Passive Reachability Measurement for Inline Service Chains
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for verifying the correctness of inline service chains in networks are inadequate, as they often rely on active packet injection, which can confuse or drop packets, leading to performance degradation and incorrect path verification.
Innovation Solution
A method that passively observes data flows by creating copies of packets at switches and forwarding them to a controller for path correctness testing, allowing for the reconstruction of the path traversed by the flow and verification of reachability without impacting the middleboxes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If active packet injection is used to test network path reachability, then path verification can be performed, but the middleboxes may drop the injected packets or their internal states may be confused, leading to performance degradation
Solution Approach 1:
The patent uses packet copying (mirroring) at switches to obtain copies of data packets traversing the service chain. These copies are sent to the controller for path verification without the original packets being affected. This resolves the contradiction by enabling measurement without injecting test packets that could disrupt middlebox operations.
Solution Approach 2:
The patent introduces switches as intermediary devices that mirror packets to the controller. The controller acts as a mediator that receives packet copies and performs path verification independently, without directly interacting with the middleboxes. This intermediary approach allows measurement while protecting the reliability of inline services.
2Measurement precision
If packet copies are mirrored to the controller for path verification, then accurate path monitoring is achieved, but control channel overhead increases
Solution Approach 1:
The patent applies local quality by selectively mirroring packets based on their relevance to path verification. Not all packets need to be copied - the system can filter and select only those packets that are necessary for verifying service chain paths, thereby reducing control channel overhead while maintaining monitoring accuracy.
Solution Approach 2:
The patent implements partial action by mirroring only a subset of packets rather than all packets in the service chain. This selective approach provides sufficient information for path verification without overwhelming the control channel, balancing measurement precision with resource conservation.
Data Source
AI summary
A method is implemented by a computing device in communication with a network having a plurality of switches. The method performs path correctness testing where the path defines a service chain associated with a flow of data. The method configures the plurality of switches using a flow control protocol. The method includes the steps of selecting a switch from the path to configure path correctness testing, sending a configuration request to the selected switch to instate at least one rule for testing path correctness, receiving copies of flow packets from at least one switch in the path, and comparing the copies of flow packets with a service chain configuration of the flow to confirm path correctness.