Programmable Packet Processing Graph for Dynamic Network Flow Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network architectures struggle to dynamically manage network resources and packet flows in virtualized environments, requiring reconfiguration of VLANs and network infrastructure, which is difficult with traditional equipment where control logic is co-located with switching logic, limiting flexibility in managing server-to-server traffic changes.
Innovation Solution
A programmable packet data processing system with a reconfigurable packet processing component arranged in a graph structure, allowing nodes to forward packets based on a graph configuration responsive to controller requests, enabling dynamic creation, update, and deletion of packet processing resources, and using match-action classifiers for packet processing functions like validation, modification, and multiplexing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If control logic is co-located with switching logic in conventional network equipment, then device structure is simplified, but flexibility in dynamically managing network resources and packet flows deteriorates
Solution Approach 1:
The patent segments the network device into distinct control plane and data plane components. The control logic is separated from the switching logic, allowing independent configuration and management. This segmentation enables flexible reconfiguration of packet flows without affecting the underlying device structure, resolving the contradiction between simplified structure and management flexibility.
Solution Approach 2:
The patent introduces a flow manager as an intermediary component that mediates between the control logic and switching logic. This intermediary enables dynamic configuration of packet flows by translating high-level flow specifications into low-level switching operations, providing flexibility while maintaining clear separation of concerns between control and data planes.
2Reliability
If traditional VLAN configuration methods are used in virtualized environments, then network security is maintained, but ability to dynamically reconfigure network resources deteriorates
Solution Approach 1:
The patent implements dynamic flow configurations that can be modified in real-time without reconfiguring underlying VLAN structures. Flow rules can be added, removed, or modified dynamically to accommodate changing virtual machine placements and traffic patterns, while VLAN security boundaries remain static and intact, resolving the contradiction between security stability and configuration flexibility.
Solution Approach 2:
The patent establishes VLAN security configurations in advance as static frameworks, then uses dynamic flow rules to handle specific traffic management needs. This preliminary configuration of security boundaries allows subsequent flexible adjustments to packet flows without compromising the established security architecture.
3Measurement precision
If flow tables with multiple entries are used to manage packet flows, then packet classification accuracy is improved, but processing time and system complexity increase
Solution Approach 1:
The patent performs flow classification and matching operations in advance, caching results and pre-computing flow paths. By preparing flow tables with pre-evaluated rules and using flow managers to anticipate traffic patterns, the system reduces real-time processing requirements while maintaining high classification accuracy through pre-established matching criteria.
Data Source
AI summary
Embodiments of the present disclosure relate to software-defined networks, and particularly, but not exclusively to programmable packet data processing systems, methods and computer readable products for use therein.


