Everyone knows ping. But over the decades, the networking community has quietly built an entire family of specialized variants — each solving a problem that standard ICMP couldn't.
A few examples of why you'd reach for something else:
tcping — when firewalls eat your ICMP and you need to test port availability
arping — L2 diagnostics and duplicate IP detection, no IP stack needed
fping — scan a /24 in seconds, all hosts in parallel
OWAMP — when you actually need one-way latency, not just RTT
dnsping — when the slowness lives in your resolver, not the network
I put together a comparison table of the most useful ones, across protocol, OSI layer, platform, multi-host support, and root requirements.
The OSI layer column alone tells you a lot — if you're reaching for ping to debug something that lives at L4 or L7, you're probably using the wrong tool.
What about Google? Anyone has any insights on their unit economics since they own the models and the infrastructure (which is also custom TPUs)? Are they doing better or are they in the same money losing business?
That's the point. On the GPU side they use all the 5000+ cores to parallelize the algorithm (they use the hardware to its full potential). On the CPU side they use just one core (at least there is no mention around the cores used on the CPU). It's like saying a Camry beat a Ferrari in maximum speed, but you don't mention that the Ferrari was only in the first gear for that specific race.
How about if you did the same on 8 or 16 core CPU that can have much more than 16 GB of memory and is not as expensive to move data around its own memory?
If you use iperf for troubleshooting or testing you may have thought about the Raspberry Pi as an end point. The RPi is cheap and easy to deploy, but there are other single board computers out there with higher performance.
You might have seen the recent project on network speed testing on a Raspberry Pi which tweets to Comcast when the speed goes below 50 Mbps. This is an alternative or complementary option.
The Raspberry Pi software is installed as a Debian package. The data is displayed on a dashboard. You can receive alerts by email when the connection goes down, the Ping packet loss goes above x%, you can't access Facebook, or can't resolve times.com. No tweeting capabilities though.
Hi everyone, I am the author of the post. Network tomography is a very intriguing idea and very promising. However, it hasn't taken off outside the academic/research realm.
I am wondering what networking folks think about it.