Dynamic Cloud Network Control via Utility Weights
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed cloud networks face challenges in dynamically adapting to changing service demands and network conditions due to centralized optimization complexities and significant network delays, especially in lightly loaded scenarios, while existing solutions provide only expected time-averaged performance guarantees.
Innovation Solution
The proposed dynamic cloud network control algorithm incorporates a Lyapunov drift-plus-penalty method with a shortest transmission-plus-processing distance bias, which computes processing utility weights based on queue backlogs and processing costs to optimize resource allocation and flow rate assignments, ensuring probability-1 time average cost and occupancy bounds, reducing network delay without compromising throughput or overall cloud network cost.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If centralized optimization is used to allocate resources and route flows in distributed cloud networks, then service demands can be met with minimum cost, but network delays become significant and system complexity increases
Solution Approach 1:
The patent segments the centralized optimization problem into distributed decision-making at individual cloud locations. Each location independently computes processing utility weights based on local queue backlogs and processing costs, eliminating the need for centralized coordination and reducing network delay while maintaining service demand fulfillment.
Solution Approach 2:
The patent implements local quality by having each cloud location make autonomous decisions based on local information (queue backlogs and processing costs). The processing utility weight computation is performed locally at each node, allowing rapid adaptation without waiting for centralized optimization, thus reducing network delay while meeting service demands.
2Loss of time
If distributed dynamic control is implemented to reduce network delay, then response time improves, but achieving optimal resource allocation and cost minimization becomes more difficult
Solution Approach 1:
The patent changes the control parameter from centralized optimization variables to local processing utility weights computed from queue backlogs and processing costs. This parameter transformation enables distributed decision-making that reduces network delay while the mathematical structure of utility weights ensures optimal resource allocation and cost minimization are achieved.
Solution Approach 2:
The patent implements feedback through the processing utility weight computation, which continuously uses local queue backlog information and processing costs to make real-time resource allocation decisions. This feedback mechanism enables distributed control to achieve optimal performance without centralized coordination, reducing network delay while maintaining allocation efficiency.
3Reliability
If existing control algorithms are used, then expected time-averaged performance guarantees are provided, but probability-1 bounds and delay reduction in lightly loaded scenarios are not achieved
Solution Approach 1:
The patent implements dynamic control by having each cloud location continuously compute processing utility weights based on real-time local queue backlogs and processing costs. This dynamic distributed approach provides probability-1 time average cost and occupancy bounds while reducing network delay in lightly loaded scenarios, improving upon static expected time-averaged performance guarantees.
Data Source
AI summary
Various exemplary embodiments relate to a network node in a distributed dynamic cloud, the node including: a memory; and a processor configured to: observe a local queue backlog at the beginning of a timeslot, for each of a plurality of commodities; compute a processing utility weight for a first commodity based upon the local queue backlog of the first commodity, the local queue backlog of a second commodity, and a processing cost; where the second commodity may be the succeeding commodity in a service chain; compute an optimal commodity using the processing utility weights; wherein the optimal commodity is the commodity with the highest utility weight; assign the number of processing resource units allocated to the timeslot to zero when the processing utility weight of the optimal commodity is less than or equal to zero; and execute processing resource allocation and processing flow rate assignment decisions based upon the optimal commodity.