Active-Active Stateful Network Service Cluster Traffic Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Stateful network services face bottlenecks in scaling due to high CPU intensity, particularly in virtual environments, where traditional active-active clustering methods lead to significant traffic redirection, reducing throughput and increasing CPU load.
Innovation Solution
A method where a first service node identifies a second service node to process data messages and forwards state information for subsequent messages, allowing efficient routing and minimizing traffic redirection by using an overlay network and consistent hashing to manage service node memberships and state synchronization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If active-active clustering is used to scale stateful network services, then scalability is improved, but traffic redirection overhead increases significantly
Solution Approach 1:
The patent pre-establishes mapping relationships between client addresses and service nodes before traffic arrives. When a service node joins or leaves the cluster, the controller proactively updates the mapping information in the controller's table, so that when traffic arrives, the controller can directly forward it to the correct service node without requiring redirection. This preliminary action eliminates the need for real-time traffic redirection and its associated overhead.
Solution Approach 2:
The patent introduces a controller as an intermediary component that maintains a mapping table between client addresses and service nodes. The controller receives traffic from the load balancer, looks up the appropriate service node in its mapping table, and forwards traffic to the correct node. This intermediary approach centralizes the complexity of traffic routing and eliminates the need for service nodes to perform complex redirection operations themselves.
2Reliability
If traffic redirection is used to handle service node failures, then reliability is improved, but throughput decreases due to CPU overhead
Solution Approach 1:
The patent pre-establishes backup mappings in the controller's mapping table. When a service node fails, the controller already has alternative service node mappings ready, allowing it to immediately redirect traffic to backup nodes without requiring complex real-time failure detection and decision-making. This preliminary preparation of backup paths ensures both reliability and maintains throughput by avoiding CPU-intensive real-time redirection decisions.
Solution Approach 2:
The system implements health check mechanisms where service nodes report their status to the controller. The controller uses this feedback to maintain an updated mapping table that reflects the current operational status of service nodes. When failures are detected through feedback, the controller automatically updates its mapping table to route traffic to healthy nodes, ensuring reliability while maintaining efficient throughput through automated rather than manual redirection.
3Stability of the object's composition
If consistent hashing is used to select service nodes, then session state consistency is improved, but most traffic requires redirection to the selected node
Solution Approach 1:
The controller pre-calculates and stores the mapping between client addresses and service nodes based on consistent hashing in its mapping table. This preliminary calculation ensures that session state consistency is maintained while eliminating the need for real-time redirection. When traffic arrives, the controller simply forwards it to the pre-determined service node, maintaining both consistency and throughput.
Solution Approach 2:
The controller maintains a copy of the consistent hashing mapping relationship in its mapping table. This copy allows the controller to quickly determine the appropriate service node for each client address without requiring the receiving node to perform consistent hashing calculations or redirect traffic. The mapping copy enables direct forwarding while maintaining session consistency.
Data Source
AI summary
For a managed network, some embodiments provide a method for a set of service nodes in an active-active service node cluster in conjunction with a host computer hosting a destination data compute node (DCN) to improve the efficiency of directing a data message to a service node storing state information for the flow to which the data message belongs. a first service node receives a data message in a particular data message flow for which it does not maintain state information. The first service node then identifies a second service node to process the data message and forwards the data message to the second service node. The second service node sends state information for the particular data message flow to the first service node, for the first service node to use to process subsequent data messages in the particular data message flow.