Lock-lease management for data center maintenance scheduling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Large data centers face challenges in scheduling upgrades of hardware and software resources without causing service disruptions, especially as the number of resources grows exponentially, leading to conflicts and potential failures due to insufficient resources during maintenance operations.

Innovation Solution

A platform and framework that enables simultaneous and safe execution of data center hardware and application maintenance, deployment, and optimization operations by using a lock-lease approach and calculating effective application availability, distinguishing between stateless and stateful applications, and incorporating machine learning to determine safe application availability limits.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If upgrades are scheduled to reduce service disruption, then service availability is improved, but the scheduling complexity and resource coordination requirements increase significantly

Engineering Contradiction:

Improveservice availabilityVSAvoidscheduling complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces a lock manager as an intermediary component that mediates between upgrade operations and application instances. The lock manager assigns locks to upgrade operations and manages their lifecycle, preventing conflicts without requiring complex manual scheduling. This intermediary abstraction simplifies the scheduling complexity while ensuring service availability through automated coordination.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements feedback mechanisms where application instances report their availability status to the lock manager, and the lock manager adjusts lock assignments based on real-time availability information. This feedback loop enables dynamic scheduling that adapts to changing system conditions, improving service availability while reducing the need for complex predetermined schedules.

Inventive Principle:

Principle #23Feedback

2Quantity of substance

If the number of resources scales up to tens of thousands or hundreds of thousands, then computing capacity is improved, but the likelihood of upgrade conflicts and service disruptions increases

Engineering Contradiction:

Improvenumber of resourcesVSAvoidservice stability

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The patent segments the large-scale resource management problem into finer granules by assigning individual locks to specific application instances or groups. Instead of managing all resources as a single unit, the system divides upgrade operations into discrete, manageable segments that can be coordinated independently. This segmentation reduces the complexity of managing conflicts across tens of thousands of resources while maintaining service stability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies preliminary action by having the lock manager pre-assign locks to upgrade operations before they execute, and by having application instances pre-report their availability status. This preliminary coordination prevents conflicts from arising in the first place, enabling the system to scale to large numbers of resources without proportionally increasing service disruption risk.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If traffic failouts are used to handle critical situations, then critical upgrade operations are prioritized, but non-critical updates are left unaddressed and resource utilization is inefficient

Engineering Contradiction:

Improvecritical operation throughputVSAvoidresource allocation complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements dynamics by allowing the lock manager to adjust lock assignment priorities based on real-time availability information and operational conditions. Rather than using static traffic failouts, the system dynamically allocates resources to upgrade operations based on actual system state, enabling both critical and non-critical updates to be handled efficiently according to current conditions, thereby reducing overall resource utilization complexity.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS11803362B2Lock-lease management of hardware and software maintenance, deployment, and optimization operations using effective application availaility

Publication Date: 2023.10.31 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

In an example embodiment, the concept of effective application availability is introduced. In a data center, there may be a minimum amount of a particular resource that is needed to continue proper functioning of an application. This may be termed a safe application availability limit. The safe application availability limit may be obtained or determined using a safe application availability limit rule. An effective application availability may be calculated by taking a current application availability and subtracting it by an availability loss from a requested operation. The effective application availability may then be compared to the safe application availability limit in order to determine whether to permit the operation to be performed at the requested time, or whether to request that the operation be attempted again at some later time.