APNIC Pty Ltd.

07/08/2024 | Press release | Distributed by Public on 07/07/2024 17:58

How deploying IPv6 at NITK led me to IETF

IETF Internet Standards Ambassadors (Vanessa is fourth from left) with Foundation Head of Programs and Partnerships Sylvia Cadena.

I'm an undergraduate student at the National Institute of Technology Karnataka (NITK), India, and part of the IPv6 migration team there. Our deployment efforts culminated in presenting our IPv6 deployment findings and testbed experiments at the IPv6 Operations (v6ops) Working Group session at IETF 118 and the subsequent v6ops side meeting at IETF 119. The feedback we received was invaluable, providing many insights and suggestions for the future direction of our project.

Last year, I was honoured to be selected as an APNIC Fellow for APNIC 56 in Kyoto, Japan. Attending the APNIC workshops and conference allowed me to learn so much more about deployment and network security. I am also really grateful to have had the opportunity to learn from and engage with some of the most influential figures in the Internet community.

Our team received generous support from the APNIC Foundation through the IPv6 Deployment Grant, enabling us to attend IETF 118 in Prague. I'm grateful for the sponsorship from the APNIC Foundation that allowed us to attend IETF 119 in Brisbane as Internet Standards Ambassadors.

I also had the opportunity to be part of the team conducting an IPv6 workshop at NITK (PDF) in March 2024. This workshop was attended by students and faculty from nearby universities. I delivered a session on the different IPv6 address assignment techniques. Kavya and I also conducted a hands-on lab session where participants were required to set up a functional IPv6-only testbed comprising a DHCPv6 server and a DNS resolver.

Participating in IETF

From 4 -10 November 2023, we attended the IETF 118 meeting for the first time in Prague. We were excited to be a part of the vibrant worldwide community that is the IETF, which is responsible for developing standards that shape the Internet. We presented on the progress of IPv6 deployment at NITK, and also discussed our project on Neighbor Discovery Enhancements. We presented our findings, learnings, and difficulties with the v6ops community.

Our journey at IETF 118 in Prague commenced with the collaborative energy of the Hackathon on 4 November 2023. Setting the groundwork for projects, we participated in the vibrant atmosphere of innovation. On the following day, 5 November, we completed registrations, attended a new participants overview, and engaged in Quick Connections. During HotRFCs, a presentation on IPv6 traffic progress sparked my interest in measuring IPv6 packet loss rate.

Monday was a day filled with meetings and social connections. Engaging in the BPF/eBPF (bpf) Working Group, we explored the eBPF Instruction Set (draft-thaler-bpf-isa-00) and the Instruction-level BPF Memory Model (slides, PDF). The day continued with a session that discussed the charter for the Detecting Unwanted Location Trackers proposed Working Group, followed by the TCP Maintenance (tcpm) Working Group meeting.

On Tuesday, active participation continued with meetings. Amidst the energy of IETF 118 in Prague, Tuesday unfolded as a day of insightful discussions. The day kicked off with the Congestion Control Working Group (ccwg) agenda, where we discussed the criteria for evaluating congestion control algorithms.

The main focus of the day was our active participation in the v6ops meeting. Here, we formally presented our ongoing work on IPv6 Deployment at NITK. In our presentation, we shared the current progress and some initial observations based on the gathered statistical data (slides, PDF). This presentation was a key moment, allowing us to connect with the v6ops community. The ensuing discussions provided varied feedback on our work, and we were fortunate to receive helpful suggestions for future projects and experiments.

My involvement in the NITK IPv6 project

One of the objectives of our project was to set up a testbed at NITK. The testbed involved the deployment of an OPNsense testbed router, configuring a Generic Tunnelling InterFace (GIF), switch setup, DNS server configuration using BIND9, IPv6 address assignment with KEA DHCPv6, and end client address configurations.

After setting up the testbed, my experiments focused on the DHCPv6 address assignment and the setting up of a DHCPv6 filter on the intermediate switch. I began exploring the filter's behaviour with and without relay, which led us to some interesting observations that we presented at IETF 119. I also looked into SLAAC for the testbed end clients and other Router Advertisement (RA) options on the OPNsense router. Part of my work was migrating the DHCPv6 server from mem file storage to a persistent database and exploring configurations for the production environment.

At NITK, our data centre is the backbone of the critical infrastructure, hosting routers, switches, firewalls, servers, and storage systems. We successfully migrated the data centre VLAN to IPv6, then enabled IPv6 routing for essential segments such as the Central Computing Centre (CCC) staff VLAN and the CCC lab VLAN. To efficiently manage IPv6 address allocation, we deployed a DHCPv6 server within the data centre VLAN, currently servicing the CCC staff and lab VLANs.

How we did it

Initially, we formed a team of faculty members, advisors, and students to work on the IPv6 deployment project.

To set up the testbed, we obtained the necessary equipment and pretty much started the setup by reading the manuals. The official manuals of the equipment/software we were using were helpful, but we also benefitted greatly from the tutorials and labs offered by APNIC Academy.

Getting the testbed to work end-to-end was not an easy effort - we faced more than a few issues that we would spend quite some time debugging. But luckily, we had great mentorship throughout! We followed a format of weekly/biweekly calls to discuss updates, any problems we faced, and different approaches to tackling them. Once the testbed was up and running, it served as a playground for us to test anything we wanted to put into production.

Setting up the testbed for end-to-end functionality was challenging, requiring significant time to debug various issues. Fortunately, we had excellent mentorship throughout the process. We held weekly and biweekly calls to discuss updates, troubleshoot problems and explore different approaches. Once operational, the testbed became a valuable playground for testing everything we planned to put into production.

Throughout all our tasks and experiments, we dedicated significant effort to documenting our process. We believed it was crucial to track every step we took, the issues we encountered, and how we addressed them. As a result, we produced several technical documents, including the Testbed Setup documentation (PDF) and DC VLAN Migration documentation (PDF). We hope these will serve as a valuable knowledge base for future IPv6 integration efforts both within and beyond NITK.

What's next?

Coming up, the main priority is to migrate the remaining campus VLANs as well. We are currently in the process of tagging all switches and access points across the NITK campus to support Dynamic VLANs. The plan is to initially configure IPv6 routing on a designated VLAN and add trusted users for IPv6 functionality testing. Here we will actively identify and address any issues encountered (and document them as well). If all goes well, we plan to enable IPv6 across all VLANs on campus.

As of now, most of our internal applications, like IRIS (our Management Information System (MIS)), have been migrated to support IPv6. We will gradually shift our focus towards evaluating the performance of IPv6 vs IPv4, comparing latency and overall performance.

I am committed to continuing my involvement in migrating our campus network to IPv6 and also mentoring the upcoming batch of students who will join the migration effort. With the testbed, I will continue conducting more experiments and gathering observations that we will require for the production environment. In particular, as we proceed with migrating VLANs, I am eager to look into IPv6 vs IPv4 performance testing and perform statistical data analysis.

Vanessa Maria Fernandes is a student at NITK Surathkal and an APNIC 56 fellow with an interest in networking.

The views expressed by the authors of this blog are their own and do not necessarily reflect the views of APNIC. Please note a Code of Conduct applies to this blog.