Cisco upgrades target Kubernetes, cloud, and AI/ML

Cisco has added new features to its core cloud and computing packages to better manage and support distributed applications.

The enhancements affect Cisco’s Intersight cloud management system, UCS X-Series server and HyperFlex hyperconverged system.

The idea is to provide tools that offer flexibility and manageability while increasing performance and reducing the costs of modern cloud-based apps and workloads, said DD Dasgupta, vice president of Cisco’s Cloud & Compute Product Management group. 

Cisco is extending its Intersight support for Kubernetes, which had managed only Cisco Kubernets and only on-premises. The upgrade, Intersight Kubernetes Service Attached Clusters, provides a single place for IT pros to look at and manage all their Kubernetes clusters, including those running on Microsoft Azure and AWS cloud platforms, with plans to add Google Cloud support in the future.

“Many times IT doesn’t know these clusters exist because some part of the organization spun-up a Kubernetes cluster for some specific workload and didn’t tell anyone. This Intersight feature will eliminate that from happening and causing problems,” Dasgupta said.

Another new feature, Cisco Intersight integrations with Amazon Elastic Compute Cloud (Amazon EC2) lets customers manage the inventory and automation of virtual machines running in AWS cloud and on-prem environments. “Our goal is to enable our customers to achieve consistency of operations and unified experiences across data center and different public clouds,” Dasgupta said.

On the server side, Cisco has integrated graphics processing unit (GPU) nodes into its flagship UCS X system and bolstered the unified switching fabric of the X blade system to support up to 14.84Tbps of aggregate bandwidth.

Introduced almost a year ago, the UCS X system is the fastest ramping server in Cisco’s history and features a modular hardware architecture that can be upgraded over time to support new generations of processors, storage, nonvolatile memory, accelerators, and interconnect.

The idea is for the system to support a wide-range of duties  from virtualized workloads, traditional enterprise applications, and databases to private cloud and cloud-native applications, Dasgupta said, and all of these tasks are controlled and managed by Intersight.

The new capacity and GPU support will help to better handle increasing and demanding AI and ML use cases, Dasgupta said. “Big Data forces have very different traffic pattern in and out of, you know, the compute cycles and the GPU cycles—GPU cycles within and outside the box,” Dasgupta said . “We’re all seeing the rise of container-based workloads, too, as the world grows more virtualized containers. GPUs are being required for many of these modern workloads.”

Cisco has also created HyperFlex Express, a new lower priced configuration of its HyperFlex hyperconverged system that offers computing, networking and ration called storage resources in a single system.

“HyperFlex Express is going to be almost 50% lower than any HyperFlex that we’ve sold in the past,” Dasgupta said. “We’re also coupling it with some very specific partner programs that lets customers deploy the system by almost 80% faster. By simplifying the product in these partner programs, we believe HyperFlex Express lowers the entry point for adopting hyperconverged solutions.”

Cisco Intersight Kubernetes Service Attached Clusters and Cisco Intersight integrations with Amazon Elastic Compute Cloud (Amazon EC2) are available now. GPU Node with Cisco UCS X-Fabric Technology will be available by May.  Cisco HyperFlex Express is available now.