Placement Policy-Based Computing Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current distributed computing environments lack user control over resource provisioning, leading to automation, performance, and usability issues, particularly in IaaS services, where users cannot specify resource location or grouping, resulting in security, performance, and isolation challenges.
Innovation Solution
A resource management system that allows users to define policies for provisioning computing resources based on user preferences, including geographical location, security configurations, and performance considerations, enabling fine-grained control over resource allocation and management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If computing resources are automatically provisioned without user control, then provisioning speed is improved, but user control and customization are lost
Solution Approach 1:
The system dynamically adjusts resource provisioning between automated and user-controlled modes based on service type and user preferences. Placement policies can be dynamically created, modified, and deleted to adapt to changing requirements while maintaining both automation efficiency and user control where needed.
Solution Approach 2:
The resource provisioning system is segmented into different policy types (placement policies, affinity policies, anti-affinity policies) that can be independently configured. This allows specific aspects of resource allocation to be automated while others remain under user control, resolving the contradiction between automation and customization.
2Device complexity
If computing resources are pre-created in static groups, then resource allocation is simplified, but adaptability to changing demand is reduced
Solution Approach 1:
Resource groups are transformed from static to dynamic structures that can be automatically created, modified, and dissolved based on service demand. The system continuously monitors resource utilization and automatically adjusts group compositions without requiring manual intervention or fixed pre-configuration.
Solution Approach 2:
The system implements self-service resource provisioning where the infrastructure automatically creates and manages resource groups based on service requests and demand patterns, eliminating the need for manual group management while maintaining adaptability to changing requirements.
3Productivity
If resources are shared in multi-tenant environments, then resource utilization efficiency is improved, but security and isolation concerns increase
Solution Approach 1:
The system segments resource sharing into different levels using affinity and anti-affinity policies. Resources can be shared with specific tenants or groups of tenants while maintaining isolation from others. This selective segmentation enables controlled sharing that improves utilization while preserving security and isolation requirements.
Solution Approach 2:
Different isolation and sharing characteristics are applied to different resource groups based on security requirements and tenant preferences. The system creates localized sharing arrangements where resources are shared within specific tenant groups while maintaining strict isolation from other groups, achieving both efficiency and security.
Data Source
AI summary
Techniques are disclosed for managing and allocating resources based on resource policies in response to user requests. A resource management system can receive a request for a service. A request may indicate preferences for allocation resources (e.g., a resource definition) to enable the service. A resource definition may indicate a topology of the computing resources to allocate for the user. The topology may indicate what computing resources to allocate and how to allocate those computing resource. Based on the information indicated by a request, the resource management system may determine a placement policy for allocation of computing resources indicated by the request. A placement policy may indicate a placement of one or more computing resources requested by a user. The placement policy may indicate where and how computing resources are to be placed once allocated. The computing resources may be allocated based on the placement policy and the resource definition.