Scalable Policy Management in Edge Virtual Bridging

The patent segments the monolithic TCAM rule replication approach into a hierarchical structure with a global policy store and per-VM policy caches. The global policy store maintains the master policy definitions, while individual VMs maintain only their specific policy instances in local caches or soft state tables, dramatically reducing the complexity growth from linear O(N) to logarithmic or constant time complexity for policy lookups.

The patent introduces a new dimensional approach by adding a temporal and hierarchical dimension to policy management. Instead of flat TCAM replication across all VMs, policies are organized in a multi-level hierarchy with global policies at one level and VM-specific policies at another, with time-based soft state expiration mechanisms. This dimensional transformation allows the system to handle large numbers of VMs without proportional increases in TCAM resources.

The patent merges the policy storage function into a centralized global policy store that is shared across all VMs, eliminating the need for separate TCAM rule copies for each VM. The global policy store maintains a single instance of each policy definition, which is then referenced or cached by individual VMs as needed, reducing total resource requirements from O(N) to O(1) for the master policy definitions.

The patent implements selective copying where only necessary policy portions are copied to VM-specific caches or soft state tables, rather than replicating entire TCAM rule sets. The copied data structures are maintained in soft state with expiration mechanisms, allowing the system to use minimal local copies while maintaining policy enforcement, significantly reducing the quantity of resources required compared to full replication.

The patent introduces dynamic policy management through soft state data structures with time-based expiration and refresh mechanisms. Policy information is dynamically cached at VM level with automatic expiration and renewal, allowing the system to scale to large numbers of VMs without requiring permanent TCAM rule copies for each one. This dynamic approach transforms the static O(N) replication model into a scalable model where resource usage does not grow linearly with VM count.

The patent changes the fundamental parameters of policy storage by transitioning from hard-state permanent TCAM replication to soft-state cached policies with expiration. By changing the persistence parameter from permanent to temporary with automatic renewal, the system achieves scalability while maintaining policy control. The parameter change from synchronous TCAM updates to asynchronous policy caching and expiration enables the system to handle large-scale VM environments efficiently.