Red Hat announced several new capabilities to Red Hat OpenShift Data Science, our managed cloud service for data scientists and developers of intelligent applications.
Red Hat OpenShift Data Science now includes new features for deeper data analysis and better collaboration between ITOps, data scientists, and intelligent application developers. Furthermore, customers can now use committed AWS to spend to purchase and run Red Hat OpenShift Data Science directly through AWS Marketplace, providing easier accessibility and flexibility for deployments. This is included as part of Red Hat’s latest announcement for its expanded portfolio of open solutions publicly available in AWS Marketplace.
New feature highlights include:
Support for custom Jupyter Notebooks
Red Hat OpenShift Data Science already includes Jupyter Notebook images and common libraries, including TensorFlow, Pytorch, Scikit-learn, CUDA and more. The notebook images for TensorFlow and Pytorch support both CPUs and GPUs.
Now, we have augmented the solution’s features to allow customers to use their unique images and custom Jupyter Notebooks. This gives customers additional flexibility and control and allows them to customise their data further modelling for more accurate and tailored results.
Customised NVIDIA GPU support
No two data science workloads are the same; some may require greater GPU processing power than others. To ensure customers have the power they need, Red Hat OpenShift Data Science now includes customised NVIDIA data centre GPU support, allowing them to choose accelerated computing clusters pre-configured to their specific workloads.
The inclusion of customised GPU support exemplifies Red Hat’s continued commitment to providing customers with the ability to easily scale workloads up or down, depending on their needs. Additionally, the NVIDIA AI Enterprise software suite is certified to run on Red Hat OpenShift and includes key enabling technologies for rapidly deploying, managing and scaling AI workloads.
Red Hat has also launched a new learning path in the developer portal for Red Hat OpenShift Data Science to help customers activate this new capability and configure a Jupyter notebook to use GPUs for AI/ML modelling.
New collaboration with Pachyderm
The Red Hat OpenShift partner ecosystem comprises some of the world’s most popular data science technologies, all with certified Red Hat OpenShift Operators. A select number of these commercial partner products have been integrated directly within Red Hat OpenShift Data Science to provide more advanced data lifecycle capabilities.
That integration continues to grow with the addition of Pachyderm. Pachyderm provides automatic data versioning and allows data science teams to build and scale containerised, data-driven machine-learning pipelines with guaranteed data lineages. Customers can more easily track changes to models and be assured that the right models are put into production.
This partner offering is the latest addition to current integrations with Starburst, Anaconda Commercial Edition, IBM Watson Studio, Intel OpenVINO, and AI Analytics Toolkit, all of which can be accessed directly within the Red Hat OpenShift Data Science dashboard.
Enhanced MLOps
A data science projects UI has been added to the AI cloud service, which provides the ability to organise model development and artefacts into multiple projects to foster increased collaboration. Extended MLOps capabilities such as data science model serving based on open source projects, will help with the production deployment of models across the various footprints where OpenShift can run. The initial underpinnings of model monitoring will allow data scientists and IT operations teams to track the health and performance of models as they are rolled into production.
HIPAA-ready support for healthcare organisations
Several Red Hat Cloud Services, including Red Hat OpenShift Data Science, are now HIPAA-ready and able to support HIPAA-covered entities in building healthcare applications. Built on Red Hat OpenShift, these HIPAA-ready Red Hat Cloud Services can help reduce operational complexity by allowing organisations in the healthcare industry to focus on what matters – building innovative solutions and applications that can improve patient care and services.
What’s next?
Red Hat OpenShift Data Science is currently available in AWS Marketplace and on Red Hat OpenShift Service on AWS (ROSA) as a limited release, with full general availability in the coming weeks. In the months ahead, Red Hat will continue to introduce new features and functionalities, including enhanced data science pipeline capabilities and support for additional public cloud services.
Each new feature will reflect our commitment to providing an easy-to-use common data science platform. This platform allows teams to consistently export models to production across all cloud and edge environments and provision data science projects, regardless of their underlying infrastructure.
As we prepare to introduce new capabilities, you can learn more about how long-time customers like Boston University have benefitted from Red Hat OpenShift Data Science since its inception. The University has been using the solution since last year in conjunction with Red Hat OpenShift Service on AWS to create new educational tools for hundreds of students in the school’s Computer Science and Computer Engineering departments.
We look forward to supporting our customers as they grow their data science initiatives, now and in the future.
