Title in Chinese: 互联网国际路由的风险分析和绕避博弈研究

There has been evidence that some nations have deployed surveillance systems on core Internet transit links that reside within their borders, which means international traffic flows of other nations which traverse these links to reach their destinations are under the security risk of being analyzed. Definitely this kind of cases will be more prevalent in the future. Therefore, it is very important to investigate the following questions.

1) In the current Internet routing, how significant is the exposure of its own international traffic flows by routing infrastructure to other nations? Can this nation make all its international traffic flows routing around its adversary nations successfully?

2) When the international situation becomes tense, every nation will try its best to revise the Internet routing decisions via any techniques and strategies. The final routing of traffic flows is determined by all participants, which means the actions of these nations are interacted with each other. When every nation cannot improve its situation by revising its own actions, a stable Internet would be reached. It is valuable to know the situation of each nation in the routing of the stable Internet.

We believe the research efforts of this project will promote research progress of Internet topology and routing inference, and have high significance for understanding and improving the routing security of traffic flows of our nation.

Title in Chinese: 跨域端到端弱确定性通信的网络加速技术研究

It is well-known that current interdomain routing in the Internet is mainly for ISPs' business considerations, and performance is not the first priority when making routing decisions. As a result, the resulting paths may not be optimal in terms of performance, such as latency, jitter, loss and throughput. Meanwhile, the Internet is a "best effort" network, which means it does not provide any performance guarantee. However, users of most services/applications are performance-senstive, so applications in the Internet do need a kind of end-to-end communication services with optimal performance or with guaranteed performance.

About two decades ago, researchers proposed the concept of "application-layer routing" or "apllication-layer traffic optimization" (ALTO) and tried to help applications improve their communication performance. At that time, the main concern was how to find a better overlay path and avoid the interference between underlay routing and overlay routing. In recent years, there have been more and more computation/storage resources in the Internet (aka clouds), and more and more applications become computation-intensive. As a result, application-layer traffic optimization does not only involve application-layer routing (which is to make proper decisions on networking resources scheduling for networking performance), but also involves caching (trade storage for better networking performance) and coding (trade computation for better networking performance).

We aim to provide Internet-scale end-to-end communication services with soft guarantees for various applications, especially for latency-sensitive applications and the content delivery of VR live streaming.

This research is supported by Huawei Ltd. (Overlay Routing Strategies Towards Cross-domain End-to-End Communication with Soft Guarantees, starting from Aug. 2021)

Title in Chinese: 面向IPv6的网络空间国际治理联合研发与示范

This is an interdisciplinary research project. See cyberspace governance for more information.

I am particularly interested in the following problems.

1) Optimizing the Performance of Geo-Distributed Video Analytics for Smart Cities.

2) Accerating Distributed Machine Learning.

3) Mechanism Design to Efficiently Allocate Computing Resource.