Controller-less Peer-to-Peer Distributed Switch via Sync Daemon
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current distributed virtual switches require centralized controllers, leading to complex configurations and single points of failure, making it difficult to dynamically change network topology and add or remove switches without major changes, and causing system degradation or failure if the centralized controller is down.
Innovation Solution
A controller-less peer-to-peer distributed switch system using a sync daemon to synchronize flow and action tables across peer nodes via a distributed hash table, eliminating the need for centralized control and enabling dynamic network adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If centralized controllers are used in distributed virtual switches, then control and management is simplified, but system reliability deteriorates due to single points of failure
Solution Approach 1:
The patent segments the centralized control function into distributed control plane instances that run on each peer node. Each peer maintains its own flow table and action table, eliminating the single centralized controller and its associated single point of failure. The control functionality is divided and distributed across multiple independent nodes.
Solution Approach 2:
The patent introduces a distributed hash table (DHT) as an intermediary mechanism that enables peer nodes to synchronize their flow and action tables without requiring a centralized controller. The DHT acts as a decentralized coordination layer that allows peers to discover and communicate with each other autonomously.
2Ease of operation
If centralized controllers are used to manage distributed virtual switches, then configuration updates can be pushed centrally, but adaptability deteriorates when the controller is down
Solution Approach 1:
Each peer node in the patent is equipped with autonomous capabilities to discover other peers, synchronize its flow and action tables, and make forwarding decisions independently. The peers self-organize into a functional network without requiring external control, enabling the system to adapt dynamically even when individual nodes fail or join.
Solution Approach 2:
The patent implements a dynamic peer-to-peer network where peers can join and leave the network on-demand. The network topology and peer memberships are dynamically updated through the DHT synchronization mechanism, allowing the system to adapt to changing network conditions without centralized reconfiguration.
3Productivity
If flow tables and action tables are maintained on each peer node, then distributed forwarding is enabled, but synchronization complexity increases
Solution Approach 1:
The distributed hash table serves as an intermediary synchronization mechanism that simplifies the maintenance of flow and action tables across peers. Each peer queries the DHT to obtain the current state of these tables, and the DHT automatically handles the coordination of updates and consistency across all nodes.
Solution Approach 2:
The patent uses copying to maintain consistency across distributed peers. Each peer maintains local copies of the flow table and action table, and these copies are synchronized through the DHT. When updates occur, the changes are propagated to all peers through the DHT, ensuring all nodes have consistent forwarding information.
Data Source
AI summary
In one embodiment, a device (e.g., a sync daemon) connects to peer nodes in a communication network. Each of the peer nodes has a switch that supports a flow table and an action table. The device facilitates a connection between a switch of a first peer node and a switch of a second peer node, and maintains the flow table and the action table of each of the peer nodes, such that the flow tables and the action tables are kept in synchronization with one another across each of the peer nodes via a distributed hash table.