Host Pool Desirability Index for Virtualized Resource Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud-based networks face challenges in managing compute capacity due to backwards-incompatible updates, which limit live migration of virtualized resources, leading to overwhelmed pools and reduced opportunities for maintenance or heat management.
Innovation Solution
The implementation of desirability indices to selectively manage the placement and migration of virtualized resources, where hosts with backwards-incompatible updates are deprioritized as target hosts, and resources are preferentially migrated to hosts with higher desirability indices based on update status, capacity, and compatibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hosts are updated with backwards-incompatible updates, then system reliability and security are improved, but virtualized resource migration capability deteriorates
Solution Approach 1:
The system performs preliminary actions by maintaining a pool of hosts at the previous version that are specifically reserved for receiving migrated virtualized resources. Before updating all hosts, the system ensures that updated hosts have reduced desirability indices, and preliminary migration paths are established to non-updated hosts, preventing migration deadlocks before they occur.
Solution Approach 2:
The system applies local quality by creating heterogeneous host pools with different update states. Specific hosts are maintained at different software versions based on their role and migration needs. The desirability index is locally adjusted for each host based on its update status, allowing the system to optimize migration decisions on a per-host basis rather than applying a uniform update policy across all hosts.
2Productivity
If virtualized resources are concentrated on updated hosts, then resource utilization efficiency is improved, but system maintainability deteriorates
Solution Approach 1:
The system segments the host fleet into multiple pools based on update status, with distinct functionality assigned to each segment. Updated hosts handle workloads that require new features or security fixes, while non-updated hosts serve as migration targets and maintenance buffers. This segmentation allows independent maintenance of different host groups without affecting the entire system.
Solution Approach 2:
The system implements dynamic host pool management where the composition and capacity of updated versus non-updated host pools changes over time. As hosts are updated, their desirability indices are dynamically adjusted, and the system continuously rebalances resource allocation. This dynamic approach allows the system to adapt resource distribution to current maintenance needs while maintaining high utilization efficiency.
3Reliability
If all hosts are updated simultaneously, then system security is improved, but operational stability deteriorates
Solution Approach 1:
The system implements periodic, staged updates rather than simultaneous updates across all hosts. Hosts are updated in batches or individually over time, with each update cycle followed by a stabilization period where migration desirability indices are adjusted. This periodic approach allows the system to incrementally improve security while maintaining operational stability through controlled transitions.
Solution Approach 2:
The desirability index system acts as an intermediary mechanism that mediates between updated and non-updated hosts during the transition period. It provides a quantitative framework for making balanced migration decisions that consider both security improvements and operational stability. The intermediary mechanism ensures smooth transitions by preventing over-concentration of resources on either updated or non-updated hosts.
Data Source
AI summary
Various techniques for managing heat and backwards-incompatible updates in cloud-based networks are described. In an example method, a virtualized resource, is identified. At least one first host may include an updated version of an element and at least one second host may include a previous version of the element. The updated version may be incompatible with the previous version. A first desirability index corresponding to the at least one first host may be less than a second desirability index corresponding to the at least one second host. The virtualized resource may be live-migrated from the source host to a target host among the at least one first host.