Virtual Platform Cluster Resource Sharing with Write Order Fidelity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtualized environments, managing and coordinating the sharing of resources across geographically distributed storage nodes with synchronous and asynchronous latencies is challenging, particularly in maintaining data coherence and ensuring proper write order fidelity to address errors and failures.
Innovation Solution
A method and system for dynamically managing shared resources among virtual platform clusters, utilizing a non-transitory computer-readable medium with executable code to provide access to distributed resources under different latencies and operating conditions, including synchronous and asynchronous operations, with a management component distributed across clusters to ensure data coherence and write order consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is cached at directors to improve access speed, then access speed is improved, but data coherence with modifications at other directors becomes difficult to maintain
Solution Approach 1:
The system implements a feedback mechanism where directors broadcast write operations to all other directors in the array. When a director modifies data, it sends a broadcast message containing the modified data and its new location. Other directors receive this feedback and update their caches accordingly, ensuring coherence across the distributed system while maintaining fast local access.
Solution Approach 2:
The system performs preliminary actions by pre-establishing a communication protocol and broadcast mechanism before data modifications occur. Directors are pre-configured to listen for broadcast messages and automatically update their caches when changes are announced, eliminating the need for complex real-time coordination during data access operations.
2Reliability
If synchronous transfers are used to ensure data reliability, then data protection is improved, but transfer time increases significantly for geographically distant sites
Solution Approach 1:
The system dynamically adapts the transfer protocol based on the operational requirements and network conditions. For geographically distant sites where synchronous transfers would cause excessive delay, the system allows asynchronous operations while maintaining reliability through the broadcast mechanism and write-order-fidelity protocol, which ensures that all directors see writes in the correct sequence without requiring synchronous confirmation.
Solution Approach 2:
The system changes the timing parameter of data transfers based on distance and requirements. Instead of using fixed synchronous transfers, the system adjusts the transfer timing to be asynchronous for distant sites, while still maintaining data protection through the broadcast-based coherence protocol that ensures all directors have consistent views of data without requiring synchronous acknowledgment.
3Speed
If asynchronous transfers are used to reduce transfer time, then transfer speed is improved, but maintaining proper write ordering becomes difficult
Solution Approach 1:
The system uses feedback through broadcast messages to maintain write order fidelity in asynchronous transfers. When a director performs a write operation, it immediately broadcasts the write details to all other directors. This feedback mechanism allows the system to track and enforce write ordering across asynchronous transfers, ensuring that all directors apply writes in the correct sequence even though transfers occur independently.
Solution Approach 2:
The broadcast mechanism acts as an intermediary that coordinates write operations across asynchronous transfers. Instead of relying on direct synchronization between pairs of directors, the broadcast serves as a central coordination mechanism that announces all writes to the entire system, allowing asynchronous transfers to proceed at high speed while maintaining proper ordering through the intermediary broadcast protocol.
4Ease of operation
If centralized management infrastructure is used to control virtual machines, then management control is improved, but system complexity and single point of failure risk increase
Solution Approach 1:
The system implements self-service by enabling directors to autonomously perform data caching, coherence maintenance, and write ordering without requiring constant centralized control. Each director independently manages its own cache and participates in the broadcast protocol to maintain coherence, reducing the burden on the virtual center and distributing intelligence throughout the system.
Solution Approach 2:
The system segments management functions by distributing autonomy to individual directors while the virtual center handles higher-level virtual machine management. This segmentation reduces the complexity burden on any single component - directors handle local cache coherence independently, while the virtual center manages virtualization policies, creating a layered architecture that simplifies overall system complexity.
Data Source
AI summary
Sharing resources in a virtualized environment includes providing access by a plurality of virtual platform clusters to distributed resources of the virtualized environment, where each of the virtual platform clusters has a director that manages resources for one or more virtual machines in a corresponding one of the virtual platform clusters. Sharing resources also includes dynamically managing sharing of the distributed resources among the plurality of virtual platform clusters according to a sharing arrangement, where the distributed resources include a plurality of objects. A first subset of the plurality of objects is shared among the plurality of virtual platform clusters differently than a second subset of the plurality of objects. At least some of the objects that are shared are also copied locally to different ones of the virtual platform clusters.