A comparision of public cloud pricing
There are many things to consider when thinking of moving workloads to the cloud. There is a common misconception that cloud is cheaper than whatever you are doing right now. There is a chance that it isn’t cheaper but there are other benefits to factor in to the price. However price is still a factor which is why I conducted my own cloud comparison.
It is nearly impossible to compare apples with apples when it comes to comparing virtual machine resources between different clouds. But I’ve tried to simplify the comparison by choosing a standard sized VM which would be suitable for a mid-sized Windows Server.
The comparisons are based on an unmanaged VM with 4vCPU, 8GB of Memory, 100GB System Partition, 1TB data partition and a basic backup.
When viewing these comparisons please be aware that not everything is equal. There are slight discrepancies between:
- The Number of CPUs allocated
- The speed of the CPUs allocated
- The amount of memory allocated
- The type of storage allocated
- The contention ratio of CPU and Memory
Other factors to consider are location of public clouds. Some cloud offering have different pricing based on regions etc. These comparisons are based on VMs being located in the APAC region. All pricing is current at the time of writing (20th August 2014) and can be reproduced by using the online calculators from each vendor.
From my comparisons, most IaaS offerings are close in price and in this example Softlayer was actually the cheapest. But I’m sure that if I changed the size of my VM there would be a different result.
All of these Cloud providers offer calculators online so you can work out your own pricing.
There are other studies going around comparing performance differences on these Cloud platforms which is a whole other story.
In Summary, Cloud isn’t cheap but it does provide you with features that aren’t always available in your on premises private cloud. Some good features and some bad, things to consider ; scalability, high availability, disaster recovery, performance, automation, data sovereignty, location, bandwidth requirements, contract lock in, spend commitment, compatibility, integration, data retention… (and the list goes on).