SD-WAN Group Policy Determination via Inter-Node Query
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Software-Defined (SD)-Wide Area Networks (WANs), determining the identity of a destination group for routing data packets is challenging when this information is not readily available at the source network node.
Innovation Solution
The source network node sends a request for the destination group identifier to a second network node, which communicates with a local fabric control plane to retrieve and send back the necessary group identifier.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If the source network node has all destination group identities locally stored, then policy determination is fast and simple, but the network node requires large memory and the system is difficult to scale
Solution Approach 1:
The patent divides the group identity information into two segments: source group identities are stored locally at the source network node, while destination group identities are stored at the destination network node or controller. When a data packet needs policy determination, the source node queries the destination node or controller for the destination group identity if not locally available. This segmentation reduces the memory burden on individual nodes while enabling complete policy determination.
Solution Approach 2:
The patent introduces an intermediary mechanism (destination network node or controller) that stores and provides destination group identities upon request. Instead of every source node storing all possible destination group identities, the intermediary holds this information and responds to queries, reducing overall system memory requirements while maintaining fast policy determination capability.
2Quantity of substance
If the source network node queries the destination network node for group identity, then memory requirements are reduced, but additional communication overhead and latency are introduced
Solution Approach 1:
The patent implements a hybrid approach where the source network node attempts to determine policy using locally available source group identity first (partial action). Only when the destination group identity is not locally available does it initiate a query to the destination node or controller (excessive action). This partial/excessive action strategy balances memory efficiency with minimal communication overhead, as queries are performed only when necessary.
3Adaptability or versatility
If group identity information is distributed across multiple network nodes, then system scalability is improved, but the complexity of managing and synchronizing this information increases
Solution Approach 1:
The patent designates the destination network node or controller as an intermediary that maintains the authoritative store of destination group identities. This intermediary approach simplifies information management compared to a fully distributed system, as the intermediary is responsible for storing and providing this information on demand, reducing the complexity of synchronization and consistency management across the network.
Solution Approach 2:
The destination network node or controller maintains its own local store of destination group identities and serves requests from source nodes autonomously. This self-service mechanism allows each destination node to independently manage its group identity information without requiring complex centralized coordination, improving scalability while keeping management complexity localized.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one embodiment, a method includes receiving a data packet from a first host located in the first site, where the data packet may be destined to a second host located in a second site that may be different from the first site, determining that an identity of a second group to which the second host belongs is not available at the first network node, sending a request for an identifier of the second group to a second network node, where the request may comprise an address of the second host, receiving a response comprising the identifier of the second group from the second network node, determining that the second group is a destination group, applying one or more policies associated with the destination group to the data packet, and causing the data packet to be routed to the second host.