Virtualized Cluster Communication System Witness Node
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtualized computing systems face challenges in maintaining redundancy and performance due to high I/O bandwidth consumption during continuous replication and the potential disruption of 'live' migration services by hardware or software faults, leading to service loss.
Innovation Solution
A virtualized cluster communication system is implemented, where each node executes a specialized instance of an operating system privileged to control a hypervisor, enabling VMs to be consolidated and managed as cluster resources through a cluster stack, allowing peer communication and automatic failover in case of node failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If continuous replication of primary VMs on secondary server is implemented to provide redundancy, then reliability is improved, but I/O bandwidth consumption increases leading to decreased productivity
Solution Approach 1:
The system segments the redundancy function by introducing a dedicated witness node that stores only configuration and state information, separate from the primary and secondary servers that handle actual VM workloads. This segmentation allows redundancy to be maintained without the full I/O bandwidth consumption of continuous replication, as the witness node requires minimal data synchronization.
Solution Approach 2:
The witness node acts as an intermediary in the failover process. Instead of direct continuous replication between primary and secondary servers consuming excessive I/O bandwidth, the witness node mediates the failover decision-making process by monitoring heartbeats and triggering failover only when necessary, thereby reducing unnecessary data transmission and maintaining system performance.
2Reliability
If VM live migration is used to transfer VM copy to second physical server, then reliability is improved, but the service may be disabled in event of severe hardware or software fault at primary server
Solution Approach 1:
The system performs preliminary action by pre-configuring the witness node with cluster configuration information and state data before any failure occurs. The witness node continuously monitors the health of primary and secondary servers through heartbeat mechanisms, so when a severe fault occurs, the failover can be immediately triggered without requiring complex live migration operations, ensuring service continuity even when live migration would be disabled.
3Reliability
If second physical server with identical virtualized environment is introduced for redundancy, then reliability is improved, but device complexity increases
Solution Approach 1:
The system extracts the critical redundancy function from the complex requirement of maintaining identical virtualized environments on secondary servers. By separating the witness node's role to only store configuration and state information rather than full VM replicas, the system reduces device complexity while maintaining reliability. The witness node contains only essential metadata, not complete VM images, thereby simplifying the overall system architecture.
Data Source
AI summary
A method includes executing, in each of a number of nodes of a cluster communication system, a specialized instance of an operating system privileged to control a corresponding hypervisor configured to consolidate one or more VM(s) on a system hardware. The one or more VM(s) is configured to be associated with a non-privileged operating system. The method also includes providing a cluster stack associated with the specialized instance of the operating system on the each of the number of nodes to enable communication between peers thereof in different nodes, and controlling the one or more VM(s) as a cluster resource through the cluster stack.