Distributed Simulation Nodes with Local Data Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed simulation environments, the sharing of a single storage node among multiple simulation apparatuses can lead to a bottleneck phenomenon, where tasks must wait for others to complete before proceeding, especially during extensive regression testing, due to high storage access demands.
Innovation Solution
A distributed simulation system is implemented with simulation nodes that each have their own allocated data storage and a data updater to ensure data consistency, allowing each node to receive data only once from the storage node before executing tasks, and using advanced storage technologies like HDD, SSD, DRAM, SRAM, FRAM, or flash memory, and file systems such as AFS or Hadoop, with a management node allocating tasks based on computation power and load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multiple simulation apparatuses share one storage node, then storage resource utilization is improved, but storage access bottleneck occurs
Solution Approach 1:
The patent divides the shared storage node into multiple data storages, with each simulation apparatus having its own dedicated data storage. This segmentation eliminates the storage access bottleneck by allowing parallel access without contention, while still maintaining high storage resource utilization through the distributed architecture.
2Stability of the object's composition
If simulation apparatuses repeatedly receive simulation data from storage, then data consistency is improved, but storage access overhead increases
Solution Approach 1:
The patent implements a data storing policy that pre-loads and caches simulation data (such as DUT data and simulation input data) into local data storages before execution. This preliminary action ensures data consistency is maintained while significantly reducing storage access time during actual simulation execution, as the data is already available locally.
3Productivity
If parallel task execution is increased, then productivity is improved, but storage bottleneck phenomenon worsens
Solution Approach 1:
By segmenting the storage system into multiple independent data storages distributed across different simulation apparatuses, the patent enables true parallel task execution without storage coordination complexity. Each apparatus accesses its own local storage independently, eliminating the need for complex storage access coordination while maximizing productivity through parallel execution.
4Productivity
If local data storage is implemented, then storage access speed is improved, but data consistency management complexity increases
Solution Approach 1:
The patent implements a data updater that actively monitors and compares data across distributed data storages, providing feedback mechanisms to maintain consistency. This feedback system automatically detects and resolves data inconsistencies, managing data consistency complexity through systematic comparison and update protocols while preserving the speed benefits of local storage access.
Data Source
AI summary
A simulation apparatus and a distribution simulation system are disclosed. The simulation apparatus, according to one example, includes a simulation executer configured to execute a simulation task, a data storage configured to store data related to the simulation task based on a data storage policy that is set in advance of the execution of the simulation tasks, and a data updater configured to update the data stored in the data storage to most recent data by comparing the data stored in the data storage with data stored in another simulation apparatus.