In the Cloud Performance is Instrumented as Cost

About 5 years ago, I was at a conference presentation where a slide contained this statement.  We all looked it at, thought about and said to ourselves 'yes, of course, that is obvious'.  However, it has taken until this year for me to have conversations that reflect that idea.

In the cloud, you rent hardware as an operational expense, rather than purchasing it as a capital expense. Every architectural decision you make has an immediate cost consequence. You can have as much performance as you are willing to pay for. This drives out different behaviour. 
  • If you are not short of CPU, you are probably spending too much on your Cloud subscription. 
  • If you are short of CPU, then you need to configure the resource manager to prioritise higher-priority processes over lower-priority processing. 
This session will look at how that played out at one customer.  They improved performance and saved money on their cloud subscription.   The single most important thing they did was to clearly and quantitatively state their definition of acceptable performance.  
That set the ground rules for sizing and configuring both their database and their application, implementing various database features, including defining a resource manager plan that reflected the customers stated priorities.

I am giving this presentation online for SPOUG at 17.00CET on 12th March.

The presentation is available in a pdf file in pdf format.