Forum Discussion
- legacy1Alessandro Volta
what we don't know is if it upstream or downstream with that latency test I'm guessing downstream
Why yes their would be excessive latency given how the test works it floods the downstream with the hub rate limiting and buffer your connection.
It why VM should not be doing auto speed tests on hub in router mode.- horsemanAlessandro Volta
legacy1 wrote:what we don't know is if it upstream or downstream with that latency test I'm guessing downstream Why yes their would be excessive latency given how the test works it floods the downstream with the hub rate limiting and buffer your connection.It why VM should not be doing auto speed tests on hub in router mode.
Whilst seeking to keep an open mind and thus still appreciating any post unaccompanied by any evidential data the use of “we” and “guessed” does’nt add any credibility……. IMHO….
Since “my” continued results of latency exceeds 10s (and jitter >200ms) are “apparently” using “SamKnows” servers for Hub3 yet also WiFi but unsupported (yet) by any other non VM. Connect diagnostic then obvious “comment/query” will also include:
1. Is this a SK ISSUE?
2. Or a VM “overlay” on specific Hub?
3. Does a v11 Hub3/different NMS soft/firmware/or Hub4-5 experience similar Excess latency/jitter?
Upgrade to IOS 16.3 still results in same excess….. AND ….
AFTER 7 ELAPSED,5 WORKING DAYS READERS WILL NOTE THAT VM FT HAVE NOT YET REPLIED!
Enquiring minds would still appreciate the courtesy of official VM acknowledgement of symptom even if potential answer or escalation exceeds FT diagnostic expertise and/or authorisation!
- David_BnForum Team
Thanks for your post on our Community Forums horseman and welcome back to our Forums.
Sorry to hear of the latency issues that have been flagged on the Connect Application.
Can you please confirm if you experience latency on just this one device or multiple?
Having ran a flow on your hub, no issues appeared to be present.
Kindest regards,
David_Bn
- Client62Alessandro Volta
Kurzsignalbuch verloren?
- SephirothAlessandro Volta
Nein - Mann hatt Enigma benuetzt!
- asim18Rising star
OK I have read this entire thread and I cannot make heads or tails of anything.
1) The VM connect app is certainly broken. But does it need to be fixed? Not in my opinion. I'd rather VM completely forget about this app and focus their money and attention on their network infrastructure, not some pointless app.
2) Latency is typically measured using ICMP packets, this is a completely different protocol to usual protocols such as TCP/UDP/SIP. ICMP will typically be given the lowest priority by networking hardware including the hub, which is why you are seeing ridiculously high latency.
3) If you need to measure latency, use the ping command and ping bbc.co.uk or google.co.uk, or other important peer IP. Measuring latency to your local speedtest server using a proprietary mobile phone app is pointless. You need to measure your latency to an actual endpoint which you are communicating with using a proper ping utility, definitely not using an iPad app lol.
- TudorVery Insightful Person
"If you need to measure latency, use the ping command and ping bbc.co.uk or google.co.uk," where will that get you? Ping uses ICMP.
- asim18Rising star
It'll get you past the dodgy VM Hubs. The issue is not ICMP. The issue here, in my opinion, is how the Hubs handle ICMP when the Hub itself is either the source or destination of the probe.
When you use a separate LAN device, that LAN device becomes the source as far as the Hub is concerned. The ICMP packet is simply another packet it needs to forward; it won’t respond itself. I have not experienced any issues when the VM HUB is simply routing/forwarding ICMP packets and is not the source/destination of the ICMP probe.
This can be tested using the Hub3's built in ping utility. Try to ping a local IP, every now and then you will get some ridiculous figure like "14323ms", this is not a network issue, it's because the hub's firmware...
You can also try pinging the Hub itself, locally. All Ethernet devices on my network are less than 1ms, most are around 0.4ms. But the Hub3 hovers around 3ms.
- SephirothAlessandro Volta
Nothing wrong with ICMP. Indeed, Tracert tells you exactly which node(s) is/are busy with priority traffic - which is a clue as to what's going on.
But yes, the CONNECT app is useless and do let Horsey have his fun with VM.
Related Content
- 2 months ago
- 3 months ago
- 17 days ago
- 8 months ago
- 20 days ago