distributed-ndn

Node hello processes

• Nodes running as processes
• Each node has a UDP server thread which listens for incoming packets and a main thread which sends hello to all peers

Graph simulation

• 1000x1000 grid space where each node will have a random physical coordinate
• 20 nodes
• Every node will form links with 5 nearest neighbors
• Balancing NUM_NODES and CONNECTIVITY (k) will give a graph which will be connected even when some nodes die.

How to run graph simulation

1. Install packages

pip install -r requirements.txt

2. Execute main.py

cd simulations && python3 main.py

Network Layer (Theoretical)

1. Simulate wireless network using dynamic node positions
   • a central coordinate for every group
   • nodes running in group get coordinates within the defined radius of the group
   • neighbor discovery: every node listens for neighbors: calculate received signal strength as a function of euclidean distance and noise
   • A common file for all node to figure out its coordinates and neighbors
2. All communication using UDP packets
3. Advertisements to all nearby nodes.
   • Content
     • Data information
     • Current timestamp
     • Proof of identity signed data + timestamp
     • public key
   • When received and verified, node registers as a neighbor and adds to neighbor table as a link
   • Node has an advertisement period
   • If neighbor fails to send 3 times (dampening), then it is removed from table
4. Node:
   • Neighbor table -> FIB
   • Cache -> Content Store
   • Pending Interest Table -> PIT
5. Node - either sensor or collecter
6. Longest prefix match