One of the main reasons to move to the cloud is the promise of a decreasing Total Cost of Ownership. But the calculation of the TCO (IaaS) is not always easy to do. Luckily some cloud-providers provide also the so-called TCO-calculators for comparising the TCO of your IT on-premise with the Cloud-solutions. I was already aware of Amazon’s AWS TCO calculator, but it came to my attention that Oracle has got on too. This post is about a small test to show some differences.
Let’s assume we own a company and our subject of attention is 4 outdated servers, with 2 quad core cpu’s each, and we want to know if there’s a business case to go to the cloud. And I will keep in mind I’m using TCO calculators from providers… 🙂
Start filling in the numbers, at first with the AWS’s TCO calculator. Default it is in ‘basic’ mode, with an option to go ‘advanced’. Used the basic mode for this purpose with default metrics. In the advanced mode you are able to finetune, adapting the figures to fit your own situation.
TCO Cloud calculator AWS
That’s quite an easy exercise. That’s it. The result at AWS:
But the total amount of server-costs seems quite high. About 40k per year. This can’t be right, can it?
An overview-look at the most interesting component, the server on-premises component:
A lot of Overhead has been calculated. Under the button ‘Calculations’ a more detailed story of this overhead:
Ok. Let’s take a look at Oracle.
TCO Cloud calculator Oracle
Let’s run the same test with Oracle’s TCO calculator. Here you must first take a decision .
A bit confusing for me, does it differ what to choose?
Clicked two of them: ‘Migrating your Data Center’ and ‘Lifting and Shifting Infrastructure Operations’. The only difference I could detect is slightly increased storage costs when chosing for ‘Migrating Your Data Center’.
Quite a different start anyhow.
No need to fill in the number of cores here….
Next step, storage 1 TB
Next step, default network requirements:
Next step, what kind of service:
Next step, what kind of storage service:
Next step, what kind of networking service:
Next step, the result:
This seems a lot cheaper, at least on-premises. Big difference: the costly part here is IT Labor instead of Overhead / Facilities at on-premises:
Offered Cloud solutions
A big difference what you get in the cloud: at Oracle we get one instance with 6 OCPU for our 4 servers with in total 32 cores:
And with AWS we get 4 instances with 32 virtual CPU’s:
This makes it very difficult to compare the two methodologies for me, and even harder to judge.
I ran another test with 1 server. So now we have two cases to compare.
What is the outcome per case ?
AWS TCO Calculator (basic)
|On Premise||Cloud||Overhead On-prem||Offered solution|
|1||186.512||147.534||75.089||4 x r3.8xlarge, 32 vCPU|
|2||140.260||37.925||61.540||1 x r3.8xlarge, 8 vCPU|
Oracle TCO calculator
|TCO Oracle||Cloud||Overhead On-prem||Offered solution|
|1||108.441||58.300||37.985||1 x OC6|
|2||34.996||16.398||19.240||1 x OC2M *)|
*) 2 OCPU’s with 4vCPU
But what does it mean? I tend to say ‘Absolutely nothing’. Most of all disappointment in the differences of the TCO on-premise. I know every company is different, but this is quite a big difference.
But this needs more investigation and examples to fill in and there are a lot of financial people to judge the usability of these TCO calculators. Just try it, pick the metrics you think is best for your company, it could be helpful after all.
Oracle TCO calculator: https://oracle.valuestoryapp.com/iaas/
AWS TCO calculator: https://awstcocalculator.com/
Machine shapes in Oracle Cloud: https://docs.oracle.com/cloud/latest/computecs_common/OCSUG/GUID-1DD0FA71-AC7B-461C-B8C1-14892725AA69.htm#OCSUG210