Enhancing the infrastructure options on its public cloud platform, Google has announced sole-tenant nodes for Google Compute Engine. The new instances will provide dedicated hardware for sensitive workloads.
In cloud computing, generally the virtual machine (VM) instances are shared by many customers, forming a multi-tenant environment. Google has designed the sole-tenant nodes to allow enterprises use hardware for their exclusive use. Each node will be associated with one physical server, and will be the only node running on that server.
The sole-tenant nodes will help companies meet the regulatory and compliance needs for sensitive workloads. The new instances will also be useful for customers who require more control over their cloud environments, and want to achieve the highest level of utilization.
Additionally, customers can specify the way instances are arranged on nodes, and separate the instances with sensitive workloads from the ones with non-sensitive workloads.
The search-engine giant will allow customers to launch instances of all shapes in terms of CPU and memory. Further, the algorithms will automatically select the best location across all nodes to launch the instance. Customers can also manually select the location for more control.
Google said that the instances launched on sole-tenant nodes will take advantage of live migration to eliminate the downtime during proactive maintenance.
Customers can get started with sole-tenant nodes by launching them from Google Cloud SDK (software development kit), and from Compute Engine APIs.
Also read: Google open sources Asylo framework for confidential computing
The sole-tenant nodes are currently available in beta. The pricing of these nodes will be based on amount of system memory and processing power used by customers. The nodes can be used on per-second basis, with a one-minute minimum charge.
