VM Load Balancing via Storage Controller Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current load balancing systems for virtual machines in clustered computing environments fail to account for storage system load factors, leading to bottlenecks and inefficient resource distribution.
Innovation Solution
A system and method that generate storage system load factor information and utilize it to redistribute virtual machines across physical processors, incorporating a load balancer and storage load factor information generator to balance the load based on storage controller performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If load balancing systems only use processor load information to distribute virtual machines, then the system complexity is reduced and ease of operation is improved, but storage system bottlenecks are not detected and productivity deteriorates
Solution Approach 1:
The load balancing system implements feedback by continuously monitoring storage system load information from storage controllers and using this feedback to dynamically adjust virtual machine distribution. The system collects storage load metrics, compares them against thresholds, and redistributes virtual machines accordingly to prevent storage bottlenecks while maintaining operational simplicity.
Solution Approach 2:
The patent introduces storage load information as an intermediary element between the virtual machines and storage controllers. This intermediary provides detailed storage system status data to the load balancing component, enabling informed decision-making about virtual machine placement without increasing operational complexity for end users.
2Productivity
If storage system load information is integrated into load balancing, then storage-related bottlenecks are detected and productivity is improved, but device complexity increases
Solution Approach 1:
The load balancing component is designed with multi-functionality, serving both traditional processor load balancing and the new storage load balancing functions. By integrating storage monitoring capabilities into the existing load balancer, the system avoids adding separate dedicated hardware or software components, thereby limiting the increase in overall system complexity.
Solution Approach 2:
The patent merges the storage load monitoring function with the existing load balancing component. Instead of creating a separate storage management system, the storage load information generator and load balancer are combined into an integrated system that handles both processor and storage considerations in a unified manner.
3Productivity
If virtual machines are redistributed based on storage load information, then storage resource optimization is achieved and productivity is improved, but loss of information about storage controller status increases without proper monitoring
Solution Approach 1:
The system performs preliminary action by proactively monitoring storage controller load information before bottlenecks occur. The storage load information generator continuously collects storage metrics and makes this information available to the load balancer in advance, enabling preventive redistribution of virtual machines before performance degradation happens.
Solution Approach 2:
The patent implements continuous feedback loops where storage controllers report their load status to the storage load information generator, which in turn provides this information to the load balancer. This feedback mechanism ensures that no critical storage status information is lost and enables real-time adaptive decision-making for virtual machine placement.
Data Source
AI summary
Methods and systems for generating storage related load factor information for load balancing of multiple virtual machines operable in a cluster of multiple physical processors (such as a blade center). Load factor information is generated within a storage system relating to operation of the storage system as a whole and relating to each of multiple storage controllers in the storage system. The information so generated in the storage system is communicated to a load balancing element associated with the multiple virtual machines. The load balancing element then utilizes the storage related load factor information, optionally in combination with other load factor information, to distribute or redistribute the operation of the multiple virtual machines over the plurality of physical processors.