Distributed Software-Defined Network
Accelerated Raft Consensus SDN
🏆 Top 8/200+ teams – IEEE Kalpana Hackathon 2023
A distributed SDN control plane that stays consistent and scalable under failure via accelerated Raft consensus.
A distributed SDN model with an accelerated Raft consensus algorithm that keeps multiple controllers consistent while scaling protocol handling. Top 8 of 200+ teams at the IEEE Kalpana 30Hr Hackathon.
The Problem
Centralised SDN controllers were a scalability and single-point-of-failure bottleneck for network control.
System Architecture
Architecture design
Illustrative paths, not live telemetry.
Architecture overview. No node selected.
Full network reference
Network Apps
Control intent
Kind: client
Incoming (0)
None in this diagram
Outgoing (2)
- Controller Aflow
- Controller Bflow
Controller A
Replica
Kind: service
Incoming (1)
- Network Appsflow
Outgoing (2)
- Raft Consensusflow
- Controller Bfeedback / replicate
Controller B
Replica
Kind: service
Incoming (2)
- Network Appsflow
- Controller Afeedback / replicate
Outgoing (1)
- Raft Consensusflow
Raft Consensus
Accelerated
Kind: agent
Incoming (2)
- Controller Aflow
- Controller Bflow
Switch
Data plane
Kind: data
Incoming (1)
- Raft Consensusdata
Outgoing (0)
None in this diagram
Switch
Data plane
Kind: data
Incoming (1)
- Raft Consensusdata
Outgoing (0)
None in this diagram
What I Engineered
- Implemented a distributed SDN model with an accelerated Raft consensus algorithm.
- Replicated controller state across nodes to remove single points of failure.
- Enhanced network scalability for distributed-systems protocols.
Impact & Scale
- Consistent control plane across replicated controllers
- Accelerated Raft for faster leader election and failover
- 🏆 Top 8 of 200+ teams at IEEE Kalpana 2023
Stack
Portfolio context
Open to AI Software Development & Platform Engineering roles
Need someone who ships systems like this into production?