Modern IT teams can now experience the direct access to their own cloud at the root level. This prescriptive to start, root-level model allows teams to modify VM attributes unlike public clouds can offer. Modify VM settings, make customer VM flavors, configure default setting, tune CPU to vCPU ratios, and so much more.
Key Features:
- The default hypervisor is KVM, the industry standard type 1 hypervisor. OpenStack has support for other hypervisors as well.
- Common VM images (OSs) are preloaded, plus easily create your own and remove images you don’t want available.
- Guided support for having your Windows VMs.
- Typical resource flavors come preconfigured and you can create your own and remove ones that don’t fit for you.
- Tune your CPU and RAM to vCPU and vRAM ratios to optimize your costs. Most cloud compute in public cloud is underutilized though the cost is the same regardless. You can dramatically lower cost and/or improve performance by understanding actual utilization.
Sizing Your Private Cloud Core by Compute
RAM and CPU will be presented in your private cloud as vRAM and vCPUs. Your OpenMetal cloud will come with prebuilt virtual machine flavors typical to the public and private cloud marketplace. This is just to provide an easy starting point. As this is your cloud, you can configure the VMs and flavors to fit your situation. Your workloads may also be more or less intense than typical workloads. See below for how to estimate the intensity of your workloads.
To get the best value, we recommend either testing against a trial cloud with a simulated workload or stepping into your private cloud with a portion of your workload. Once a workload is running in the cloud compute/virtual machines it is trivial to see the actual utilization from the hardware node.
Compute Workload Intensity Estimation
These workloads are often net low resource consumers:
- Development pipelines that have simulated production environments without the production traffic.
- Non-production Kubernetes deployments as Kubernetes will claim resources as if it is a production environment.
- Part time usage, particularly off hours, workloads.
These 4 workloads are often high resource consumers:
- Bulk rendering of images or video.
- Simulations or other heavy calculation based workloads.
- High traffic applications intentionally made dense.
- Automated tests that require full environment creation on the fly.
- Large scale data manipulation including big data processing and storage.
If you know your workloads are either CPU or RAM dominant and don’t match the typical ratios, you may need to calculate your capacity needs manually.
OpenMetal Cloud Compute FAQs
What is the OpenStack Compute Project?
OpenStack Nova is the OpenStack project that provides a way to provision compute instances (aka virtual servers). Nova supports creating virtual machines, and bare metal servers (through the use of Ironic) and has limited support for system containers. Nova runs a set of daemons on top of existing Linux servers to provide that service.
What tools are available for using OpenStack Nova?
Horizon, OpenStack Client, and Nova Client are tools available for Nova.
What makes up the OpenStack Nova architecture
Nova requires the following OpenStack services for basic functions such as Keystone, Glance, Neutron, and Placement. It can also integrate with other services to include persistent block storage, encrypted disks, and bare metal compute instances.
How does Ceph work on the on-demand OpenMetal private cloud?
- Nova-API daemon is the heart of Nova with the purpose to accept and fill incoming API requests.
- Nova-scheduler takes a virtual machine instance request from the queue and determines where it should run.
- Nova-compute accepts actions from the queue and performs. one or a series of virtual machine API calls to carry them out while updating the state in the database.
- Nova-volume manages the creation, attaching, and detaching of persistent volumes to compute instances.
- Nova-network worker daemon accepts networking tasks from the queue and performs system commands to manipulate the network.
- Nova-queue provides a central hub for passing messages between daemons, and Nova-database stores most of the configuration and run-time state for cloud infrastructure.
More OpenMetal Platform Components
Block Storage
Resilient, scalable, and highly available block storage is available by default to your virtual machines or containers.