Computer Cluster Disaster Recovery via Dual Compartment Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computer clusters for hosting server-based applications lack customization, dynamic resource resizing, and high availability, with standard virtual machine offerings restricted to a single server and limited network settings, which hinders scalability and reliability.
Innovation Solution
A computer cluster architecture with two compartments of physical nodes, each with multiple partitions for infrastructural and guest virtual instances, allowing for monitoring, redundancy, and dynamic resource management, enabling high reliability and availability with disaster recovery capabilities, and allowing customers to manage resources without noticing hardware changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If standard virtual machine hosting is provided on a single server, then the system is simple to manage, but it lacks high availability and cannot withstand hardware failures
Solution Approach 1:
The system is divided into multiple physical nodes (first and second compartments), each running independent virtualization environments. This segmentation ensures that failure of one node does not affect others, providing high availability while maintaining manageable complexity through modular architecture
Solution Approach 2:
Each physical node is configured with specific local resources (mass storage devices with designated partitions) and local virtual instances. This local quality allows each node to independently provide services, ensuring continuity of operation even when other nodes fail
2Adaptability or versatility
If resources are dynamically resized and customized, then customer flexibility improves, but hardware management complexity increases
Solution Approach 1:
A management system acts as an intermediary between customers and the physical hardware infrastructure. This intermediary handles the complexity of hardware management, resource allocation, and dynamic resizing, allowing customers to access customized resources without directly managing hardware complexity
Solution Approach 2:
The management system provides universal functionality across multiple physical nodes, handling resource provisioning, monitoring, and customization tasks centrally. This multi-functional approach enables adaptability while consolidating hardware management complexity into a single system
3Reliability
If infrastructural virtual instances monitor each other across compartments, then disaster recovery capability improves, but system complexity increases
Solution Approach 1:
The monitoring functionality is merged into the infrastructural virtual instances themselves, which are already present in each compartment. By combining monitoring capabilities with existing infrastructural components, the system achieves disaster recovery without adding separate monitoring systems that would increase complexity
4Productivity
If mass storage devices are partitioned for multiple purposes, then resource utilization improves, but risk of failure impact increases
Solution Approach 1:
Mass storage devices are segmented into multiple partitions with distinct purposes (boot partition, infrastructural resources, guest resources). This segmentation isolates failure impacts - a failure in one partition does not affect others - while still achieving high resource utilization through shared infrastructure
Data Source
AI summary
A computer cluster includes: a first compartment, further including a plurality of first physical nodes configured to provide first resources for a first infrastructural virtual instance and second resources for guest virtual instances; and a second compartment, further including a plurality of second physical nodes configured to provide third resources for a second infrastructural virtual instance and fourth resources for guest virtual instances. Each of the plurality of physical nodes further includes a mass storage device, the mass storage device further including: a partition provided for booting the physical node; a partition as part of resources assigned to an infrastructural virtual instance; and a partition as part of other resources. The first infrastructural virtual instance and the second infrastructural virtual instance are configured for monitoring each other.