Cisco’s latest unified computing system (UCS) server targets artificial intelligence (AI) and machine learning in the enterprise. The company is also validating machine learning tools from Anaconda and Google’s Kubeflow as well as software from Cloudera and Hortonworks on the new server.

The new C480 ML M5 uses two Intel Xeon Scalable central processing units and eight Nvidia Tesla V100 Tensor Core graphics processing units (GPUs). These GPUs accelerate machine learning software stacks.

Customers across all industries are experiencing a “tremendous explosion of data,” said Cisco’s Todd Brannon, marketing director for data center solutions. He cited a Forrester report, which predicts “insight-driven businesses” — these are companies that use software to create a competitive advantage and help them make use of all of this data — will by 2020 steal $1.2 trillion annually from their less-informed peers.

“If data is the fuel for the fire, there is plenty of fuel developing,” he said.

But deep learning also requires software tools to make sense of the data. To that end Cisco is working with software providers and other technology companies to validate machine learning tools on top of the new server. “We’re working to demystify these AI and machine learning stacks, working with ISVs to curate top to bottom stacks,” Brannon said.

This includes contributing code to the Google-led open source project, Kubeflow. Companies running Kubeflow on the new UCS server will be able to use the same machine learning tools on-premises and on Google Cloud.

“We believe the power of machine learning should be available for all organizations, whether in the cloud or on-premises, and we’re excited to continue our collaborative efforts with Cisco,” said David Aronchick, product manager at Google Cloud, in a statement.

Cisco is also validating Anaconda Enterprise on top of the UCS server. Some 6 million data scientists use this software platform to develop and score machine learning models.

Additionally, Cisco has been working with big data software vendors including Cloudera, Hortonworks, and MapR. It recently published a validated design with Cloudera Data Science Workbench that integrates an existing big data CVD for Cloudera with deep learning frameworks such as TensorFlow and PyTorch.

With Hortonworks, Cisco is working to validate Hadoop 3.1 in a design where the UCS server is part of the big data cluster, said Han Yang, product manager for the UCS portfolio at Cisco. This will support Docker containers running analytic workloads such as Apache Spark and Google TensorFlow that require both CPUs and GPUs, he explained. “It’s able to support much more variety of software, and the particular software stack can oftentimes be curate by the data scientist themselves for particular needs,” he said.

“And NetApp and Pure Storage have expressed interest in working with us to extend the FlexPod and FlashStack [converged infrastructure] portfolios to create converged infrastructure that combines this deep learning system with flash storage,” Brannon said. Cisco partners with NetApp on FlexPod products and Pure Storage on FlashStack.