Project Overview

Project Name: Perfect ARP and NDP
Student: Zhiheng Dong
Mentors: Tommaso Pecorella, Ameya Deshpande, Manoj Kumar Rana
Google page: https://summerofcode.withgoogle.com/programs/2022/projects/DdUT09pZ
Project Goals: The project aims to implement a “Perfect” version of ARP and NDP which produce caches that contain all needed IP address–MAC address mapping. ARP and NDP are sometimes intrusive in high level testing, it will introduce unexpected noise to the simulation (unwanted delays, packet drops, etc.). Perfect ARP/NDP can ensure every ARP/NDP look-up hits the cache so that no ARP messages or Neighbor discovery messages will be generated.
Repository: https://gitlab.com/ZhiHengD/ns-3-dev
About Me: I’m a first-year computer science graduate student from University of Southern California, US. I got my bachelor’s degree in computer science from University of Wisconsin- Madison, US. I began my graduate study in August 2021. I’m interested in computer networks and have good code experience in C and C++, I also did projects with Python, Java and JavaScript.
Project Plan: link
Final Report:link

Milestones and Deliverables

Phase 1

Duration: Week 1 to Week 7

Phase 2

Duration: Week 8 to Week 10

Week1

Duration: June 13 - June 19

Start working on IPv4 perfect Arp.
Implemented an approach of creating subclass of ArpL3Protocol and ArpCache to dynamically create shared perfect cache.
Handled interface up/down event in that approach.
Tested IP neighbor cache flushing behavior about permanent entries in Linux.
Determined to implement perfect cache in a more concise way, which is create static perfect cache first after discussing with mentors.

Week2

Duration: June 20 - June 26

Week3

Duration: June 27 - July 3

Implemented a subclass of Icmpv6L4Protocol for disable DAD
Create two example for perfect ARP/NDP, the one is simple csma, another is double wifi. Meet problems when creating global routing in double wifi example, fix by using nix-vector routing instead.
Draft Merge Request of static perfect cache with two examples. link

Week4

Duration: July 4 - July 10

Add test for perfect ARP/NDP.
Add user-level API for updating perfect cache after IP address changes.
Update MR based on comments. Fix mistakes and improve iteration logic. link
Continue and refine the work on the dynamical perfect cache. Perfect cache updating will base on IP addresses added/removed instead of link up/down.

Week5

Duration: July 11 - July 17

Update MR based on comments.
Implement dynamical perfect ARP on locally shared cache per network.
Add new perfect-arp-cache and perfect-arp-l3-protocol.
Add addAddressCallback and removeAddressCallback to make it perfect caches updated dynamically by IP addresses change.
Use a new data structure for perfect cache, std::unordered_map<Ipv4Address, Cache *, Ipv4AddressHash> PerfectCache; The key is network address and the value is the pointer to the unorder_map of normal caches. This new structure is used to differentiate a case that when an interface has multiple IP addresses from different networks, its perfect cache would contain entries from different networks.
Overwite ArpCache::Lookup,ArpCache::Add,ArpCache::Remove for the new data structure of perfect ARP cache.

Week6

Duration: July 18 - July 24

Week7

Duration: July 25 - July 31

Refactor Icmpv6L4Protocol::ReceiveLLA and Icmpv6L4Protocol::HandleNA, change the nesting of if statement to switch case of NUD_STATE
Add new NUD_STATE: STATIC_AUTOGENERATED. The entries generated by perfect ARP/NDP will use the new state
Update printout in test and example, because of the new state and ARP/NDISC cache reorder.

Week8 - Week9

Duration: August 1 - August 14

I'm not fully involved in GSoC program these days because of some personal issue

Update MR based on comments.
add function to flush all auto-generated entries
add remove callbacks in ipv4interface to enable dynamical perfect cache
delete the double wifi example and modify the simple example to a detailed one to show all API and sending a data stream to show how perfect cache reduces the packet drop rate in some situation.