Google announces enhanced AI Platform Prediction

Google announces enhanced AI Platform Prediction

The product now includes new hardware options, increased reliability, flexibility, security and ease of use.

Google has announced the general availability of AI Platform Prediction, a service that allows companies to train their machine learning (ML) models at scale, to host their trained models in the cloud, and to use their models to make predictions about new data.

The service is based on a Google Kubernetes Engine (GKE) backend designed for improved reliability and flexibility. It does this through new hardware options (Compute Engine machine types and NVIDIA accelerators), reduced overhead latency, and improved tail latency.

Cloud-based architecture adds security

The new architecture allows companies to create a perimeter around the ML models to isolate them from the rest of their cloud environment. Administrators can then configure the perimeter to give the model access to only those tools and resources it needs to operate.

This in turn makes it more difficult for hackers to move around within the environment if they breach the corporate network.

The perimeter functionality is called VPC-Service Controls. These controls ensure that calls to the CAIP Online Prediction APIs are made from within the perimeter. Private IP will allow VMs and Services within the restricted networks or security perimeters to access the CMLE APIs without having to traverse the public internet.

New features enhance artificial intelligence flexibility

In addition to standard features such as autoscaling, access logs, and request/response logging, several updates improve robustness, flexibility, and usability.

For example, AI Platform makes it simple to deploy models trained using XGBoost / scikit learn models on high-mem/high-cpu machine types.

Additionally, the service includes Resource Metrics to measure characteristics such as GPU, CPU, RAM, and network utilization. These metrics can help make decisions about what hardware to use to minimize latencies and optimize performance.

Speed, ease of use and simplicity

Google engineers Bhupesh Chandra and Robbie Haertel explained the overall benefits in a blog post. “All of these features are available in a fully managed, cluster-less environment with enterprise support. No need to stand up or manage your own highly available GKE clusters,” they wrote.

“These features of our managed platform allow your data scientists and engineers to focus on business problems instead of managing infrastructure.”

Also read: What is Google Anthos? Is this the modern cloud infrastructure you’re looking for?