I have possibly gone through all the available articles published. But still not convinced.
I am trying to do capacity planning for our company. around 2000 employees will use SharePoint 2013 globally. User operations do vary from document authoring to wikis, to custom lists and Office Web Apps.
All the articles on the internet suggest that I should focus on RPS rather than number of the users. However, they provide RPS calculating formula to calculate RPS. In one of the TechNet articles, it is stated that 4 cores CPU and 14 GB memory server can generate around 50 RPS. What does this really mean? They have not mentioned the user load at all? What RPS does give me here then? RPS is the throughput that server generates in a sec. but 50 RPS looks too low..
I am using my common sense now, a page view, say home page, can make around 30-40 requests to server depending on the components on the page. How does this scenario fit above statement?
Has anyone done proper real world capacity planning? If so, where did you start? Did you get it right first place? How did you monitor the performance post-release?
Many thanks for your help and direction.