ROUND-TRIP TIME (RTT) – Understanding Network and Security for Near-Edge Computing

Also known as latency, this represents the amount of time in milliseconds that a packet takes to travel from sender to receiver and back again.

Assuming the path between the sender and the receiver is a straight line, there is no way to reduce RTT because it is limited by the speed of light. The speed of light is fast, but finite – for every 100 km (60 miles) of distance a signal travels, 1 millisecond is added to its RTT.

Maximum segment size (MSS)

MSS is the maximum number of bytes that may be contained in the payload section of a TCP segment.

To calculate the appropriate MSS value, you must take the standard Maximum Transmission Unit (MTU) of 1,5001 and subtract any overhead involved.

1 The path MTU over the internet is almost always 1,500 bytes.

Here is an example

Layer 4: TCP header of 32 bytes2

2 TCP headers are 20 bytes natively, but these days, it is a safe assumption that TCP timestamps are in use, which raise the value to 32 bytes.

Layer 3: IP header of 20 bytes

Layer 3: 78 bytes overhead if an IPSEC VPN tunnel is being used

In this example, the MSS value would be 1,500 – 32 -20 -73 = 1,375 bytes

The MSS value is set in the receiver’s operating system. That value is announced by the receiver during the TCP three-way handshake, telling the sender it is the largest payload it can accept. This can be restricted to a desired value by a stateful device in the middle, such as a firewall or load balancer, via MSS clamping.

Mathis equation

The maximum throughput that can be achieved by a TCP connection can be calculated as follows:

Here, we have the following:

t is the effective throughput in bits per second

m is the MSS value in bits

r is the RTT value in milliseconds

p is the percent packet loss as a decimal number

Figure 2.2 below illustrates how the interaction of RTT and packet loss affects the effective throughput of a connection. The lighter line shows a loss of one packet in a million, while the darker line shows a loss of one in a hundred thousand.

Figure 2.2 – Effective throughput estimated by the Mathis equation

Figure 2.2 illustrates how the interaction of RTT and packet loss affects the effective throughput of a connection. The lighter line shows a loss of one packet in a million, while the darker line shows a loss of one in a hundred thousand. Remember, under these conditions, it does not matter how fast your internet connection is – these are hard limits. As you can see, it doesn’t take much packet loss to severely curtail an end user’s experience.