How does dynamic RAM differ from RAM?

Originally, RAM was an array of flip-flops, each being two transistors, with the ability to be set, reset, and interrogated. The contents were static, leading to the term Static RAM, and would stay valid as long as power was maintained.
As the demands for memory density increased, the design changed to that of a single transistor and capacitor. The capacitor stored the bit value, 1 or 0, and the transistor allowed interrogation, set, and reset.
The problem with a capacitor is that it discharges when interrogated, and also it discharges by itself due to internal leakage paths. This led to the requirement for refreshing the whole system periodically.
Refreshing is a process where each capacitor is interrogated, and then "refreshed" back to the set or reset state that it had. Since the capacitors are very small, this refresh interval was very short, in the order of milliseconds.
This meant that it was impossible to stop the refresh cycles without losing the contents of memory. Most modern designs do this automatically in hardware. Some designs utilize the processor and a periodic interrupt to do this. In that latter case, not even the processor can be halted without losing memory.