Computer Networks Education in the New Era: Challenges, Opportunities, and What Should SIGCOMM Do?

Computer Networks Education in the New Era: Challenges, Opportunities, and What Should SIGCOMM Do?

Host: Qiao Xiang (Xiamen University)

Panelists: Kenneth Calvert (University of Kentucky), Cristel Pelsser (UCLouvain), Ying Zhang (Meta), Alan Liu (University of Maryland, College Park), Zili Meng (Hong Kong University of Science and Technoloy)

Scribe: Jinghui Jiang (Xiamen University), Ziyi Wang (Xiamen University), Xing Fang (Xiamen University)

Introduction
Every year, we SIGCOMMers organize many research conferences, workshops, and seminars to share new research ideas and results. However, there are few such venues to talk about education issues in computer networks. This side event proposes to discuss (1) the challenges and opportunities of computer network education, at both undergraduate and graduate levels, in the new era of emerging technologies (e.g., AI, quantum, and new computing architectures) and (2) what should SIGCOMM do in this new era to better prepare students with sufficient competencies in computer network for their future career.

Q&A:
Q1: How do we draw students interest in network, and prepare them for a related career?

Ying: She shared her experience from three aspects: The first aspect is about impact, which can introduce students to the importance of our research field from the perspective of commercial services, ensuring that these services are up and running, which is very meaningful. The second aspect is about the technical stack. Let students actually understand and learn the underlying technical stack of the system, which can help them have a better understanding of the technical stack. The third perspective is about mentorship. Mentors who can provide guidance and growth opportunities for young people are more attractive.

Kenneth: He thinks there are different kinds of students, and everybody is different in some way, but some students want to really understand how things work. He thinks that students who are happy to find out what’s going on behind the scenes are the ones you’d really like to attract. “Remember, the goal isn’t to dissuade students from AI but to show them how networking and AI complement.”

Cristel: Her experience is that we can add some computer science courses or activities for high school students, helping students understand network problems, protocol design, different layers, and how networks work in reality to prepare them for university-level thinking.

Alan: He thinks that seeing real things is important; we should think about how to give them information about what networking devices really are for the first time. These real-world demonstrations are very helpful in drawing students’ interest in the network. And in terms of preparing for a career, he thinks we should definitely invite what people like to talk about. It’s such a great life in the networking industry, and, professor here, we are doing networking research; we have also been doing well.

Zili: He recommends that professors simplify network courses, reduce the number of concepts and projects, and let students explore on their own. Passionate students will dive deeper into what interests them. As for preparing for industry, he believes we don’t need to over-prepare them. As long as we provide basic knowledge, they will figure out the rest through continued learning.

Q2: What does LLM bring to teaching networks? What is your experience?

Zili: He encourages his students to use LLM and write code using Copilot. Using LLM is more like problem engineering, and students need to know what prompts to use and how to ask their questions specifically to GPT. Writing code with Copilot is fine for students who don’t know where to begin.

Alan: He encourages the use of LLMs but in a slightly different way. He allows his students to use LLMs to learn concepts but not to generate assignment code. If students use LLMs to get reference codes, they need to provide the prompts they used and the citations.

Cristel: She usually doesn’t push students to use LLMs, and she tries to prevent them from using LLMs. She tries to prevent students from using LLMs because she is not convinced it improves their learning. It’s a good tool for improving writing and finding keywords for searches, but beyond that, she is not convinced.

Kenneth: He thinks that things evolve at different speeds and in different states, such as immutability, routing paradigms, etc. There is so much knowledge now. But the underlying basics are still basically the same. So, can LLM help students understand what is really going on? Can they explain what is happening behind the scenes? This is what he is interested in trying.

Ying: She thinks there are two aspects. One aspect is that in broader education, LLMs are useful for onboarding new engineers. Networking is full of acronyms, and LLMs can efficiently deliver this meta-knowledge to newcomers. Another aspect is making solid reports from the National Academies more accessible. They still generate paper-based reports that few read. These are high-quality reports from various scientists, so can we use LLMs to make them accessible to the public, even to high school students? This could be an area where LLMs help with education in general.

Q3: What educational activities should SIGCOMM organize?

Cristel: She thinks we should invest time in designing very simple exercises for kids or middle and high school students. For example, in computer science, we can find many exercises, but there are very few in networking. She thinks investing time in designing some exercises that can be done with just paper and pencil, such as larger routing exercises, would be very useful for attracting more people to participate in the class.

Kenneth: There are actually a lot of educational resources on the SIGCOMM website, but they are probably not being fully utilized. So try linking them to the main sigcomm page, which may help to find some resources that should be there.

Audience question: SIGCOMM has a tradition of rotating conferences around the world, which is great. But it also brings a problem. Many students are international students, and SIGCOMM rotating around the world makes it difficult for many international students to attend. It’s a community problem—what can we do about it?

Host: Qiao thinks there are two aspects: first, maybe we can provide more travel funding for students who can attend; on the other hand, SIGCOMM is also working hard to make online participation more interesting and to make the discussions and learnings from the conference accessible to those who cannot attend.

This year’s SIGCOMM Education Forum discussed many educational issues that need to be addressed in the network field. As LLM becomes increasingly popular, how to maintain talent investment in basic fields such as computer networks and how to use LLM to assist in talent training in the network field have become some issues that need to be concerned about. This year’s SIGCOMM Education Forum showed the views of the experts in the field on these issues, which has broad and sufficient significance for talent training in the network field.

1 Like