Advance Support
Table of contents
Scaling
To process large datasets, you can apply the following strategies:
- You may deploy BQAT-Stateless as a cluster of containers across your array of servers.
- You may split the dataset into smaller chunks and process them in parallel using BQAT-CLI or BQAT-API (only vertical scaling is supported by default).
Depend on your use case, you might need customised scaling solution in the cloud or on-premise. We will help to design and implement horizontal scaling for BQAT-CLI and BQAT-API (BQAT-GUI).
Optimisation
By default, BQAT is not optimised to use all available CPU cores and memory for robustness and statability concern. We will need to look at the spec of your machine to suggest the best configuration for you.
On the other hand, GPU accereleration is not enabled in BQAT. We will need to look at the spec of your GPU and driver version to suggest the best configuration for you. This will speed up a series of machine learning components of the algorithm.
Security
We continuously monitor CVEs and available patches for dependencies. If you have specific security requirements, we’re here to help you mitigate any risks in your production environment.
Compliance
We are aiming to provide quality metrics which can help to evalutate the quality of your data for international standard compliance. For instance, regarding face, ISO 29794-5 is the international standard we are looking at. If you would like to add other quality metrics, we will help you to customise BQAT to meet your compliance requirements.
Customisation
- Data IO customisation (e.g. base64 handling, output formating)
- API endpoint customisation (adapt to upstream and downstream system)
- EDA report customisation (advance visualisation)
- Web access control for remote user
- Backend data storage management
- Metric score calibraion and labeling customisation (adapt to different compliance standard)
Contact us for advance support.