More and more cloud-computing experts are talking about "multicloud". The term refers to an architecture that spans multiple cloud environments in order to take advantage of different services, different levels of performance, security, or redundancy, or even different cloud vendors. But what sometimes gets lost in these discussions is that multicloud is not always public cloud. As machine learning (ML) continues to pervade enterprise environments, we need to understand how to make ML practical on multicloud -- including those architectures that span the firewall. Let's look at three possible scenarios.

PowerAI is IBM's machine learning framework for companies that use servers based on its Power processors and NVIDIA's NVLink high-speed interconnects that allow for data to pass extremely quickly between the processor and the GPU that does most of the deep learning calculations. Today, the company announced that PowerAI now supports Google's popular Tensorflow machine learning library. While TensorFlow has only been available for a little over a year, it has quickly become the most popular open source machine learning project on GitHub. IBM's PowerAI already supported other frameworks and libraries like CAFFETheano, Torch, cuDNN, and NVIDIA DIGITS, but Tensorflow support was sorely missing from this lineup. IBM clearly sees the combination of PowerAI with Nvidia's NVLink interface and Pascal P100 GPU accelerators as a way to differentiate itself from the competition -- and in this case, the competition it is gunning for is clearly Intel (though it's worth noting that Intel and Google also recently teamed up to improve TensorFlow performance on its CPUs).