Geo-Distributed Coflow Routing Under WAN and Location Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for data center communication networks fail to efficiently manage and optimize the transmission of data across geographically distributed networks, particularly in addressing bottlenecks and compliance with location-dependent constraints in optimizing the flow completion times and bandwidth utilization overheads.
Innovation Solution
Implementing a technical solution by mapping and optimizing the coflow to a plurality of geo-distributed computer sites via wide area network, subject to location-dependent constraints, using a mathematical optimization to identify multiple candidate data paths and adjust bandwidth to ensure efficient flow completion times and resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If coflows are transmitted across geo-distributed computer sites via WAN, then data processing can be distributed across multiple locations, but total flow completion time increases due to network distance and bandwidth limitations
Solution Approach 1:
The patent segments the coflow into multiple independent flows that can be transmitted through different network paths simultaneously. Each flow is routed independently through optimal paths identified by the mathematical optimization, allowing parallel transmission and reducing overall completion time despite geo-distribution
Solution Approach 2:
The patent introduces multiple routing dimensions by identifying multiple candidate data paths between source-destination pairs. The mathematical optimization selects paths across different network dimensions (multiple routes, different hops), transforming the single-path limitation into a multi-dimensional path selection problem that reduces completion time
2Loss of time
If multiple candidate data paths are identified and mathematical optimization is performed, then flow completion time is minimized, but computational complexity increases
Solution Approach 1:
The patent performs preliminary identification of multiple candidate data paths before the actual data transmission. The mathematical optimization is executed in advance to determine the optimal path set, allowing the actual transmission to proceed without real-time computation overhead, thus minimizing flow completion time while managing complexity through pre-computation
3Reliability
If location-dependent constraints are enforced for geo-distributed sites, then data sovereignty and compliance requirements are met, but available bandwidth and routing options are reduced
Solution Approach 1:
The patent applies local quality constraints by enforcing location-dependent rules at each geo-distributed site. The mathematical optimization incorporates these local constraints into the path selection process, ensuring that data routing respects location-specific requirements while still achieving optimal overall flow completion time within the constrained environment
Data Source
AI summary
A coflow is mapped to a plurality of geo-distributed computer sites that can communicate via wide area network (WAN), where the mapping is subject to one or more location-dependent constraints. Multiple candidate data paths are identified for each of a plurality of source-destination pairs of the plurality of geo-distributed computer sites. A mathematical optimization is performed to find a set of paths from the candidate data paths based on total flow completion time and at least one additional objective of the coflow.