Sole-tenant nodes to Google Cloud’s rescue

    6759
    SHARE
    Google Cloud

    In a bid to put Google in line with Amazon and Microsoft, Google Cloud has announced that it’s using sole-tenant nodes in its Compute Engine to help customers run multiple instances of various sizes without sharing hardware hosts with other projects.

    In regards to compliance regulations, the cloud giant hopes that the sole-tenant nodes will help regulated industries separate their compute resources in the cloud.

    Accommodating virtual machines

    The n1-node-96-624node type is available in multiple zones, and has 96 vCPUs and 624 GB of system memory.




    The node size can accommodate virtual machine (VM) instances with up to 96 vCPUs and 624 GB of system memory, and customers will be able to fill the node with multiple smaller VM instances with various sizes including custom machine types and instances via extended memory.

    The instances that the nodes are run on must have at least two vCPUs, and when a node is full, customers will not be able to schedule additional instances.

    Compute Engine features

    According to Google Cloud’s blog post, the sole-tenant nodes are compatible with the existing Compute Engine features:

    • If your node’s host system requires maintenance, the node and all of the instances on the node continue to operate while they live migrate to updated host hardware.
    • Sustained use discounts and committed use discounts reduce the costs of your sole-tenant nodes. Read sole-tenant node pricing to learn how discounts apply to sole-tenant nodes.
    • VPC networks work with instances running on sole-tenant nodes the same way that they work with normal VM instances. You can use VPC networks to establish network connections between sole-tenant instances and normal VM instances.
    • Use custom machine types or predefined machine types to create instances on your sole-tenant nodes. Because you already pay for the vCPUs and system memory of the node itself, you do not pay extra for these instances.
    • Create managed instance groups on your node groups. Your managed instance groups can use autoscaling while running on sole-tenant nodes, but the node groups cannot automatically scale.
    • Combine VMs with multiple machine types on each node. You can use a mix of different machine types and custom machine types on the same node until the node reaches its vCPU and system memory limit, which is defined by the node type.

    Companies that aren’t highly regulated may also find nodes helpful for picking their location setting and deciding where they want their instances to run via user-defined labels.

    Written by Leah Alger