Server Cluster Migration Scheduling Using Genetic Algorithms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In server network environments, balancing server load and transferring data for tenants is complex due to arbitrary tenant/server relationships leading to load imbalances, which conventional exhaustive enumeration methods are ineffective in solving efficiently.
Innovation Solution
A computer system with a tenant replacement manager that uses a genetic algorithm to schedule migration events for server resources within a server cluster, considering load balance and data transfer constraints, by generating and optimizing migration scheduling schemes to balance server load and minimize data transfer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If exhaustive enumeration of all tenant-server combinations is used to find global optimal placement, then load balance and data transfer optimization can be achieved, but the computational complexity becomes extremely high and the problem becomes unsolvable in practical time
Solution Approach 1:
The patent segments the tenant placement optimization problem into multiple independent server resources. Instead of enumerating all possible tenant-server combinations globally, the system processes each server resource independently, assigning tenants to servers one at a time based on current load states. This segmentation transforms the exponential complexity problem into a series of manageable sub-problems that can be solved efficiently using greedy algorithms with load balance and data transfer constraints.
2Ease of manufacture
If arbitrary assignment of tenant-server relationships is used, then the system is simple to implement, but load imbalance occurs which deteriorates network/server performance
Solution Approach 1:
The patent implements dynamic tenant placement adjustment rather than static arbitrary assignment. The system continuously monitors server load states and dynamically assigns or migrates tenants to servers based on real-time load balance conditions and data transfer requirements. This dynamic approach maintains implementation simplicity while ensuring performance stability through adaptive decision-making that responds to changing system states.
3Productivity
If tenant migration scheduling is performed without considering constraints, then the migration process is fast and simple, but load balance and data transfer requirements are not met
Solution Approach 1:
The patent changes the parameters used for tenant migration scheduling from simple speed-based criteria to multi-constraint optimization parameters. The system evaluates migration candidates based on load balance constraints (server resource availability, load distribution) and data transfer constraints (storage demand, migration cost). By incorporating these constraints into the scheduling decision parameters, the system achieves both efficient migration speed and precise constraint satisfaction through weighted evaluation functions.
Data Source
AI summary
In accordance with aspects of the disclosure, systems and methods are provided for scheduling migration events for server resources in a server cluster relative to a time interval while considering load balance constraints and data transfer constraints of the server cluster. The systems and methods may include generating one or more potential migration scheduling schemes for migration of each server resource to at least one other server resource in the server cluster within the time interval while considering the load balance constraints and the data transfer constraints of the server cluster, and generating a migration event schedule for each migration event within the time interval based on the one or more potential migration scheduling schemes generated for migration of each server resource to at least one other server resource in the server cluster.