- iperf3the standard throughput test
- BDPbandwidth × delay = the window you need
- 1500bytes: the usual Ethernet MTU
- RFC 9111today's HTTP caching rules
- Testiperf3 throughput testingInstall iperf3, run TCP and UDP tests, read the results correctly and find the real bottleneck.
- TuneTCP tuningWhy a 10 Gbit/s link can still give 300 Mbit/s transfers, and how buffers, window scaling and BBR fix it.
- CacheSquid and web cachingConfigure the Squid proxy, set up interception, and understand how HTTP caches and CDNs decide what to keep.
- MeasureMeasurement toolboxping, mtr, tcpdump, Wireshark, perfSONAR and RIPE Atlas: what each tool tells you and when to use it.
- DiagnoseSlow network? Step by stepA methodical checklist from the first ping to the packet capture.
- HistoryNLANR and the research networksThe laboratory where iperf and much early web-caching work began, and the vBNS era.
Quick start: your first iperf3 test
# on the server
iperf3 -s
# on the client: 10-second TCP test, then the reverse direction
iperf3 -c server.example.net
iperf3 -c server.example.net -RThen read how to interpret the numbers.
About this site
NLANR.net keeps the address of the former National Laboratory for Applied Network Research useful: the guides are new, written for today's tools and checked against official documentation. Old links to iperf, Squid and measurement pages lead to the matching guide, and the laboratory's old papers and files are still available at their original addresses.
The project is free, has no ads and is kept online with the support of IBAN. More about the site.