Think the original poster has managed to stir up a religious debate.
The job is to manage/execute/spec the project.
Like it or not linux is an OS, its not going to earn your company money. Its support.
So put in in terms of money:
1. What are the switching costs and reduced:
a. what are the scripts you need to rewrite - in terms of man hours, anticipated outsourcing costs
b. What standard tools used in the organization will break, great if it is none but any experienced migrator will tell you otherwise
c. What is the cash saved in the life cycle of the project and what costs are saved in the extended life
d. What additional sales expected?
2. What is the value of your stream of updates( you have data, quantify it in terms of $$$)
a. In Engineering terms what is the cost of remedy and what is the cost of defect prevention (Past data where a patch is not present and how much activity the repair cost the organization, how much does it cost the organizaiton)
3. Disaster recovery
a. What happens if the OS becomes a blocker (e.g. some obscure library file provided by the latest RHEL on the dev pc and not on CentOS)
b. What happens if the applications behave differently(not likely but Ive seen it happen)
c. No offense, but what if you as key frontman(assuming here) are not able to solve a CentOS issue and you need help
d. You are proven right in the end you need a stream of updates, so how much will it cost to setup one
4. Suppliers
a. Do your organizations or partners have the capability and experience to implement rapid/ acceptable deployment for OS (provided you need multiple farms)
b. What will they charge and feed it back to 1
5. Customers
a. Do you marketing people go around promoting quality and talk about your RHEL, imagine the liability if it your customer faces an issue and get pissy
b. Will they accept CentOS, redhat has a lot of pull in the enterprise linux world. Like it or not, people will not like it when you switch away(like you!)
c. Will your customers choose another supplier if you switch, customers can be fickle, if they can switch on color of a GUI they can swith on an OS
6. HR needs (not really your call from the sound of your post , but still its in your domain and you are holding the bucket, so better to voice out)
a. Do your anticipated maintenance staff have the proper staff and certifcation needs(some organizations require a certain % of staff to be certified)
b. Do you need to hire more staff(project/contract/temps) to enable this project
This is a very short summary of what you can do(ill write more if you pay me =) ).All feeds back to 1. And ultimately its the CIO's call as some posters have made the point.
Ive seen organizations dragged down by such issues, where one engineering group goes off pushing their own distro and another group pushes their own. Lots of wasted resources and time. If your CIO makes the call, he makes the call. If you as SME know the difference, present in a way so that he can make the call. Supporting redhat will not help your organization or you. Put it in neutral terms and show what RH has to offer and if you got time pull in other companies. If he is open to CentOS why not suggest the full spectrum and let him make the decisions. pad it with costs from trade magazines(alright bad source but still better than nothing) , studies and most importantly your company's history. His call, his decision , his responsibility. In any case, if it hits the fan, you are covered and your organization has a plan from day 1. Good luck