Defense in depth is a security strategy that calls for controls to be applied in multiple layers. Similar controls may be applied redundantly, or controls of different types may be applied in line with one another.
Defense in depth is generally regarded as a best practice. Defense in depth is sometime contrasted with perimeter security, which in its purest form concentrates all security controls at the network edge. A colloquial way of describing the difference in these approaches is to call defense in depth the "onion" model and to call perimeter security the "lollipop" model.