Distributed HA Resource Reservation in Cloud Clusters

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing high-availability (HA) models for cloud services inefficiently utilize resources due to strict node-to-storage relationships and half-capacity utilization to tolerate single-node failures, leading to wasted resources and potential downtime in case of multiple failures.

Innovation Solution

An improved HA resource reservation approach that distributes reserved capacity across a cluster of nodes to avoid service disruption while tolerating a desired number of node failures, using a scheduling algorithm to assign units of compute to primary nodes and reserve capacity for failover across multiple secondary nodes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If reserved capacity is concentrated on specific HA partner nodes to tolerate single-node failures, then system reliability is improved, but resource utilization deteriorates due to half-capacity utilization requirements

Engineering Contradiction:

Improvesystem reliabilityVSAvoidresource utilization

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent segments the reserved capacity requirement into distributed portions across multiple secondary nodes rather than concentrating it on single HA partners. Each secondary node holds a portion of the reserved capacity, collectively providing the full failover capacity needed. This segmentation allows better resource utilization while maintaining reliability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent makes secondary nodes universal by enabling them to serve dual purposes: normal workload execution and failover capacity provision. Secondary nodes can dynamically switch between these roles based on system conditions, allowing the same infrastructure to support both productivity and reliability requirements without dedicated standby capacity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Stability of the object's composition

If strict node-to-storage relationships are enforced in traditional HA models, then data integrity is improved, but resource flexibility deteriorates leading to wasted resources

Engineering Contradiction:

Improvedata integrityVSAvoidresource flexibility

Core Design Contradiction:

Stability of the object's compositionVSAdaptability or versatility

Solution Approach 1:

The patent introduces a software layer (orchestration system) as an intermediary that manages data integrity and failover logic. This intermediary enables flexible resource allocation across the cluster while maintaining data integrity through coordinated management, removing the need for strict physical node-to-storage relationships.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transitions from a single-dimension model (strict node-to-storage binding) to a multi-dimensional model where resources are allocated across the cluster based on multiple factors including workload, capacity availability, and failover requirements. This dimensional change enables both data integrity and resource flexibility.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

3Reliability

If capacity is reserved on dedicated HA partner nodes, then failover reliability is improved, but cluster productivity deteriorates due to reduced available capacity

Engineering Contradiction:

Improvefailover reliabilityVSAvoidcluster productivity

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent merges the failover capacity function with the workload execution function by allowing secondary nodes to serve both purposes simultaneously. The reserved capacity on secondary nodes is not idle but can be utilized for workloads when not needed for failover, thereby improving cluster productivity while maintaining failover reliability.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces dynamic resource allocation where the role of nodes (primary, secondary, or standby) and their capacity allocation can change based on system conditions. This dynamic approach allows the cluster to optimize between reliability and productivity by adjusting resource allocation in response to workload demands and failure conditions.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20250181463A1Cost-effective, failure-aware resource allocation and reservation in the cloud

Publication Date: 2025.06.05 NETAPP INC

AI summary

Systems and methods for an improved HA resource reservation approach are provided. According to one embodiment, for a given HA cluster of greater than two nodes in which a number (f) of concurrent node failures are to be tolerated, more efficient utilization of resources may be achieved by distributing HA reserved capacity across more than f nodes of the cluster rather than naïvely concentrating the HA reserved capacity in f nodes. As node failures are not a common occurrence, those of the nodes of the HA cluster having HA reserved capacity may allow for some bursting of one or more units of compute executing thereon unless or until f concurrent node failures occur, thereby promoting more efficient utilization of node resources.