speed networks (e.g., 40Gbps, 100Gbps) for scientific data movement. This is mainly because the processor overhead caused by the current software network protocol implementations increases considerably as the network speeds increase. However, there are usually big differences between software and hardware implementations for the same network protocol due to the unique hardware design constraints. Therefore, it is critical to test the interoperability between heterogeneous network protocol implementations to make sure that they can correctly interact with each other.This project proposes a new class of testing methods to efficiently and systematically check the interoperability of emerging heterogeneous network protocol implementations under two types of packet parameters: packet dynamic parameters (e.g., packet delay and loss), and packet semantics parameters (e.g., formats and meanings of packet fields), as different network protocol implementations communicate and interact with one another using packets. There are three research thrusts. The first research thrust focuses on interoperability testing with the packet dynamic parameters, the second focuses on the packet semantics parameters, and the third focuses on the mixed parameters.This project will improve not only the performance but also the stability of networks, especially ultra-high-speed networks where hardware implementations play critical roles. This project will benefit a large group of ultra-high-speed network users, such as scientific users who often need to move a huge amount of data over ultra-high-speed networks. In addition, the research problems identified in this project are related to three communities: networking, hardware, and software engineering; this may stimulate more possible collaborations among these three communities.The project documents, source code, and related data will be available on website https://hinterop.github.io, which will be maintained for at least three years after the end of the project.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.