Decentralized Task Management via Distributed Process Slicing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In decentralized data networks with nodes of limited memory capacity, it is challenging to execute large computing processes due to insufficient memory for swapping processes.
Innovation Solution
A method where computing processes are suspended, broken down into slices, and distributed across network nodes using a mapping rule, such as a hash function, creating a distributed process image that can be resumed by finding and reassembling the slices on suitable nodes, allowing for the execution of larger processes even with restricted memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If processes are executed on network nodes with limited memory capacity, then device complexity is reduced, but the ability to execute large computing processes deteriorates
Solution Approach 1:
The patent divides the process image into multiple slices that can be distributed across different network nodes. Each slice is a portion of the overall process memory image, allowing the process to be executed across nodes with limited individual memory capacity while collectively providing sufficient memory space for large processes.
Solution Approach 2:
The patent transitions from executing processes within the memory constraints of a single node to distributing process slices across multiple nodes in the network. This adds a spatial dimension to memory utilization, allowing the system to overcome individual node memory limitations by leveraging the collective memory space of the network.
2Adaptability or versatility
If processes are distributed across multiple network nodes, then the ability to execute large processes is improved, but device complexity increases
Solution Approach 1:
The patent introduces a process management system that acts as an intermediary between the process execution and memory management functions. This mediator handles the complexity of creating process images, dividing them into slices, distributing them across nodes, and managing their reassembly during process resumption, thereby shielding the individual nodes from the complexity of distributed process execution.
Solution Approach 2:
The patent creates a process image copy that is divided into slices and distributed across network nodes. Instead of requiring each node to hold the complete process image, only the necessary slices are copied to each node, reducing the memory burden on individual nodes while enabling large process execution.
3Quantity of substance
If process images are broken down into slices and distributed, then memory requirements per node are reduced, but the complexity of process resumption increases
Solution Approach 1:
The patent performs preliminary actions by creating the process image, dividing it into slices, and distributing them to appropriate nodes before the process needs to be resumed. This preparation phase handles the complexity of slice distribution and node selection in advance, so that during process resumption, the system only needs to retrieve the slices and reassemble them, significantly simplifying the resumption operation.
Data Source
AI summary
In a method for the management of tasks in a decentralized data network with a plurality of nodes for carrying out the tasks, resources are distributed based on a mapping rule, in particular a hash function. A task that is to be suspended is distributed by dividing the process image of the task into segments and by distributing the segments over the nodes using the mapping rule. Thus, a distributed swap space is created so that tasks can also be carried out on nodes whose swap space is not sufficient on its own. The method can be used in embedded systems with a limited storage capacity and/or in a voltage distribution system, wherein the nodes are, for example, switching units in the voltage distribution system. The method can also be used in any other technical systems such as, for example, a power generation system, an automation system and the like.