Storage Topology Manager for Distributed Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed computing systems, existing technologies face challenges in dynamically managing and maintaining shared storage topologies, which affects high availability and application placement, especially in environments where applications depend on specific shared storage devices that may not be accessible by all nodes.
Innovation Solution
A method for dynamically constructing and maintaining a storage topology representation in a distributed computing system by aggregating information from nodes about storage devices, updating the topology in response to changing conditions, and using this information to govern application placement and drive dynamic placement decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dynamic storage topology management is implemented, then high availability and application placement accuracy are improved, but system complexity increases
Solution Approach 1:
A storage topology manager acts as an intermediary component between nodes and storage devices. It dynamically receives information from nodes about connected storage devices, constructs a global storage topology representation, and uses this information to govern application placement decisions. This mediator approach centralizes topology management functionality, reducing the complexity burden on individual nodes while enabling high availability through informed placement decisions.
2Measurement precision
If dynamic topology construction is implemented, then application placement accuracy is improved, but information aggregation overhead increases
Solution Approach 1:
The storage topology manager implements a feedback mechanism where nodes dynamically send information about storage devices when conditions change. The manager receives these updates, updates its topology representation, and uses this current information to make accurate placement decisions. This feedback loop ensures placement accuracy reflects current system state while managing information aggregation through event-driven updates rather than continuous polling.
Solution Approach 2:
The system performs preliminary actions by having nodes proactively send storage information updates before application placement decisions are made. The topology manager maintains an up-to-date representation by receiving advance information about storage device connectivity changes, allowing rapid placement decisions without requiring extensive real-time data gathering during the placement process itself.
Data Source
AI summary
Defining a storage topology of a distributed computing system including a set of machine nodes. A method includes dynamically receiving from a number of nodes in the distributed computing system information about storage devices. Each node sends information about storage devices connected to that particular node. The information is sent dynamically from each node as conditions related to storage change and as a result of conditions related to storage changing. From the received information, the method includes dynamically constructing a storage topology representation of the distributed computed system.