Dynamic Resource Allocation for Storage Replication

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In clustered storage systems, resource management for replication transfers and client traffic is challenging due to resource limitations, which can lead to negative impacts on data transfer rates and latency if not properly balanced based on recovery point objectives (RPO) and service level objectives (SLO).

Innovation Solution

A system and method for controlling resources by using a token-based approach to allocate tokens based on SLOs, prioritizing replication transfers and client traffic, and scheduling replication transfers based on RPOs to ensure efficient resource utilization and maintain performance characteristics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If replication transfers are performed frequently to maintain data backup, then data reliability is improved, but resource consumption increases and affects client traffic performance

Engineering Contradiction:

Improvedata reliabilityVSAvoidresource consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system dynamically adjusts replication transfer frequency and resource allocation based on real-time conditions. The resource control module monitors system state and modifies replication behavior adaptively, changing transfer schedules and resource allocation ratios according to current workload and performance metrics, thereby resolving the contradiction between maintaining high reliability and reducing resource consumption.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes key parameters such as replication transfer frequency, resource allocation ratios, and priority levels based on system conditions. By adjusting these parameters dynamically, the system can maintain data reliability while optimizing resource consumption to prevent negative impacts on client traffic.

Inventive Principle:

Principle #35Parameter changes

2Productivity

If more resources are allocated to replication transfers, then data backup performance is improved, but client traffic performance deteriorates

Engineering Contradiction:

Improvedata backup performanceVSAvoidclient traffic performance

Core Design Contradiction:

ProductivityVSSpeed

Solution Approach 1:

The system allocates resources partially to replication transfers based on priority levels and system conditions, rather than consistently allocating maximum resources. The resource control module determines appropriate allocation ratios that provide sufficient backup performance while leaving adequate resources for client traffic, avoiding excessive resource consumption by replication operations.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The resource control module acts as an intermediary between replication transfers and client traffic, managing resource allocation and scheduling. It mediates resource distribution by controlling token issuance, adjusting replication transfer schedules, and balancing resource availability between backup operations and client access, thereby preventing client traffic performance deterioration.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Reliability

If replication transfers are prioritized to ensure data backup, then reliability is improved, but latency increases

Engineering Contradiction:

Improvedata backup reliabilityVSAvoidlatency

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system implements periodic replication transfers with dynamically adjusted intervals rather than continuous or fixed-schedule transfers. The resource control module determines optimal transfer timing based on system conditions, using periodic actions that ensure data backup reliability while minimizing latency impact by spacing transfers appropriately and avoiding excessive frequency.

Inventive Principle:

Principle #19Periodic action

Data Source

PatentUS9983958B2Techniques for dynamically controlling resources based on service level objectives

Publication Date: 2018.05.29 NETAPP INC

AI summary

Various embodiments are generally directed an apparatus and method for receiving a recovery point objective for a workload, the recovery point objective comprising an amount of time in which information for the workload will be lost if a failure occurs, and determining a service level objective for a replication transfer based on the recovery point objective, the replication transfer to replicate information on a destination node to maintain the recovery point objective. Various embodiments include dynamically controlling one or more resources to replicate the information on the destination node based on the service level objective and communicating information for the replication transfer from the source node to the destination node.