SDN Flow Entry Aggregation for Switch Storage Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of flow entries in switches due to application-aware routing in software-defined networking (SDN) exceeds storage capacity, necessitating a method to reduce the number of entries stored.
Innovation Solution
Classifying flow entries into partitions using indicators and employing bit merging or subset merging to compress entries, with each entry represented by ternary strings to indicate network requirements, allowing for adaptive threshold-based compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If application-aware routing function is implemented to allocate suitable network routes, then quality of service requirements are fulfilled, but the number of flow entries stored in switches exceeds storage capacity
Solution Approach 1:
The patent segments flow entries into multiple partitions based on matching criteria (such as prefix length, packet type, or other indicators). Each partition contains flow entries with similar characteristics, enabling efficient aggregation. This segmentation allows the system to manage large numbers of flow entries by organizing them into manageable groups that can be compressed together.
Solution Approach 2:
The patent merges multiple flow entries into aggregated flow entries by identifying common matching criteria across multiple entries. Using bit merging techniques, the system combines flow entries that share common prefix bits or matching patterns, creating a single aggregated entry that represents multiple original entries. This merging process significantly reduces the total number of flow entries stored in switches while preserving the ability to match incoming packets correctly.
2Quantity of substance
If flow entries are compressed using bit merging or subset merging, then storage capacity is optimized, but compression operations may be unnecessarily performed
Solution Approach 1:
The patent implements a dynamic threshold mechanism that adapts to current switch conditions. The threshold for triggering compression operations is not fixed but adjusts based on factors such as current storage usage, flow entry characteristics, and network traffic patterns. This dynamic approach ensures compression is performed only when beneficial, avoiding unnecessary compression operations that would waste processing time.
Solution Approach 2:
The system incorporates feedback mechanisms that monitor the effectiveness of compression operations and adjust future compression decisions accordingly. By tracking metrics such as compression ratio, matching accuracy, and processing overhead, the system learns from past operations and optimizes when to perform compression. This feedback loop prevents unnecessary compression operations by identifying conditions under which compression would be beneficial versus when it would be wasteful.
Data Source
AI summary
A flow entry aggregation method of a network system includes classifying a plurality of flow entries into a plurality of partitions according to a plurality of indicators of the plurality of flow entries, wherein each flow entry utilizes ternary strings to represent at least one field of the flow entry and the plurality of indicators are utilized to indicating network requirements corresponding to the plurality of flow entries; and utilizing bit merging or subset merging to compress the flow entries in the same partition.