Tunnel Keep-Alive Transmission via Intra-Node Multicast
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional networks that incorporate a heartbeat/keep-alive mechanism face scalability issues due to the control processor's need to individually create and transmit keep-alive packets for each tunnel, leading to high CPU cycle usage and congestion, which impede other basic protocols.
Innovation Solution
An intra-node multicast mechanism is used where a single packet trigger transmission is employed by the forwarding engines of the network switching device to send keep-alive messages for multiple tunnels, reducing the load on the control processor and utilizing multicast capability to distribute the load across multiple forwarding engines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the control processor individually creates and transmits keep-alive packets for each tunnel, then tunnel reachability monitoring is achieved, but CPU cycle usage increases and network congestion occurs
Solution Approach 1:
The patent merges multiple individual keep-alive packet transmissions into a single multicast packet. Instead of the control processor creating separate packets for each tunnel, a single multicast packet is created and distributed to all forwarding engines, which then replicate it for their respective tunnels. This combining approach maintains monitoring reliability while dramatically reducing CPU overhead.
Solution Approach 2:
The patent introduces forwarding engines as intermediaries between the control processor and the tunnels. The control processor sends a single trigger packet to the forwarding engines, which then handle the replication and transmission of keep-alive messages to multiple tunnels. This intermediary approach offloads the heavy processing burden from the control processor to the forwarding engines.
2Reliability
If the control processor transmits keep-alive packets on all tunnels, then tunnel availability is monitored, but network congestion increases and other protocols are impeded
Solution Approach 1:
The patent combines multiple individual keep-alive transmissions into a single multicast transmission. By merging the packet creation and initial transmission steps into one action, the total number of packets traversing the network is reduced, thereby decreasing network congestion while still achieving comprehensive tunnel monitoring.
Solution Approach 2:
The patent segments the keep-alive transmission function across multiple forwarding engines. Each forwarding engine receives a single multicast trigger packet and independently replicates it for its managed tunnels. This segmentation distributes the transmission load and reduces overall network traffic compared to individual unicast transmissions from the control processor.
3Reliability
If individual packet construction is performed for each tunnel, then accurate tunnel monitoring is achieved, but the mechanism does not scale to large numbers of tunnels
Solution Approach 1:
The patent creates a universal multicast packet format that serves all tunnels simultaneously. A single packet structure with multicast addressing can be replicated by forwarding engines to monitor multiple tunnels across the network. This universal approach allows the system to scale to large numbers of tunnels without requiring individualized packet construction for each tunnel.
Solution Approach 2:
The patent uses packet replication by forwarding engines instead of individual packet construction. Each forwarding engine receives a template multicast packet and copies it for its specific tunnels, inserting appropriate identifiers. This copying mechanism is highly scalable and maintains monitoring accuracy while reducing the control processor's workload.
Data Source
AI summary
A method, apparatus and computer program product for efficiently transmitting tunnel keep-alive messages is presented. A tunnel keep-alive trigger message is provided from a control processor of a multi-processor network switching device to at least one forwarding engine of the network switching device. The at least one of the forwarding engines produces a tunnel keep-alive packet as a result of receiving the tunnel keep-alive trigger message from the control processor. The at least one forwarding engine forwards the tunnel keep-alive packet on all tunnels the forwarding engine is servicing.