A good cliché is that DR is achieved by 3P's: Planning, Performing, and Practice; of a strategy that meets predefined organizational goals and objectives for business continuity.
Because every organization is unique with the demands of the business, do not use a one size fits all approach for a solution. Each environment needs to be approached from the perspective of identifying an appropriate approach.
Design for high availability and disaster recovery first, then for performance and capacity.
Understand the costs associated with a DR plan, not just servers but manpower too. Remember that farm administrators are busy; they generally do not have spare time to be proactive and think about what should be done, unless senior IT management allocates time for them to do these activities.
Design from the bottom to up: server capacity, network devices, installation, content databases, and then site collections.