Packet Event Tracking via Set Membership Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network troubleshooting in Software-Defined Networking (SDN) environments is challenging due to the scale and complexity, making it difficult to identify specific issues such as packet loss across hundreds or thousands of logical routers and switches.
Innovation Solution
Implementing a logical forwarding element that uses a set membership filter, such as a Bloom filter, to track packet events and determine if specific packets have been dropped by updating and querying packet signatures across the network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional network monitoring methods are used in SDN environments, then network traffic can be monitored, but it becomes difficult to identify specific packet loss issues across hundreds or thousands of logical routers and switches due to scale and complexity
Solution Approach 1:
The patent segments the network troubleshooting function by introducing set membership filters at individual logical forwarding elements. Each filter independently tracks packet events for specific packet sets, dividing the complex global troubleshooting problem into manageable local tracking tasks. This segmentation enables precise packet loss identification without requiring centralized analysis of entire network traffic.
Solution Approach 2:
The patent introduces set membership filters as intermediary structures between packet forwarding and troubleshooting analysis. These filters act as mediators that track packet events and provide troubleshooting information without interfering with normal packet forwarding operations. The filters translate complex network states into simplified membership queries that can be efficiently answered.
2Measurement precision
If set membership filters like Bloom filters are implemented to track packet events, then packet loss identification accuracy improves, but memory usage and filter maintenance overhead increase
Solution Approach 1:
The patent employs Bloom filters, which are space-efficient probabilistic data structures that use minimal memory to track packet memberships. While Bloom filters have probabilistic properties and may have false positives, they provide sufficiently accurate packet event tracking at very low memory cost, making them ideal for deployment at numerous logical forwarding elements without significant resource burden.
Solution Approach 2:
The patent changes the parameter representation from tracking individual packets to tracking packet signatures and event types. By hashing packet characteristics into compact representations and using set membership filters to track these transformed parameters, the system achieves accurate packet event tracking with reduced memory requirements compared to storing complete packet information.
3Ease of operation
If packet event tracking is implemented across all logical forwarding elements, then network troubleshooting capability improves, but the system complexity and overhead of managing filters across the network increases
Solution Approach 1:
The patent implements self-service mechanisms where logical forwarding elements autonomously maintain their own set membership filters and generate troubleshooting information locally. Each forwarding element independently tracks packet events and can respond to troubleshooting queries without requiring centralized filter management or coordination, significantly reducing system complexity while maintaining comprehensive tracking capability.
Solution Approach 2:
The patent establishes feedback mechanisms where set membership filters continuously monitor packet events and provide real-time troubleshooting information back to the control plane or monitoring systems. This feedback loop enables dynamic packet loss identification and network issue detection without requiring manual intervention or complex centralized analysis.
Data Source
AI summary
Example methods and systems are provided for packet event tracking. One example method may comprise: in response to a logical forwarding element detecting a packet event for a first packet, determining a first packet signature of the first packet; and tracking the packet event by updating a set membership filter based on the first packet signature to indicate that the first packet is a member of a set associated with the packet event. The method may also comprise: in response to receiving a query as to whether the packet event has been detected and tracked for a second packet, determining a second packet signature of the second packet. Based on the second packet signature, the set membership filter may be applied to determine whether the second packet is a possible member of the set associated with the packet event.