Distributed Network Scheduling via Token-Based Circuit Switch Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data center network scheduling methods, such as dynamic and fixed scheduling, face inefficiencies due to reliance on centralized controllers for traffic demand prediction and resource allocation, leading to potential over-allocation of circuit switch resources and reduced network throughput.
Innovation Solution
A distributed reconfigurable data center network (RDCN) scheduling method that determines traffic demands independently at each electronic device using count-min sketch (CMS) and allocates circuit switch resources efficiently by using a token-based system to avoid redundant allocations and optimize circuit switch usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If dynamic scheduling is used to obtain traffic demand information from a centralized network controller, then optimized scheduling can be performed, but it takes a long time to obtain traffic demand information and calculate optimal schedule
Solution Approach 1:
The centralized scheduling function is segmented and distributed to individual ToR switches. Each switch independently determines its own traffic demand using local CMS data structures, eliminating the need to query a centralized controller for traffic information. This segmentation resolves the contradiction by enabling fast local decision-making while maintaining scheduling optimization through distributed intelligence.
Solution Approach 2:
Each ToR switch performs self-service by autonomously determining its own traffic demand using locally maintained CMS data structures. The switches independently calculate their burst levels and make scheduling decisions without external intervention, thus eliminating the time-consuming process of obtaining traffic demand information from a centralized controller while still achieving optimized scheduling.
2Loss of time
If fixed scheduling method is used to perform repetitive round robin connections, then time burden for predicting traffic demands is reduced, but unnecessary waste of circuit switch resources occurs
Solution Approach 1:
The system implements dynamic scheduling at the distributed level, where each ToR switch continuously monitors its traffic demand using CMS data structures and adjusts its circuit switch allocation dynamically. This replaces the static round-robin approach with a dynamic mechanism that adapts to actual traffic conditions, reducing resource waste while maintaining fast decision-making.
Solution Approach 2:
Each ToR switch changes its scheduling parameters (burst level, traffic demand thresholds) based on real-time traffic conditions observed in its CMS data structures. This parameter adaptation allows the system to avoid fixed round-robin allocations and instead allocate circuit switch resources dynamically according to actual traffic needs, reducing waste while keeping prediction overhead low.
3Productivity
If distributed scheduling is used to avoid centralized controller intervention, then controller load is reduced, but circuit switch resource allocation duplication may occur
Solution Approach 1:
The system implements feedback mechanisms where each ToR switch monitors its traffic demand using CMS data structures and adjusts its circuit switch allocation based on observed traffic patterns. This distributed feedback loop enables autonomous decision-making without centralized control while preventing resource allocation conflicts through local awareness of traffic conditions and coordinated token-based access.
Solution Approach 2:
A token-based intermediary mechanism is introduced to coordinate circuit switch access among distributed ToR switches. The token acts as a mediator that grants exclusive access rights to the circuit switch, preventing simultaneous allocations and resource conflicts. This intermediary enables distributed scheduling while maintaining resource allocation reliability through structured coordination.
4Productivity
If token-based system is used to distribute traffic demand to circuit switch, then resource allocation is optimized, but complexity of coordination increases
Solution Approach 1:
The token-based coordination mechanism serves multiple functions simultaneously: it grants circuit switch access rights, coordinates traffic demand distribution, and prevents resource allocation conflicts. This multi-functional token system optimizes resource allocation efficiency while avoiding the need for separate complex coordination protocols, thereby limiting the increase in overall system complexity.
Data Source
AI summary
An electronic device for distributed scheduling of a network and a method of operating the same are provided. The electronic device includes a processor and a memory configured to store instructions, wherein the instructions, when executed by the processor, may cause the electronic device to determine a traffic demand for a network connection, based on the traffic demand, determine a burst level for traffic of a predetermined time period and transmit the burst level to a control device, by comparing the burst level with a threshold value determined by the control device, determine whether the traffic demand is in a traffic over-demand state, receive a token for connecting to a circuit switch to distribute the traffic demand, and when the traffic demand is in the traffic over-demand state, connect to the circuit switch through the token.