IPv6 GlobalRouting
Student: Shashwat Patni
Mentors: Tommaso Pecorella, Manoj Kumar Rana
Organisation: ns-3 Network Simulator
Project Wiki: link
Repository: Repository
📖 Project Overview
The goal of this project was to extend ns-3’s native GlobalRouting protocol to support IPv6 while minimizing code duplication from the existing IPv4 implementation. Global Routing is based on the quagga implementation of the OSPF Routing Protocol (RFC 2328) and uses the Djikstra Shortest path algorithm to compute paths for a single area network configuration. The problem is that GlobalRouting doesn’t work for IPv6, and that’s a huge limitation. The goal of the project was to fix that limitation . Over the summer , we extended and refactored the code , to make globalRouting Also Work for IPv6 , using the same fundamental Algorithm, and tweaks to handle IPv6 vs IPv4 Differences.
Along the way we faced some obstacles:
One of the main goals was to come up with a way to minimize code duplication. During the Project we Explored different programing alternatives to the problem, namely static polymorphism vs dynamic vs a hybrid approach. Each Design came with its own Pros and Cons . We felt that Template Meta-Programming offerred the best alternative. The tradeoff between dynamic and Static Polymorphism was runtime Performance vs Compile time Checks and complexity with the code base . We finally decided to go with the Static Route due to its various advantages. We actually developed both versions halfway , to better understand each in GlobalRoutings Context. This way we could understand that using Runtime Checks was not good enough for our use case.
After discussing with Mentors it was decided to use the Static Polymorphism approach to Generalize IPv4GlobalRouting to the generic GlobalRouting using template Meta-Programming. Using Compile Time Checks , we instantiated Two versions of the template , one for each of the address class.
Along the way I found and fixed some CRITICAL bugs lurking in the protocol and also added a few performance improvements.
Merge Requests and Commits Summary
Features Implemented
- Refactor Ipv4 Implementation of Global Routing To a generic API to allow for the further extention to IPv6
- Handle link-local addresses alongside global unicast addresses. Link Local Addresses are not routed by definition but are used as Next Hop entries in the Forwarding Tables for IPv6.
- Ensure minimal code duplication by reusing existing IPv4 logic where possible.
- Provide route visualization and printing functionality.
- Add unit tests to validate correctness of IPv6 routes.
- Handle Non-onlink addresses by asserting when encountering them.
- Handle the route Forwarding Differences between IPv4 and IPv6 for the network 3 Layer
Milestones
The Project was divided into three Phases and two Milestones:
Phase 1:
- Port existing Ipv4GlobalRouting to generic GlobalRouting
- Generalize ipv4-global-routing-test-suite and global-route-manager-impl-test-suite
- Create tests and verify GlobalRouting works for Ipv4 after templatization
- Design Tests for Ipv6 Specific Properties for GlobalRouting
- Introduce Type Aliases for IP classes
Phase 2:
- Extend GlobalRouting to IPv6
- Handle the Implementation differences of IPv6
Phase 3:
- Work On suggested Changes
- Add Documentation For IPv6 GlobalRouting
- Prepare edge Testcases, Check for Bugs , fix any remaining issues
| — | Deliverables | — |
|---|---|---|
| Milestone 1 | Generalize Ipv4GlobalRouting to Generic Global Routing Classes using Templates | Source Code |
| Milestone 2 | Add Ipv6 Support to Global Routing | Source Code |
My Experience
Acknowledgements
I would like to thank all my mentors and the ns-3 developers for providing guidance and advice throughout the project, and for the opportunity to contribute to ns-3 through GSoC ‘25. GSoC taught me alot. I went from being clueless about open source , to my very first commit , to actually contributing in a meaningful way. Thank you to everyone from the community as well as from Google Summer of Code. A very special thanks to Tommaso for the constant support and encouragement. The weekly meetings were as fun as they were a learning experience. Discussing bugs and Design choices with the seasoned Pros of ns-3 was a surreal and amazing learning experience.
Challenges Faced
There were definitely times of doubt. When i felt like this bug will not go away, or there is something wrong with my design. For example: when i found the first few bugs in the implementation or when i was looking into the dynamic Polymorphic approach. Discussing with mentors and ns-3 devs helped a lot. These challenges helped me learn, as well. For starters, the Bitter/sweet Truth about programming in general. There are no dead ends, there are always alternatives leading to similar answers. Do NOT be Afraid to ask for help .
Suggestions for Future Contributors
- GSoC is as fun as you make it. Do not treat it as work. See it as a learning oppurtunity at every curve
- Do not be afraid to ask for help.
- Be honest with your mentors. They are there to help You.
Other Important Links
- GSoC Project Page: Link