JMeter sends requests immediately, with no latency between samplers or requests. Your server will be overburdened if you perform load/stress testing immediately. Then it won't be able to PROVIDE you with realistic FINDINGS and won't be able to mimic real-world user traffic. JMeter Timers are the answer to all of these ISSUES. To apply a wait between each sampler/request, a timer element can be added to a test plan.

Both Gaussian and Poisson Timers are based on a mathematical FORMULA with a constant delay and offset. 

• JMeter Gaussian Random Timer is used to produce and add a random delay before executing a sampler, as the name implies. Each user request is delayed for a random period of time using the Gaussian Random Timer element. Based on the Gaussian CURVE distribution, it has a random deviation around the Constant Delay Offset. The sum of the Gaussian distributed value (with mean 0.0 and standard deviation 1.0) times the deviation value you choose and the offset value is the delay (think) time.
• JMeter Poisson Random Timer is also used to produce and add a random delay before executing a sampler. The Poisson Distribution Function is used to create this timer. The total of the Poisson distributed value multiplied by the defined lambda value and the offset value is the delay (think) time.