Virtualized File Server Scaling via Sharding Map
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtualized file server systems face inefficiencies in resource utilization and performance due to the lack of effective mechanisms for scaling and load management across multiple virtual machines, leading to suboptimal storage access and distribution.
Innovation Solution
The implementation of a virtualized file server architecture that includes File Server Virtual Machines (FSVMs) and Controller/Service Virtual Machines (CVMs), which manage storage access operations and distribute storage items across multiple host machines using a sharding map to optimize storage access and reduce latency by localizing data access operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple virtual machines share physical resources to improve utilization, then resource efficiency improves, but storage access performance and latency deteriorate due to resource contention and distribution complexity
Solution Approach 1:
The patent segments the virtualized file server into multiple File Server Virtual Machines (FSVMs), each responsible for specific storage operations. This segmentation allows parallel processing of storage requests across multiple FSVMs, improving both resource utilization and access speed by distributing the workload effectively.
Solution Approach 2:
The patent introduces Controller/Service Virtual Machines (CVMs) as intermediaries between user virtual machines and storage resources. CVMs manage storage access operations, handle load balancing, and coordinate data distribution, thereby improving storage access performance while maintaining efficient resource utilization across the virtualized environment.
2Quantity of substance
If virtual machines are scaled to handle increased loads, then system capacity improves, but resource distribution efficiency deteriorates due to lack of effective scaling mechanisms
Solution Approach 1:
The patent implements dynamic scaling mechanisms where FSVMs and CVMs can be automatically created, removed, or migrated based on system load conditions. This dynamic approach allows the system to scale capacity while maintaining resource distribution efficiency through automated load balancing and resource allocation algorithms.
Solution Approach 2:
The system incorporates feedback mechanisms that monitor resource utilization and storage access patterns. Based on this feedback, the system automatically adjusts the number and configuration of FSVMs and CVMs, ensuring optimal resource distribution efficiency as system capacity scales to handle increased loads.
3Reliability
If storage items are distributed across multiple host machines, then storage capacity and availability improve, but access latency increases due to distributed data location and retrieval complexity
Solution Approach 1:
The patent implements a sharding map that pre-determines and records the location of distributed storage items across host machines. This preliminary organization allows FSVMs to quickly locate and access distributed data without complex real-time search operations, reducing access latency while maintaining the benefits of distributed storage for capacity and availability.
Solution Approach 2:
The patent optimizes storage access by localizing data operations through the sharding map, which enables FSVMs to access storage items from local caches or nearby nodes when possible. This local quality approach reduces the need for remote data retrieval, thereby maintaining high storage availability while minimizing access latency.
Data Source
AI summary
Examples described herein include virtualized environments including a virtualized file server. Examples of file server virtual machine managers are described which may setup and/or configure file server virtual machines in accordance with a target number of connections and/or a target throughput for the virtualized file server. The virtualized file server may be scaled up, scaled down, scaled in, and/or scaled out.