WAN Data Transfer Scheduling via Mixed Packing Algorithm
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wide area networks (WANs) face inefficiencies due to overlapping large data transfers, leading to excessive network taxation during peak times and underutilization of hardware resources during off-peak times, as mechanisms for time-shifting data transfers are lacking, forcing operators to over-provision costly hardware.
Innovation Solution
A controller computing device schedules data transfers based on multiple requests with deadlines, computing long-term and short-term schedules to optimize resource utilization, using a mixed packing and covering algorithm to determine data flow and resource allocation, and adjusts schedules dynamically to accommodate new requests and changes in network conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If data transfers are immediately initialized upon request, then responsiveness to user requests is improved, but network resources become excessively taxed during peak times
Solution Approach 1:
The system performs preliminary scheduling of data transfers by analyzing requested time periods and proactively shifting transfers to optimal times before actual execution. The scheduler computes schedules in advance that balance immediate responsiveness with future network load distribution, preventing excessive taxation during peak periods while maintaining user convenience.
Solution Approach 2:
The scheduling system dynamically adjusts data transfer timing based on network conditions and demand patterns. It continuously monitors network usage and modifies transfer schedules to optimize resource utilization, allowing the system to respond flexibly to changing conditions while maintaining both responsiveness and efficiency.
2Productivity
If data transfers are delayed to off-peak times, then network resource utilization is improved, but deadline compliance may be compromised
Solution Approach 1:
The system incorporates deadline constraints into the preliminary scheduling computation, ensuring that transfers with urgent deadlines are assigned to appropriate time slots before execution. By pre-calculating schedules that respect deadline requirements, the system maintains reliability while still achieving load balancing across network resources.
Solution Approach 2:
The scheduler applies different scheduling strategies to different data transfers based on their specific characteristics, particularly deadline urgency. Transfers with tight deadlines receive priority scheduling that ensures compliance, while less urgent transfers are shifted to off-peak times for optimal resource utilization, creating locally optimized solutions for each transfer.
3Reliability
If costly computer networking hardware is purchased to accommodate high-demand periods, then capacity to handle peak traffic is improved, but cost efficiency deteriorates due to underutilization during low-demand periods
Solution Approach 1:
The scheduling system performs preliminary analysis of demand patterns and network capacity to determine optimal transfer timing, enabling the network to handle peak traffic using existing resources rather than requiring additional hardware. By proactively distributing load across time periods, the system eliminates the need for over-provisioning expensive networking equipment.
Solution Approach 2:
The system changes the temporal parameters of data transfers, shifting when transfers occur rather than increasing network capacity. This parameter transformation allows the same physical infrastructure to handle variable demand efficiently, maintaining peak capacity while avoiding the costs associated with permanent hardware expansion.
4Productivity
If a centralized controller computes long-term schedules, then overall network optimization is improved, but computational complexity and processing time increase
Solution Approach 1:
The scheduling system segments the computation process into manageable components, analyzing different time periods and transfer types separately. This segmentation reduces the complexity of the overall optimization problem while maintaining global optimization benefits, allowing the centralized controller to handle large-scale scheduling without being overwhelmed by computational burden.
Solution Approach 2:
The system performs preliminary computations to establish baseline schedules and identifies critical constraints before executing the full optimization algorithm. By pre-processing demand data and identifying key scheduling parameters in advance, the system reduces the complexity of the main computation while achieving comprehensive network optimization.
Data Source
AI summary
Various technologies pertaining to scheduling network traffic in a network are described. A request to transfer data from a first computing device to a second computing device includes data that identifies a volume of the data to be transferred and a deadline, where the data is to be transferred prior to the deadline. A long-term schedule is computed based upon the request, wherein the long-term schedule defines flow of traffic through the network over a relatively long time horizon. A short-term schedule is computed based upon the long-term schedule, where devices in the network are configured based upon the short-term schedule.