Traveling Through a Network
Packet Data
Data is sent through a network in 'packets'. Data is taken and disassembled and placed into smaller chunks called packets. Each packet has a unique identifier added to it in addition to source and destination IP addresses. A packet is sent off to its destination computer. Along the way, each packet can take a different route, hopping from one server to another to find the best route to get to its destination. When the packets reach the recipient, the computer will re-assemble all of the packets into a complete message.
- Ping is a simple command to see if you can reach an end device. If this times out, the device isn't reachable on the network. This could mean a connection is down or not connected. Ping will also let you know how long a connection takes and if any packets are lost. Traceroute is a more comprehensive tool that will show you every router along the way. Traceroute can help determine where network congestion and slowdowns occur. A ping or traceroute can fail if part of the network is down. It can also fail if someone has set up security protocols so that pings are not answered.
Tools for Testing
The tools ping and traceroute can be used to diagnose issues from one point to another. The ping tool is used to verify if a destination address is reachable across a network. it measures and records the round trip time of a packet and whether there are any data losses along the way.
The traceroute tool is used to view all of the hops over routers along the way from their source to their destination. It measures the time it takes for each hop. This can help diagnose points of failure.
Pinging Common Websites
With my first ping at google.com, I let my terminal window ping the site ten times. My terminal app would keep pinging until I manually cut it off by typing Control-C. There were a total of 10 packets transmitted, ten packets received, and a 0.0% packet loss. The fastest was 23.641 ms, and the average time was 31.658 ms. The longest time was 70.138 ms, and the standard deviation was 13.105 ms.
I then decided to ping Amazon sites in the U.K. and Australia. I pinged each domain ten times as well. There was no packet loss to either site. The U.K. site had nearly taken nearly twice as long.
For amazon.co.uk, the shortest time was 151.570 ms, the average was 160.641 ms, the longest time was 210.253 ms, and the standard deviation was 16.725 ms.
For amazon.co.au, the shortest time was 87.707 ms, the average was 92.617 ms, the longest time was 103.738 ms, and the standard deviation was only 4.348 ms. The Australian site was not only quicker, but it was more consistent in speeds as well. While we are looking at what appear to be large numbers, they are only milliseconds. Both pages load very quickly in a browser for an end user.
Using Traceroute on Common Websites
I used traceroute on both Amazon sites. There was a warning on each that the domains had multiple addresses. Traceroute chose a specific I.P. address for each domain. The Amazon U.K. site had 13 hops before it landed on a server with the amazon domain name attached and then started to time out on the 14th hop. Remember that pinging the Amazon Australia site was much faster than the U.K. site, but when I used traceroute, I found that the Australian site went through many more routers to get there. The final server was at the 21st hop, although it did not have an Amazon domain name. On the 22nd hop, it started to time out.
Australia is roughly twice the distance from my location than it is to the U.K. There were certainly more routers along the way. The speed going to Australia was about half the speed going to the U.K. My conclusion is that there are either faster network cables or faster routers between the U.S. and Australia than the ones from the U.S. to the U.K.
A colleague pointed out that a "new cable was laid between the US and Australia around 2019 called the Southern Cross Cables. It also appears there hasn't been new cable laid between the US and England since 2003. That could explain the discrepancy."
