Distributed Computing Grid Using Segmented User Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current distributed computing systems face challenges in achieving high-throughput and sustainability, particularly in resource consumption and heat dissipation, as they rely heavily on centralized cloud servers and require extensive cooling systems.
Innovation Solution
A method and system for distributed computing that utilizes a customer platform to upload job specifications and split computational jobs into chunks, distributing them to user devices such as personal computers, smartphones, and smart TVs, allowing for efficient execution and heat dissipation without the need for centralized cooling, while ensuring reliability through task replication and dynamic task assignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If centralized cloud servers are used for distributed computing, then computational power and processing capability are improved, but resource consumption and heat dissipation requirements increase
Solution Approach 1:
The patent segments the centralized cloud computing infrastructure into distributed user devices (smartphones, tablets, computers) that perform computational tasks locally. This segmentation eliminates the need for large centralized servers and their associated cooling systems, reducing energy consumption while maintaining computational capability through distributed processing.
Solution Approach 2:
The patent enables user devices to serve themselves by performing computational tasks locally without requiring centralized cloud infrastructure. Users' personal devices utilize their own processing power, memory, and storage resources, eliminating the need for external cooling systems and reducing overall energy consumption while maintaining computational functionality.
2Power
If centralized cloud servers are used for distributed computing, then computational power is improved, but cooling system requirements increase
Solution Approach 1:
The patent extracts the computational workload from centralized cloud servers and distributes it to user devices. By removing the centralized server infrastructure, the associated cooling systems are eliminated, reducing device complexity and energy consumption while maintaining computational power through distributed processing across multiple user devices.
Solution Approach 2:
User devices perform computational tasks independently using their own hardware resources, eliminating the need for centralized cooling infrastructure. Each device uses its inherent cooling capabilities, removing the complexity of large-scale cooling systems while maintaining full computational functionality.
3Loss of energy
If computational jobs are distributed to user devices, then resource consumption is reduced, but task management complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where the system monitors device availability, performance, and task completion status. Based on this feedback, the system dynamically adjusts task distribution, reassigns tasks to available devices, and optimizes workload management, reducing task management complexity through automated adaptive control.
Solution Approach 2:
The patent employs dynamic task management where the system continuously adapts to changing device availability and performance conditions. Task assignment and redistribution are performed dynamically based on real-time system state, simplifying management through automated adaptation rather than static rigid scheduling.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
A method and a system (100) for high-throughput distributed computing of computational jobs (400), comprise setting up a data storage system (120) and a grid of user devices (310) that offer computational capacity. Customer entities (200) upload in a customer platform (111) first job specification parameters, and then the full job data. A partitioning scheme is selected based on the job specification and on the grid status, that is periodically updated by querying the user devices (310). Based on the partitioning scheme, the job data are split on the fly in input chunks (410), that are included in executable tasks (420) and assigned to different user devices (310). To this end, devices (310) may be selected based on computing capacity and availability parameters, and on a priority level selected for the job (400). Output chunks (430) are generated by executing the tasks (420), and after verification of timely arrival of all the required output chunks (430), they are assembled as a complete job result (500), for download by the customer (200).