Cisco is expanding its Unified Computer System portfolio with optimized systems for artificial intelligence and machine learning using Nvidia's latest GPUs.
With the move, Cisco becomes the latest server and networking vendor to optimize systems for AI, machine learning and deep learning. Dell, HPE and Lenovo have similar systems. AI is being used for finance, healthcare, media, security, retail and manufacturing.
Cisco rolled out the UCS C480 ML M5 Rack Server, which is designed for deep neural networks. The UCS C480 ML M5 Rack Server is 4U with the latest Intel Xeon processors and 8 Nvidia Tesla V100-32G GPUs with NVLink interconnects.
In addition, the UCS C480 ML M5 Rack Server is designed to work with Cisco's various servers and HyperFlex systems with GPUs. The UCS C480 ML M5 Rack Server is configurable for CPU, network, storage, memory and operating system.
Cisco is also partnered with various big data players for UCS integration. Partners include Cloudera, Hortonworks and MapR. Cisco also integrates with frameworks such as TensorFlow and PyTorch.
The networking giant also said it is collaborating with Hortonworks to validate Hadoop 3.1 so UCS C480 ML is part of a big data cluster, storing machine learning data and supporting Docker containers running Apache Spark and Google TensorFlow.
The UCS C480 ML M5 Rack Server will be available in the fourth quarter.