Stateful Service Gateway Failover in Public Cloud Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In public multi-tenant clouds, high availability logical network gateways face challenges in maintaining stateful services across active and standby instances operating in different physical locations, as the active and standby instances have different network addresses, leading to invalid connection states and service disruptions during failovers.
Innovation Solution
The centralized control plane distributes stateful service rules using public network addresses, with each instance mapping these addresses to their respective interface addresses, ensuring that the standby instance can seamlessly take over by converting connection states and service rules to its own interface addresses, maintaining service continuity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If active and standby gateway instances are deployed in different physical locations for high availability, then system reliability is improved, but connection state validity deteriorates because each instance has different network addresses
Solution Approach 1:
The patent introduces a mapping mechanism as an intermediary that translates between the network addresses of different gateway instances. The first gateway instance maintains a mapping to the second gateway instance's network address, allowing connection states to be valid across instances with different addresses. This mapping acts as a mediator that preserves connection state information while accommodating the physical separation of gateway instances.
Solution Approach 2:
The patent changes the parameter of network address representation by introducing mappings that associate different network addresses with each other. Instead of treating network addresses as fixed to physical locations, the system dynamically maps addresses between active and standby instances, allowing connection states to remain valid despite address differences in distributed deployments.
2Reliability
If centralized routing components are deployed in active-standby pairs across different datacenters, then service continuity is improved, but data traffic between datacenters increases
Solution Approach 1:
The patent implements preliminary action by pre-configuring the standby gateway instance with mapping information before a failover event occurs. The standby instance already has the mapping to the active instance's network address stored, so when failover is needed, the standby can immediately assume the active role without requiring real-time communication or reconfiguration. This preliminary preparation enables rapid failover while minimizing disruption to data traffic patterns.
3Adaptability or versatility
If gateway instances use different network addresses in different physical locations, then physical deployment flexibility is improved, but service rule applicability deteriorates
Solution Approach 1:
The patent achieves universality by creating a mapping mechanism that serves multiple purposes: it enables different gateway instances to use different network addresses for physical deployment flexibility, simultaneously maintains service rule applicability by translating addresses, and supports both active-standby operation and load balancing scenarios. The mapping infrastructure is universal enough to handle various deployment configurations without requiring separate solutions for each use case.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for a network controller that manages a logical network spanning multiple physical locations. For each physical location hosting data compute nodes (DCNs) belonging to the logical network, the method defines a centralized routing component for processing data messages between the DCNs hosted at the physical location and networks external to the logical network, assigns an active instance of the centralized routing component to operate at the physical location, and assigns a standby instance of the centralized routing component to operate at one of the other physical locations.