Introduction

Properly scaling workloads is one of the biggest concerns for experienced cloud professionals. This is due to improperly scaled workloads on public clouds that result in high costs. A workload that is not scaled big enough will have poor performance, but a workload that is scaled too big will waste money. This is a waste of money due to higher infrastructure costs thanks to public clouds’ consumption-based billing model.

Cloud architects often bring this kind of thinking to the private cloud; however, private clouds do not usually have extra costs from consuming more of your available resources. When scaling your workloads for the private cloud, you can scale your workload as big as you might possibly think you need and then scale down to fit your actual level of activity.

In this article:

Why is Scaling Different for Private Clouds?

Scaling for private clouds is different because private clouds do not use consumption-based pricing like public cloud platforms. When you have a private cloud, you have a fixed set of resources that you are paying for whether or not you use 10% or 80% of the resources.

The only impact scaling workloads has on private clouds is taking resources from other workloads running within the private cloud. Visit our OpenMetal Private Cloud Pricing Estimator for more information.

Cost-effectively supply cloud resources at scale to your company and customers with Hosted Private Cloud, powered by OpenStack.

How to Scale Workloads for Private Clouds

When scaling workloads for private clouds, one of the best approaches is to scale your workloads as large as you think you will need to meet your performance needs. If you have the resources available, definitely take advantage of them to guarantee you can serve your end-users’ requests. It is much easier to scale your workload before you start receiving traffic or processing requests for your end-users.

After you have gone live or entered production, monitor your performance and usage carefully to determine how much capacity you will actually need. Once you have a good idea of your actual resource requirements, you can scale back your infrastructure to meet the real world needs of your workloads.

How to Scale an On-Demand OpenStack Private Cloud?

On-Demand OpenStack private clouds give you the flexibility of public clouds with the security of private clouds. Scaling workloads on an On-Demand OpenStack private cloud is the same as a traditional private cloud with a few additional considerations.

It is a best practice to obtain additional hardware nodes for an On Demand OpenStack private cloud ahead of a launch or entry into production if testing or research suggests you will see very high demand.

While you can add hardware within minutes to an On-Demand OpenStack private cloud as you need, any delay can result in a negative experience for users during periods of high activity on unprepared infrastructure.  Watching disk space utilization – see Ceph in our manual.

Provisioning hardware resources ahead of time will protect your infrastructure from becoming overwhelmed. On-Demand OpenStack private clouds make it easy to return unneeded hardware resources once the rush is over, or you have a better idea of how much hardware resources you actually need.  Take a look at our Tipping Point article to see that Private Cloud Costs  are often so much less that running some additional private infrastructure is easily still less than public offerings.

 

Looking for private cloud hosting, OpenStack as a service, or bare metal servers? A OpenMetal Cloud is a powerful private cloud solution that gives you the security and performance you need to successfully run your business. Learn more about the OpenMetal Infrastructure as a Service.

If you want to learn more about Private Cloud and OpenStack check out some of our articles:

OpenStack, But Easy

Launch a Cloud in 45 Seconds >>