Cross-Cluster Application Deployment for Resource-Aware Component Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face inefficiencies in managing data storage and retrieval across multiple storage devices, particularly in flash-based systems, due to redundant processes and lack of direct control by the operating system, leading to increased latency and reduced reliability.
Innovation Solution
Implementing a direct-mapped flash storage system where the operating system directly addresses data blocks without intermediate storage controllers, offloading device management tasks and utilizing non-volatile RAM for fast data buffering, along with dual storage array controllers for failover and efficient data handling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If intermediate storage controllers are used to manage flash drives, then device management is simplified, but storage access latency increases and reliability decreases
Solution Approach 1:
The patent extracts the flash drive management functionality from intermediate storage controllers and directly integrates it into the operating system. The OS now directly addresses data blocks on flash drives without requiring intermediate controllers, eliminating the latency introduced by additional management layers while maintaining simplified management through OS-level control.
Solution Approach 2:
The patent introduces a direct addressing mechanism as an intermediary between the OS and flash drives, allowing the OS to directly locate and access data blocks on flash drives without intermediate controllers. This direct addressing pathway eliminates the mediation delay caused by traditional storage controllers while maintaining the ability to manage flash drives through standard OS interfaces.
2Ease of operation
If intermediate storage controllers are used, then device management is centralized, but system reliability decreases due to single points of failure
Solution Approach 1:
The patent extracts the storage control functionality from dedicated storage controllers and embeds it directly into the operating system kernel. This eliminates single points of failure associated with external controllers, as the OS itself becomes the management layer, distributing control across the system's existing redundant infrastructure and improving overall reliability.
Solution Approach 2:
The operating system is designed to perform multiple functions including direct flash drive management, eliminating the need for specialized storage controllers. By making the OS universal in its capability to manage storage devices directly, the system removes single points of failure and leverages the OS's inherent redundancy and fault tolerance mechanisms.
3Loss of time
If direct-mapped flash storage is implemented, then access latency is reduced, but data integrity during power failures may be compromised
Solution Approach 1:
The patent implements a power failure detection mechanism that performs preliminary actions by detecting power loss conditions and initiating data integrity protection protocols before data corruption can occur. The system monitors power status and prepares protective measures in advance, allowing it to safeguard data integrity even in direct-mapped flash storage configurations where write operations complete quickly.
Data Source
AI summary
In some embodiments, a method for software application deployment across dusters may include receiving a request to deploy a software application with a first deployable component and a second deployable component. The method may further include determining, based on the request, an availability, within a first duster of the plurality of dusters, of first computing resources capable of executing the first deployable component, and an availability, within a second duster of the plurality of dusters different from the first duster, of second computing resources capable of executing the second deployable component. The method may further include deploying, based on the availabilities of the first and second computing resources, the software application across the plurality of dusters. For example, this deploying across dusters may include deploying the first deployable component to the first cluster and deploying the second deployable component to the second duster. Corresponding methods and systems are also disclosed.