Open-Ended Policies for Virtual Instance Provisioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing process of creating separate storage policies for each data store's unique capabilities leads to an unmanageably large number of policies in systems with hundreds or thousands of virtual machines, making it impractical and cost-prohibitive to accommodate all capabilities.
Innovation Solution
Implementing open-ended storage policies that include rules with undefined values, allowing administrators to select user-defined values for each virtual computing instance, reducing the number of policies needed and enabling a single policy to serve multiple VM configurations across disparate data stores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If separate storage policies are created for each data store's unique capabilities, then each data store can be properly managed according to its specific requirements, but the number of policies becomes unmanageably large in systems with hundreds or thousands of VMs
Solution Approach 1:
The patent implements a universal storage policy framework where a single policy can be applied across multiple data stores with different capabilities. The policy engine dynamically adapts the universal policy to each data store's specific characteristics, eliminating the need to create separate policies for each data store while maintaining proper management of unique requirements.
Solution Approach 2:
The system changes parameters dynamically by allowing a single storage policy to have variable parameters that are resolved at provisioning time based on the target data store's capabilities. This enables the same policy to accommodate different data stores by substituting appropriate values for parameters like replication factor, storage tier, or access patterns, rather than requiring separate policies for each configuration.
2Reliability
If separate storage policies are created for each VM configuration, then specific requirements can be met, but it becomes impractical and cost-prohibitive to manage all the different policies
Solution Approach 1:
The patent creates a universal policy template that can serve multiple VM configurations by parameterizing the policy rules. Administrators define once a policy with placeholder parameters, and the system automatically instantiates appropriate policy instances for different VMs by substituting the parameters with values matching each VM's requirements, ensuring reliability while simplifying management.
Solution Approach 2:
The system enables self-service policy instantiation where the policy engine automatically resolves parameters and creates appropriate policy instances based on VM characteristics and data store capabilities, eliminating the need for administrators to manually create and manage separate policies for each VM configuration.
3Stability of the object's composition
If traditional storage policies with fixed values are used, then consistency is maintained, but flexibility to accommodate new data store capabilities is reduced
Solution Approach 1:
The patent transforms static, fixed-value policies into dynamic policies where parameters are not hardcoded but are resolved at runtime based on the target data store's capabilities. This allows the policy structure to remain consistent and stable while the parameter values adapt dynamically to accommodate new data store capabilities, ensuring both consistency and flexibility.
Solution Approach 2:
The system performs preliminary definition of policy structure and constraints while leaving specific parameter values to be determined later during provisioning. This preliminary action establishes the consistent policy framework in advance, while the actual parameter resolution occurs dynamically when needed, allowing accommodation of new capabilities without compromising structural consistency.
Data Source
AI summary
The systems described herein are configured to reduce the number of resource policies created and stored on a cluster for provisioning and/or managing virtual computing instances (VCIs) utilization of one or more resources. An open-ended VCI policy including at least one open-ended rule having an undefined value and a resource is selected. A set of valid values compatible with a selected resource corresponding to each open-ended rule is presented to the user. The user selects a valid value for each open-ended rule from the set of valid values via a user interface (UI). The selected valid values are assigned to each open-ended rule to create a complete VCI open-ended policy. The complete VCI open-ended policy is applied to provision one or more VCIs. The same open-ended VCI policy may be applied to provision different VCIs by assigning a different set of user selected valid values.