IBM CodeFlare simplifies the move to hybrid cloud
Enterprises are relying on data more than ever before, but that can come at a cost in terms of the time spent on building and managing the infrastructure to handle it.
In order to streamline the integration and efficient scaling of these big data and AI workflows into hybrid cloud environments, IBM Research is launching CodeFlare.
CodeFlare is Built using Ray, an open-source technology from UC Berkeley. It automates and speeds up the training, processing and scaling of ML models, enabling developers to focus more on data insights and less on system configuration.
CodeFlare is open source and available through the IBM GitHub repository. IBM is releasing CodeFlare's open-source code along with examples that run on IBM Cloud and Red Hat Operate First.
To create a machine learning model, researchers and developers first have to train and optimize the model. This can involve data cleaning, feature extraction, and model optimization. CodeFlare simplifies this process using a Python-based interface for what's called a pipeline -- making it simpler to integrate, parallelize and share data. It aims to unify pipeline workflows across multiple platforms without needing data scientists to learn a new workflow language.
CodeFlare pipelines will run on IBM's new serverless platform IBM Cloud Code Engine, and Red Hat OpenShift. It also aims to make it easier to integrate and bridge with other cloud-native systems by providing adapters to handle event-triggers (such as the arrival of a new file), plus load and partition data from a range of sources, such as cloud object storage, data lakes, and distributed filesystems.
You can find out more about CloudFlare on the IBM Research blog.