Traveling Through a Network
Ping, Traceroute & The Secret Life of Your Internet Traffic
What I learned from sending packets around the world
This week, I took a little networking road trip... well, my packets did. I ran ping and traceroute tests to three different sites: Google.com, Amazon.jp, and Google.kr. Even though these sites are scattered across the globe, the results were surprisingly similar… and the reason why is pretty cool.
Part 1: Ping - The Internet’s Quick “Are You There?”
Ping is basically the digital equivalent of tapping someone on the shoulder to see if they respond.
Here’s what I found:
Google.com
Packet size: 32 bytes
Sent/received: 4/4
Packet loss: 0%
RTT: 11–17 ms (avg 14 ms)
Amazon.jp
Packet size: 32 bytes
Sent/received: 4/4
Packet loss: 0%
RTT: 14–20 ms (avg 17 ms)
Google.kr
Packet size: 32 bytes
Sent/received: 4/4
Packet loss: 0%
RTT: 15–25 ms (avg 18 ms)
Even though Japan and Korea are much farther away than California, the latency didn’t skyrocket. Why? Content Delivery Networks (CDNs). Companies like Google and Amazon place servers all over the world, so when you ping “Google.kr,” your packets probably aren’t traveling to Korea—they’re hitting a nearby server in your region. That’s why the times stay low and consistent.
Part 2: Traceroute - Following the Packet’s Road Trip
If ping is a shoulder tap, traceroute is the full GPS route.
Google.com
Reached the destination in 11 hops. Hop 4 timed out - likely a router configured to ignore ICMP requests (a common security practice).
Amazon.jp
Reached the destination in 14 hops. Hops 7–13 didn’t respond, again likely due to routers blocking ICMP traffic.
Google.kr
Reached the destination in 14 hops with no timeouts. Hop 6 had a brief latency spike (52 ms), but everything stabilized afterward.
Part 3: What This Taught Me About Network Travel
All three sites behaved similarly, which makes sense once you understand CDNs. These networks distribute servers worldwide so your packets rarely travel to the “real” physical country. Instead, they hit the closest server in the CDN’s global network. That’s why Google.com, Amazon.jp, and Google.kr all responded with similar speeds.
Ping and traceroute together give you a powerful diagnostic combo:
Ping tells you if a host is reachable and how long it takes to respond.
Traceroute shows where delays or failures happen along the path.
Real‑World Problems These Tools Can Solve
Here’s where things get practical:
Packet loss troubleshooting: Ping different points (router, modem, ISP gateway) to see if the issue is local or upstream.
Slow VPN diagnosis: Traceroute can reveal which hop is causing the slowdown - often a congested VPN server.
DNS misconfigurations: If pinging a domain fails but pinging the IP works, you’ve got a DNS issue.
Firewall or security filtering: Timeouts in traceroute often indicate ICMP blocking, not actual connection failure.
Bonus: Different Types of Traceroute
Traceroute doesn’t always use the same type of probe:
ICMP (Windows tracert): Most common, but often blocked by firewalls.
UDP (Linux traceroute): Can bypass some ICMP filters.
TCP traceroute: Useful for testing paths to specific ports (like port 443 for HTTPS).
Different probe types can reveal different parts of the network path, especially when routers filter certain traffic.
Final Thoughts
This little experiment showed me just how much is happening behind the scenes every time we load a webpage. Ping and traceroute may seem simple, but they’re incredibly powerful tools for diagnosing connectivity issues, spotting bottlenecks, and understanding how traffic moves across the internet.
And thanks to CDNs, your packets don’t need a passport - they’re usually staying much closer to home than you think.
Comments
Post a Comment