In-Place Policy Updates via Ordered-Shifting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for updating policies on network devices often require significant memory resources and disrupt active policies, leading to inefficient utilization of limited memory and prolonged disruption windows.
Innovation Solution
A method and system for performing in-place policy updates on network devices by receiving a class modification batch, generating an operation commit sequence, and committing updates using a policy update programmer to minimize memory usage and disruption time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If make-before-break methodology is used for policy updates, then policy disruption is minimized, but memory resources are inefficiently utilized
Solution Approach 1:
The patent segments the policy update process into distinct phases: parsing the batch, validating memory sufficiency, generating a target class sequence, creating an operation commit sequence, and executing updates. This segmentation allows the system to reserve memory only for the duration of the update operation rather than maintaining duplicate policy sets, thus improving memory utilization while maintaining reliability through phased execution.
Solution Approach 2:
The patent performs preliminary actions by first parsing the batch, validating memory sufficiency, and generating the target class sequence before executing the actual policy updates. This preliminary processing allows the system to plan the update sequence and reserve memory efficiently, avoiding the need for continuous memory reservation during the update process.
2Reliability
If make-before-break methodology is used for policy updates, then policy disruption is minimized, but update speed is reduced
Solution Approach 1:
The patent implements dynamic memory reservation based on the actual update requirements. Instead of reserving a fixed, substantial portion of memory, the system dynamically allocates memory based on the batch size and update operations. This dynamic approach enables faster updates by reducing memory reservation time and overhead, while maintaining policy disruption minimization through controlled update sequencing.
Solution Approach 2:
The patent changes the parameter of memory reservation from a static, substantial portion to a dynamic allocation based on operational needs. By adjusting memory allocation parameters according to the actual update requirements, the system achieves faster update speeds while maintaining reliability through controlled policy application sequencing.
3Reliability
If substantial memory resources are reserved for policy updates, then update reliability is improved, but available memory for other functions is reduced
Solution Approach 1:
The patent extracts the memory reservation requirement from the general memory pool and ties it specifically to the update operation duration. By extracting the memory reservation to only when needed during the update process rather than maintaining it continuously, the system improves availability of memory for other functions while maintaining update reliability through targeted memory allocation during the update phases.
Solution Approach 2:
The patent implements a memory recovery mechanism where the reserved memory is released after the update operation completes. This discard and recover approach ensures that memory is not permanently consumed by update operations, thereby maintaining availability for other network functions while ensuring update reliability during the operation window.
Data Source
AI summary
A method and network device for enabling in-place policy updates. Specifically, the method and network device disclosed herein entail effecting the removal and/or insertion of policies on the network device while minimizing a window in which active policies may be disrupted by the policy updates. Further, in-place policy updates call upon the ordered-shifting of content stored across data structure elements. Through ordered-shifting, policy updates can utilize limited memory resources, on the network device, more efficiently over existing methodologies. Additionally, policy updates can also be committed faster, thereby allowing more policy updates to be processed with a fixed time window.