Cloud Scheduler Framework for Multi-Domain Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud computing resource allocation systems struggle to efficiently manage applications by balancing multiple domains such as consumable resources, networking, location, and availability constraints, leading to inadequate placement solutions that do not adapt to changing objectives and constraints over time.
Innovation Solution
A cloud computing allocation system and methodology that introduces Allocation Domains (ADs) managed by AD Agents, dynamically creating and modifying allocation policies, and utilizing a cloud scheduler framework with biasing functions to optimize placement across various domains, ensuring compliance with user and provider objectives and constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple allocation domains (consumable resources, networking, location, availability) are handled independently by traditional resource allocators, then each domain can be managed with simple allocation rules, but the system cannot satisfy all requirements and constraints across all domains simultaneously
Solution Approach 1:
The patent segments the allocation system into multiple independent Allocation Domain Agents (ADAs), each responsible for a specific allocation domain (consumable resources, networking, location, availability). Each ADA evaluates placement solutions against its domain's requirements and constraints independently, then the combined evaluation determines the final placement decision. This segmentation allows the system to handle multiple domains simultaneously while maintaining manageable complexity through modular design.
Solution Approach 2:
The patent introduces an intermediary component that coordinates between multiple ADAs and the placement solution. This intermediary aggregates the evaluation results from different allocation domains and synthesizes a final placement decision that satisfies all domain requirements. The intermediary acts as a mediator that reconciles potentially conflicting constraints from different domains without requiring a complete system redesign.
2Adaptability or versatility
If placement policies are fixed and predetermined, then the allocation system operates efficiently with simple rules, but it cannot adapt to changing objectives and constraints over time
Solution Approach 1:
The patent implements dynamic allocation policies where each Allocation Domain Agent can modify its evaluation criteria and constraints in real-time based on changing requirements. The system allows objectives and constraints to be updated without requiring complete system reconfiguration, enabling adaptive response to changing conditions while maintaining operational efficiency through incremental policy adjustments rather than complete reevaluations.
Solution Approach 2:
The patent incorporates feedback mechanisms where the allocation system continuously monitors placement outcomes and constraint satisfaction across all domains. This feedback informs dynamic policy adjustments, allowing the system to learn from past placements and adapt to changing objectives. The feedback loop enables automatic policy modification based on observed performance and changing requirements, reducing manual intervention time.
3Productivity
If traditional bin packing algorithms are used for placing virtual machines on physical machines, then resource utilization is optimized, but communication requirements and location constraints among virtual entities are not satisfied
Solution Approach 1:
The patent segments the evaluation process into separate allocation domains, with dedicated ADAs for consumable resources, networking, location, and availability. The consumable resources ADA handles traditional bin packing optimization for resource utilization, while separate ADAs independently evaluate networking constraints (bandwidth, delay), location constraints (proximity requirements), and availability constraints. This segmentation allows each domain to be optimized independently without compromising other requirements.
Solution Approach 2:
The patent merges the results from multiple independent ADA evaluations into a comprehensive placement decision. The system combines the resource utilization assessment from the consumable resources ADA with the communication requirements assessment from the networking ADA, location requirements from the location ADA, and availability from the availability ADA. This merging ensures that the final placement solution simultaneously satisfies all constraints while maintaining high resource utilization efficiency.
Data Source
AI summary
There are provided a system, a method and a computer program product for operating a cloud computing infrastructure. In one embodiment, the system and method performs allocation domain modeling and provides a cloud scheduler framework that takes as input desired optimization objectives and the workload constraints and efficiently produces a placement solution that satisfies the constraints while optimizing the objectives in a way that adjusts itself depending on the objectives. As the objectives change, e.g., due to actions from system administrators or due to changes in business policies, the system optimizes itself accordingly and still produces efficient and optimized placement solutions. The system and method constructs an Allocation Domain (AD) that is a particular facet for allocating a logical entity to a physical entity. An AD is created using: variables, functional definitions (functions of variables), and a policy specification that includes a Boolean expression (of the functional definitions).