Open Automation Software processes data with high efficiency and speed to maintain data accuracy and reliability.
OAS can run on multiple platforms with large scalability, from 100,000 tags on Raspberry Pi 4 up to 1,000,000 tags on Windows or Linux server. Deployed as an edge in a Distributed Network Architecture OAS can pre-process all incoming data and deliver to the final destination for unlimited number of tags to unlimited clients.
To emphasize the scalable performance of OAS the following video demonstrates 100,000 tags running on a Raspberry Pi 4 with all tag values changing each second and then logged to a SQL Server database.
Performance Variables
Open Automation Software addresses typical deadlocks, bottlenecks, and poor synchronization found in other applications that can cause data loss or system failure.
Factors Which May Impact Software Performance
Synchronous Processing
Synchronous or poorly-scaling architectures, forcing systems to block and limit throughput.
Single-Threaded Processing
Bottlenecks communicating with external sources, causing slow downs with additional endpoints. Processing data inline from source to completion on a single thread can cause a complete system failure if there is a breakdown at any step.
Network Traffic Bandwidth
Large data packet size and overwhelming bandwidth can slow down or even cause backups in delivery of data.
Inefficient Encoding and Decoding
Slow, or inefficient data encryption, forcing a choice between security or speed.
Single Record Processing to Database
Poorly designed database record inserts can cause backup or failure of delivery.
Network or Database Failure
If the system is not deployed as an edge solution with store and forward implemented data will be dropped and not processed. Important to cache this data properly to be able to process further incoming values continuously.
Limited Communication Channels
Most SCADA systems are limited to 200 or less channels and require separate instances to collect data from more than 200 devices.
Only Latest Value Update
Limiting what data is delivered learning engines can affect the data model.
Separate Data Source for Each Client
The number of clients can be restricted otherwise causing undue stress in overwhelming communications to the data source.
OAS Performance Features
Asynchronous Event Driven Processing
Asynchronous communications by subscription to data sources. Routines only need to execute on data change or watchdog failure.
Multi-Threaded Processing
Multi-threaded data processing while retaining first-in-first-out handling. Separate threads to receive, process, and deliver data in order it is received.
Optimized Compression
High speed data compression algorithms for encoding and decoding data with reduced packet size in comparison of open standards.
Optimized Encryption
High speed encryption to retain security without sacrificing speed. Always activated to maintain secure communications at all times.
Multi-Record Processing
Data logging and alarm logging routines perform bulk inserts to pass up to one million records per call. Eliminates data backup due to delivery error.
Store and Forward
OAS is deployed as a edge solution being able to maintain connection to the data source and can store data to disk on network or database engine failure, freeing up memory resources for continuous processing.
Unlimited Drivers per OAS Engine
OAS can connect to unlimited number of devices per server instance, proven systems in production communicating to more than 1,300 Modbus devices from one server.
First In First Out Processing
OAS delivers all values it receives from end to end, not just the latest value maintaining 100% accuracy for learning engines and reports. This makes it possible to use slower communication networks while still maintaining data accuracy up to 100 nanoseconds.
Centralized Data Namespace
Values received from each data source are centralized to OAS tag values, to be shared to unlimited number of client applications without impacting the communication rate from the data source.