Distributed Storage Control Software Failover for Load Balancing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In information processing systems with redundant server configurations, failures in multiple servers lead to load concentration on a single server, decreased I/O performance, and increased memory consumption, making it challenging to maintain availability and efficiency during failover processes.

Innovation Solution

The system employs a configuration where each redundancy group consists of one active and multiple standby storage control software units, with the standby units prioritized based on their order of influence across nodes, allowing failover without overloading any single server, thereby distributing resource consumption and maintaining performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the number of standby storage control software pieces in each redundancy group is increased to ensure availability during multiple server failures, then the availability is improved, but the load concentrates on one server and I/O performance deteriorates

Engineering Contradiction:
ImproveavailabilityVSAvoidI/O performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the standby storage control software across multiple servers rather than concentrating all standbys on a single server. Each server hosts at most one standby storage control software, distributing the backup capacity throughout the system. This segmentation prevents load concentration while maintaining the ability to handle multiple failures.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a dimensional change in the redundancy architecture by moving from a centralized standby model (all standbys on one server) to a distributed standby model (standbys spread across multiple servers). This spatial redistribution across the system dimension resolves the contradiction between availability and performance.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If the number of standby storage control software pieces is increased to ensure availability during multiple server failures, then the availability is improved, but memory consumption increases

Engineering Contradiction:
ImproveavailabilityVSAvoidmemory consumption
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent segments the memory burden across multiple servers by distributing standby storage control software instances. Instead of one server bearing the full memory load of all standbys, each server only hosts one standby instance, dividing the total memory consumption by the number of servers in the system.

Inventive Principle:
Principle #1Segmentation

3Productivity

If active storage control software is migrated to another server during failover, then resource concentration is avoided, but the I/O stop time is extended

Engineering Contradiction:
Improveresource distributionVSAvoidI/O stop time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent performs preliminary action by pre-positioning standby storage control software on multiple servers before failures occur. When a failure happens, the standby is already in place and can immediately take over, eliminating the need for migration during the failover process and thus avoiding I/O stop time extension.

Inventive Principle:
Principle #10Preliminary action

4Reliability

If standby storage control software is migrated in advance to another server, then availability is ensured, but the I/O performance of the non-failed server deteriorates

Engineering Contradiction:
ImproveavailabilityVSAvoidI/O performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the standby software distribution so that each server hosts at most one standby instance. This limits the impact of any single standby migration on overall system performance, as only one server's workload increases at a time rather than multiple standbys migrating to a single server.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS11615005B2Storage system and control method therefor
Publication Date: 2023.03.28 HITACHI VANTARA LTD
  • US11615005B2 patent drawing
  • US11615005B2 patent drawing
  • US11615005B2 patent drawing

AI summary

Each redundancy group is constituted by one active program (storage control software of the active program) and N standby programs (N is an integer of two or more). Each of the N standby programs is associated with a priority to be determined as a failover (FO) destination. In the same redundancy group, FO is performed from the active program to the standby program based on the priority. For the plurality of pieces of storage control software including the active programs and the standby programs that change to be active by FO in the plurality of redundancy groups arranged in the same node, standby storage control software that can set each of the programs as a FO destination are arranged in different nodes.