Latency vs Throughput

First let us understand one example and then we will proceed towards the actual definition:

 

Suppose, you are travelling from City A to City B and you have to cross a bridge which can carry 1000 vehicles per hour but due to heavy traffic on the bridge it can carry only 100 vehicles per hour.

 

Summarizing the situation:

- The maximum capacity of a bridge is 1000 vehicles/hour 

  We can call it as Bandwidth

- Due to the heavy traffic, it reduces to 100 vehicles/hour

  We can call it as Throughput

- And Time Taken from reaching city A to city B is 20mins (Let us assume).  

  We can call it as Latency

Now Let us learn the actual definitions:

 

1) Latency (delay):

The time taken by a data packet (or request) from source to reach the destination.

 

• We can measure the latency in following units: microsec, sec , millisec
• When Latency is Low, then the system is Good.
• When Latency is High, then the system is Bad.

 

2) Throughput:

The no of data packets that are processed during a specific time period.

 

• We can measure the Throughput in following units: bit per sec (mbps, gbps)
• When Throughput is Minimum, then the system is Bad.
• When Throughput is Maximum, then the system is Good.

 

Generally we should aim for low latency and maximum throughput.

 

Take a look at below image, it provides an easy illustration about both these concepts:

Image Credit: https://www.comparitech.com/net-admin/latency-vs-throughput/