Real Time Communication Illusion

Software engineers have agreed on a standard latency to consider real time since study shows that the human psychology can perceive a conversation as real time if they occur within approximately 100 milliseconds (0.1 seconds), if a system responds faster than this, users generally do not perceive any lag. This is practically real-time for many applications.

Actual real time communication is physically impossible since latency can not be gotten to zero because of some reasons.

Understanding the Reasons

• Ecoding/Decoding: Data needs to be encoded(converted into signals) at the sender’s end and decoded back at the receiver’s end. This takes time.

• Speed of Light: Speed of light is the most fundamental limitation since all forms of communication, whether electrical signals in wire, light in fiber optics, or radio waves through the air, travel at a finite speed, which is the speed of light or slightly slower in a medium.

• Software Processing: Application systems, application protocols (like TCP) all add processing overhead, which also implements queuing for when network traffic is high to not overload the system.

Conclusion

There are acceptable latency for different domain of applications:

Voice/Video Calls(VoIP): Latency under 150ms(one-way) is generally considered crucial for smooth natural conversions and above 300 ms, it becomes noticeable and disruptive.

Online Gaming: Latency under 30-50 ms is ideal for responsive experience, with up to 100 ms being acceptable.

Real-time communication aims for negligible latency, i.e delays that are so small they don’t significantly affect interaction.