Pricing Stability in the Cloud Era
Written By: John Chenard – Chief Technical Officer at Provdotnet, LLC
Public Cloud services are indeed a tool that every organization in need of on-demand compute, storage and connectivity should be using. The user controls and flexibility are unparalleled and built to provide immediate satisfaction. This unfortunately can lead to server sprawl and under-utilized systems that build up the overall monthly cost. Having worked intimately with most Cloud platforms I have developed a few strategies around maximizing dollars spent and building predictability.
The Uncertainty of Public Cloud
The leading public Cloud infrastructure services are priced based on consumed resources. CPU, memory, disk space and data transfer are the main items of measure. To complicate things a bit you also need to factor in performance tiers such as fast disk versus capacity disk and optimized connectivity versus best effort. Most of the service providers have calculators to determine current pricing although to obtain predictable, monthly costs, one must have an accurate estimate of each workload prior to migration and an estimate of growth over time. Unfortunately, few have these estimates since they weren’t necessary with on-premise infrastructure. As a result, buyers of IAAS often experience sticker shock when the monthly bill turns out to be much higher than budgeted.
Certain workloads are ideal for Public Cloud hosting. These workloads can be easily governed and possibly placed on hold if unused for periods of time. Other workloads are mostly static and are not prone to dynamic growth. Below is a sample list of these workloads:
- Mostly static web content hosting
- Simple database services with moderate data sets.
- Well governed development environments
- Limited Proof of Concept projects
- Copies of production workloads to spin up if issue arise.
As a rule of thumb, if your workload has large data that changes often, has CPU intensive demands at least a few hours a day or needs a large memory foot print you should look to keep that system off of Public Cloud. Additionally, it is usually safer to model a complex system off the Public cloud and move it once it is well understood.
Private Cloud as a Service
If an application’s workload and future usage patterns are not well defined, private cloud services may be your solution to deliver stable, predictable prices and monthly costs that business and IT executives want.
With private cloud services, the provider commits specific infrastructure to the buyer for a fixed monthly fee. The service fee can be locked in for a given term and the buyer receives defined and proven services while avoiding the complexities of managing infrastructure. The buyer also benefits from the flexibility to request additional capacity at a negotiated rate.
Another benefit to the private cloud is that the infrastructure used to provide the cloud service is dedicated and is not shared between other, unaffiliated organizations. This is unlike many of the solutions in the public cloud where resources are shared between different organizations. Having dedicated infrastructure can greatly reduce risk and uncertainty in an organization by allowing the buyer to control/specify many details of the underlying infrastructure, such as geographic location, manufacturer and performance specs.
Some Questions to consider when determining your company’s future cloud strategies:
- What are the future plans of your business? How agile/flexible does your infrastructure need to be to accommodate changes such as M&A activity or rapid growth? If the agility requirements are high and monthly cost stability are required, private cloud services are the best alternative.
- What is the size, talent levels and experience levels of your infrastructure team with public cloud services? If there are gaps, private cloud services may be a better alternative.
- What regulatory requirements does your organization face? If these are high (e.g. individual health data), public cloud services may not pass compliance audits.