Stateful PCE Path Key Caching for Inter-Domain TE
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current path computation techniques in computer networks, particularly across autonomous systems (AS) and areas, face challenges in maintaining optimal path segments and diversity due to limited visibility and re-computation requirements, leading to inefficiencies and potential loss of promised path costs and diversity.
Innovation Solution
A stateful path computation element (PCE) model that stores computed path segments with associated path keys, allowing for efficient retrieval and reuse across domains, thereby eliminating the need for additional path computations and preserving path diversity and confidentiality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If path computation is performed for each domain boundary crossing, then path computation completeness is improved, but computation time and complexity increase
Solution Approach 1:
The patent performs path computation preliminarily at domain boundary routers before traffic actually needs to traverse the domain. Computed paths are cached and reused for subsequent traffic flows, eliminating the need for repeated computations at each boundary crossing.
Solution Approach 2:
The patent pre-computes and caches path segments within domains before they are needed. When traffic needs to cross domain boundaries, the pre-computed path segments are retrieved from cache rather than computed in real-time, significantly reducing computation time.
2Productivity
If path segments are cached at domain boundary routers, then retrieval efficiency is improved, but memory usage increases
Solution Approach 1:
The patent divides the path computation problem into segments, where each domain boundary router caches only the path segments relevant to its own domain. This segmentation distributes the storage burden across multiple routers rather than concentrating all path information in one location.
Solution Approach 2:
The patent extracts and caches only the essential path segment information (next-hop router and exit domain) at domain boundaries, rather than storing complete end-to-end paths. This extraction reduces the amount of data that needs to be cached while maintaining retrieval efficiency.
3Measurement precision
If complete paths are transmitted across domain boundaries, then path accuracy is improved, but information loss and security risks increase
Solution Approach 1:
The patent segments the complete path information, transmitting only the necessary next-hop router and exit domain identifiers across domain boundaries. The full path details remain within the source domain, maintaining accuracy for routing decisions while protecting confidential path information.
Solution Approach 2:
The patent extracts only the essential routing information (next-hop and exit domain) from the complete path and transmits this extracted subset across domain boundaries. The full path configuration remains private to the source domain, preventing information loss and security risks associated with transmitting complete paths.
Data Source
AI summary
A technique retrieves computed path segments across one or more domains of a computer network in accordance with a stateful (“semi-stateful”) path computation element (PCE) model. The stateful PCE model includes a data structure configured to store one or more path segments computed by a PCE in response to a path computation request issued by a Path Computation Client (PCC). Notably, each computed path segment stored in the data structure is identified by an associated path-key value (“path key”). The path segment and path key contents of the data structure are temporarily saved (“cached”) at a predetermined location in the network for a configurable period of time.