Virtual Data Service Appliance for Hypervisor-Level Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional replication and disaster recovery systems are not fully virtualization-aware, leading to operational and administrative burdens and inflexibility in supporting mission-critical applications in virtualized environments.
Innovation Solution
A dedicated virtual data service appliance (VDSA) within a hypervisor provides replication, monitoring, and quality of service, using a tapping driver to intercept and analyze I/O requests, enabling journaling and asynchronous data replication across multiple sites while maintaining write order fidelity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional replication systems are implemented at the physical disk level, then hardware compatibility is improved, but virtualization awareness and operational flexibility deteriorate
Solution Approach 1:
The patent introduces a virtualization layer (hypervisor) as an intermediary between the physical disk level and the virtual server level. This hypervisor abstracts the physical storage resources and presents virtualized storage to multiple virtual servers, enabling the replication system to operate at the virtual disk level while maintaining awareness of virtualization. The hypervisor manages the mapping between virtual disks and physical disks, handling the complexity of virtualization transparently.
Solution Approach 2:
The patent shifts the replication operation from the traditional physical disk dimension to the virtual disk dimension. By implementing replication at the virtual disk level within the hypervisor, the system gains visibility and control over virtualized resources while maintaining compatibility with underlying physical hardware. This dimensional shift enables replication to follow virtual disks across different physical hosts.
2Adaptability or versatility
If data services are implemented at the physical disk level, then hardware agnosticism is improved, but service flexibility and scalability deteriorate
Solution Approach 1:
The hypervisor provides a universal platform that hosts multiple virtual servers and manages their storage resources. The data services (replication, monitoring, quality of service) are implemented within the hypervisor, making them universally applicable to all virtual disks regardless of the underlying physical hardware. This multi-functional approach allows the same service infrastructure to support diverse workloads and hardware configurations.
Solution Approach 2:
The patent implements virtual disk copying mechanisms within the hypervisor to enable replication. Instead of copying data at the physical disk level, the system creates and manages copies of virtual disks through the hypervisor's abstraction layer. This allows replication to be performed on virtual disk images, which can then be mapped to different physical storage resources, providing flexibility without duplicating physical infrastructure.
3Reliability
If replication is performed synchronously to ensure data consistency, then data integrity is improved, but recovery time objective deteriorates
Solution Approach 1:
The patent implements journaling as a preliminary action before replication. The hypervisor journals changes to virtual disks before replicating them to remote locations. This preliminary journaling captures the state of data at specific points in time, allowing for efficient recovery without requiring full synchronous replication. The journal acts as a buffer that records changes asynchronously, reducing the time impact on primary operations.
Solution Approach 2:
The system implements periodic replication of journaled data to remote locations rather than continuous synchronous replication. By periodically transferring replicated data, the system reduces the time overhead on primary operations while maintaining data consistency at acceptable intervals. The periodic nature of replication allows the system to balance between data consistency and recovery time objectives.
Data Source
AI summary
A cross-host multi-hypervisor system, including a plurality of host sites, each site including at least one hypervisor, each of which includes at least one virtual server, at least one virtual disk that is read from and written to by the at least one virtual server, a tapping driver in communication with the at least one virtual server, which intercepts write requests made by any one of the at least one virtual server to any one of the at least one virtual disk, and a virtual data services appliance, in communication with the tapping driver, which receives the intercepted write requests from the tapping driver, and which provides data services based thereon, and a data services manager for coordinating the virtual data services appliances at the site, and a network for communicatively coupling the plurality of sites, wherein the data services managers coordinate data transfer across the plurality of sites via the network.