NVLog Access Control for Storage Server Write Requests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storage servers face latency issues due to NVLog overflow during heavy loads, which prevents further write requests from being acknowledged until the NVLog is cleared during consistency point processing, impacting performance in mission-critical applications.
Innovation Solution
Implementing a method to control access to the NVLog by categorizing incoming write requests and limiting their usage based on priority, allowing only 'important' requests to log during heavy loads, thereby managing NVLog space allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the storage server holds write requests in memory and periodically saves them to mass storage devices, then disk access time is reduced and system performance is improved, but data loss risk increases in case of system failure between consistency points
Solution Approach 1:
The patent implements preliminary action by maintaining an NVLog that preemptively stores write requests in non-volatile memory before they are flushed to disk. This ensures that even if a failure occurs between consistency points, the data is already safely stored in non-volatile memory, eliminating data loss risk while maintaining high performance.
2Reliability
If the NVLog is used to recover data after system failure, then data reliability is improved, but write request latency increases when the NVLog becomes full during heavy loads
Solution Approach 1:
The patent applies segmentation by dividing the NVLog into multiple separate logs, each dedicated to a specific client or client group. This allows write requests from different clients to be processed in parallel without blocking each other, significantly reducing latency during heavy loads while maintaining full data recovery capability for each segmented log.
Solution Approach 2:
The patent implements partial action by allowing the system to maintain only the necessary portion of the NVLog for each client rather than a single large unified log. This reduces the overall time required to process and clear logs while ensuring adequate buffer space for each client's write requests, thereby reducing latency.
3Productivity
If the NVLog is cleared after each consistency point, then memory space is reused and system efficiency is improved, but the ability to handle concurrent write requests from multiple clients is reduced
Solution Approach 1:
By segmenting the NVLog into client-specific logs, the system can clear and reuse space in one client's log independently of others. This maintains high system efficiency through space reuse while simultaneously improving concurrent request handling capability, as each client has dedicated buffer space that doesn't interfere with other clients.
Data Source
AI summary
In one embodiment of the present invention, a method and system are provided to control access to the non-volatile log (NVlog) of a storage server. By controlling access to the NVLog of a storage server the relative disk write bandwidth available to different client write requests can be controlled. The incoming write request can be categorized, and, during times of heavy load, only be permitted to use NVLog space as permitted based on the categorization of each write request. In one embodiment, the present invention includes receiving a write request from a client at a storage server, and determining whether the received write request can be presently logged in a NVlog based on a category of the write request.