Layer-3 Subnet Stretching via Gateway Tunneling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for stretching subnets across distributed computing environments often rely on layer-2 (L2) communications, which incur high overhead and latency, and layer-3 (L3) hub configurations that increase processing complexity and error risk, limiting efficiency and availability during network migrations.
Innovation Solution
Implementing an exclusive layer-3 (L3) communication method for subnet stretching, utilizing gateway machines to manage IP addresses and tunneling protocols, allowing seamless subnet extension across multiple networks without reconfiguring applications or infrastructure, thereby reducing downtime and processing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If layer-2 (L2) gateway is used for subnet stretching, then communication continuity is maintained, but communication overhead and latency increase
Solution Approach 1:
The patent introduces gateway machines as intermediaries that operate at layer-3 to forward packets between stretched subnets. These gateways use routing tables and IP addressing to mediate communication, avoiding the overhead of L2 gateway mechanisms while maintaining communication continuity through proper packet forwarding and routing.
Solution Approach 2:
The patent replaces the layer-2 gateway mechanism with a layer-3 routing-based approach. Instead of using L2 switching and bridging mechanisms, the system uses IP routing, subnet masks, and gateway forwarding, which reduces communication overhead and latency while achieving the same subnet stretching objective.
2Device complexity
If layer-3 (L3) hub with subnet mask /32 is used for subnet stretching, then routing is simplified, but processing complexity and error risk increase due to central handler
Solution Approach 1:
The patent segments the subnet stretching function across multiple distributed gateway machines rather than using a single central L3 hub. Each gateway independently handles routing for its local subnet portion, distributing the processing load and eliminating the single point of failure and bottleneck associated with central handlers.
Solution Approach 2:
The patent implements dynamic routing where gateway machines can adaptively update routing tables and forward packets based on current network conditions. This dynamic approach replaces the static, error-prone central handler configuration with flexible, distributed routing decisions that reduce communication errors.
3Reliability
If subnet stretching is implemented across multiple distributed computing environments, then application availability is improved, but reconfiguration overhead and downtime increase
Solution Approach 1:
The patent establishes routing tables and gateway configurations in advance before migration is needed. Gateway machines are pre-configured with knowledge of stretched subnets and routing paths, allowing applications to be migrated between environments without reconfiguration downtime since the networking infrastructure is already prepared.
Solution Approach 2:
The patent creates a universal subnet stretching mechanism that works across different distributed computing environments (cloud, on-premises, hybrid) using standard IP routing principles. This universal approach allows applications to maintain the same network configuration regardless of deployment location, eliminating reconfiguration overhead during migrations.
4Adaptability or versatility
If layer-2 (L2) communications are used for subnet stretching, then broadcast and multicast functionality is maintained, but overhead and performance degradation occur
Solution Approach 1:
The patent replaces L2 broadcast and multicast mechanisms with L3 routing-based solutions. Instead of relying on L2 switching for broadcast functionality, the system uses L3 routing with appropriate routing table entries to achieve similar communication patterns, thereby eliminating L2 overhead while maintaining adaptability for various communication types.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
Systems and methods for stretching a subnet that do not require level 2 (L2) communications to be handled are provided. A user may gradually migrate VMs or applications instead of migrating an entire subnet at one time, may fail-over specific VMs without failing-over an entire subnet or renumbering IP addresses, may deploy applications to the cloud without the need to create a VPN, or may enable hybrid network connectivity without modifying routes or (re)configuring edge routers, among other benefits. The domains over which the subnet are stretched include a virtual gateway which is associated with the layer-3 (L3) addresses of the other domains. L3 communications within the domain are routed within that domain, and L3 communications within the subnet in another domain are intercepted by the local gateway, are passed to the remote gateway of the other domain, and are forwarded to the destination while leveraging L3 communications.