SUSE updated its container management and edge computing platforms with a pair of releases that boost security and automation.

Rancher Prime 3.0 is SUSE's flagship commercially supported container management platform. SUSE Edge 3.0 is, as the name implies, an edge computing platform that optimizes the Rancher platform for edge-of-network deployments. Rancher Prime had its last major update back in November 2023, while it has been over a year since SUSE Edge 2.0 was released. The two new platform updates coincide with the first day of the Cloud Native Computing Foundation (cloud native computing foundation (CNCF))) Kubecon Europe conference, which is being held in Paris.

The Rancher Prime enterprise container management platform adds new capabilities for platform engineering teams to deliver self-service platform-as-a-service (PAAS) to developers. It includes the general availability of Cluster API and Cluster Classes to enable automated, scalable PAAS delivery. The Cluster API provides a declarative approach for organizations to configure, deploy and manage Kubernetes clusters.

Security also gets a boost with support for supply chain levels for software artifacts (SLSA) certification and software bill-of-materials (SBOM), which helps to provide more assurance for applications and supply chain security. The Rancher Prime Application Collection is now generally available providing organizations with access to trusted enterprise-grade applications that will run on Rancher.

Rancher Prime is not a single technology but rather a collection of multiple open-source technologies to help enable secure cloud-native cloud deployments.

A core element of Rancher Prime is the Harvester open source technology, which is now at its 1.3.0 milestone. Harvest enables hyper-converged infrastructure (HCI) deployments, which are essentially highly optimized platforms that tightly integrate compute and storage.

In the Harvester 1.3.0 update, there is now virtual GPU support for virtual machines (VMs).

“Enabling efficient resource sharing in a multitenancy environment for GPUs is relevant for the increased demand for machine learning and AI [artificial intelligence] workloads,” Peter Smails, GM of enterprise container management at SUSE, told SDxCentral. “Previously we only supported passthrough of the entire GPU to a VM.”

Another core element of Rancher Prime is security for cloud-native workloads, which is getting a boost thanks to the NeuVector Prime 5.3.0 milestone.

“The key theme is strengthening zero trust,” Smails said.

Smails said that the NeuVector update provides valuable insights into external connections from a Kubernetes cluster. He explained that developers frequently require external connections for API services, external data sources or even internet-based open-source updates. These external connections can be to internal private networks or internet services, and it can be difficult for operations and security teams to know which should be allowed and which are suspicious. He emphasized that with the prevalence of embedded malware, backdoors and crypto mining, external connections from a cluster must be properly identified and secured.

In 5.3.0, NeuVector utilizes its layer 7 application inspection capabilities for all traffic, including DNS resolutions for fully qualified domain names (FQDNs), into IP addresses to first learn externally referenced hostnames/URLs and report on external connections.

“With this knowledge, security and operations teams can determine which connections should be allowed, which are suspicious and which should be blocked,” Smails said. “Allowed connections are then codified into the zero-trust rules for external access.”