Member OP
I noticed inconsistent latency on my CloudCone VPS in Los Angeles. Ran hourly iperf3 for one week, 3 AM vs 3 PM local time. Results: 940 Mbps at night, 180 Mbps afternoon. Packet loss 0% vs 3%. Published full data at tokyo-kenji.example.com if anyone wants to verify my methodology.
¿why not test from multiple locations brother¡¡ is very strange that only you have this problem. Jaja maybe your ISP is the one with QoS, not CloudCone. I see this in the internet many times, people blame provider but is actually last mile!
They do not make them like that anymore—hardware you could hear thinking. Your methodology reminds me of those days. Consistent tooling over time matters more than any single test. I trust week-long data over provider promises.
I think CloudCone's ToS has something about acceptable use excluding excessive bandwidth consumption during peak hours. Their definition of "excessive" is not specified, which is a drafting problem. However, I believe they also reserve the right to do network management practices. Distinguish throttling from QoS: one is punitive, one is traffic engineering. You would need discovery to prove which.
RFC 2474 defines the Differentiated Services field you are probably hitting. AS64496 and below are not relevant here since you lack your own prefix. CloudCone likely de-prioritizes your traffic via DSCP at their edge during congested periods—standard practice, not malice.
[CloudCone edge] ----DSCP AF11----> [transit]
|
[your VPS] ----DSCP BE----> [congested]
Get an ASN or stop complaining. Everyone does this.
