Single-Tier SPB Routing for Data Center Traffic Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network routing mechanisms, such as overlay methods and TCAM-supported lookups, are inefficient in handling multicast and require expensive hardware, leading to wasteful network bandwidth usage and high power consumption, especially in dynamic data center environments where server and host placement is decentralized.
Innovation Solution
Implementing a single-tier routing architecture with a Layer 3 (L3) controller using standard SPB encapsulation for centralized control over a distributed datapath, which includes propagating messages with L2 VSN ID, L3 VSN ID, and gateway MAC address to optimize traffic routing and reduce encapsulation overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If overlay methods are used for routing, then network flexibility and traffic management capability are improved, but network bandwidth efficiency deteriorates due to encapsulation overhead and inefficient multicast handling
Solution Approach 1:
The patent extracts the control plane functions (routing decisions, flow management) from the data plane (packet forwarding), placing control logic in centralized controllers while switches perform simple forwarding based on flow tables. This separation eliminates the need for complex encapsulation overhead in overlay methods while maintaining traffic management flexibility through centralized control.
Solution Approach 2:
The patent implements a universal forwarding mechanism using SPB protocols that handles both unicast and multicast traffic efficiently through a single forwarding framework. The system uses a unified approach where all switches in the network participate in the same forwarding protocol, eliminating the need for separate overlay tunnel mechanisms and reducing bandwidth overhead.
2Adaptability or versatility
If TCAM-supported lookups are used for custom forwarding rules, then forwarding flexibility and custom rule implementation are improved, but hardware cost and power consumption deteriorate
Solution Approach 1:
The patent replaces the mechanical TCAM hardware system with a software-based control plane approach. Instead of using expensive TCAM components in switches for complex lookup operations, the system uses centralized controllers that compute forwarding rules and install them in simple flow tables on switches. This substitution dramatically reduces power consumption and hardware cost while maintaining forwarding flexibility through software programmability.
3Stability of the object's composition
If traditional 3-tier routing architecture is used, then network stability and hierarchical control are improved, but network efficiency deteriorates in dynamic data center environments with decentralized host placement
Solution Approach 1:
The patent implements a dynamic forwarding system where flow tables in switches are programmatically configured by centralized controllers based on real-time network conditions and traffic patterns. This allows the network to adapt dynamically to changing host placements and traffic demands in data center environments, maintaining efficiency despite the decentralized and dynamic nature of modern data center architectures while preserving network stability through centralized coordination.
Data Source
AI summary
A method, apparatus and computer program product for providing Virtual Routing and Forwarding (VRF) and gateway Media Access Controller (MAC) distribution is presented. At least one subnet associated with a Layer 2 Virtual Switching Network (L2VSN) is provided on a network device. A message is propagated to a distributed Datapath. Network devices install the message as a routable MAC address on the L2VSN for the Layer 3 Virtual Switching Network/Virtual Routing and Forwarding (L3VSN/VRF) associated with the message. Edge devices route packets on the L2VSN addressed to the gateway MAC address.