Reply
Highlighted
Aspirant

Latency increase vs. R7000

Anyone notice (at least on 5GHz band) that the pings to router seems to hang between 3-5ms?

On my R7000 it was 1-2ms.

If I connect to 2.4GHz I get 1-2ms again.

I do always set my preamble to short but that option is now missing on the R7500 on 5GHz band so not sure what it's doing or if that is related.

I've tested with Intel 7260-AC and 6300 series cards.

Just interested in some others.

This change isn't noticeable naturally BUT in the grand scheme of things when you are accessing remote resources internationally it all adds up.

Trying to figure out if it's chipset related (R7000 Broadcom vs. R7500 Atheros) or wave 1 vs. wave 2 or some missing setting(s) etc.
Message 1 of 6
Highlighted
Mentor

Re: Latency increase vs. R7000

Cybercare wrote:

This change isn't noticeable naturally BUT in the grand scheme of things when you are accessing remote resources internationally it all adds up.


In the grand scheme of things it's neither here nor there, and when you're accessing remote resources internationally it's of no relevance.

First - responding to ping requests is a secondary function and not one usually assigned a high priority - if the router's busy routing, it will ignore the ping request until it has time to respond, and this will result in higher response times - some brands of router don't even load the code required to respond to a ping until it's needed, and as a result, this brand will always drop the first request and respond to the following ones.

Second - when a ping (or other data) packet takes an average of 153 mSec to cross the atlantic do you really think you'll notice the difference between 153 & 155 mSec?

Trust me - as one that routinely uses a VPN for transatlantic access - it doesn't.

Ping is a diagnostic meant to verify connectivity - trying to use it to determine performance based on response time is an exercise in futility because you're attempting to use it for a task it was never intended.

Give a man a fish, feed him for a day
Teach a man to fish, feed him for life.
Message 2 of 6
Aspirant

Re: Latency increase vs. R7000

fordem wrote:
In the grand scheme of things it's neither here nor there, and when you're accessing remote resources internationally it's of no relevance.

First - responding to ping requests is a secondary function and not one usually assigned a high priority - if the router's busy routing, it will ignore the ping request until it has time to respond, and this will result in higher response times - some brands of router don't even load the code required to respond to a ping until it's needed, and as a result, this brand will always drop the first request and respond to the following ones.

Second - when a ping (or other data) packet takes an average of 153 mSec to cross the atlantic do you really think you'll notice the difference between 153 & 155 mSec?

Trust me - as one that routinely uses a VPN for transatlantic access - it doesn't.

Ping is a diagnostic meant to verify connectivity - trying to use it to determine performance based on response time is an exercise in futility because you're attempting to use it for a task it was never intended.


Don't take this as a flame/attack just wanted to respond in a little more detail.

You're also incorrect saying it's of no relevance for international resources. You do not have enough data points or details to make such an assumption because you have no idea how I or others may be using the connection and with what apps/services etc. Also don't assume all connections are a straight shot and that 150~ish ms is the worse one should see. It also doesn't all have to be ground base. In some instances depending what I am doing I may actually route over a sat connection at some point. Those are also very high latency where something else increasing it without a need further adds to the overall number.

Ping is and can be used for measuring latency. It's just not guaranteed to be accurate unless you fully control or know the forward and return path and also the devices handling it because as you say, some devices can deprioritize or drop it. It could also return via another path etc. etc. Ping de-prioritization is generally left for core/enterprise routers though. VERY few consumer devices do this or properly at that if they do. This is not the case/issue here though so not relevant.

You may routinely use VPN's for transatlantic access and not have an impact but that's just based on how you use the data as I stated above. I don't use my data for the same things as you. Things I do are latency dependent and people who say, well it's just 3-6ms more not anything to cry over are the same people who don't understand how to properly program, engineer or develop things. When everyone starts to accept this sort of stuff it results in sloppy/poor work. This is true for networking all the way to application developing.

Example:
I've got an ISP at another location who has us using a device that does similar for routing (not wireless). Well now place another device in path that does this and then anything outside your local side that may be doing similar. Guess what happens? It easily gets up. While even at 25-50ms latency people won't notice anything probably, when you are in the hundreds, depending what you are doing, adding 20 more can be noticeable as you are already high.

My point of this post was to keep it simple and try to gather more data. I could have started off and avoid this I guess if I said the latency is also confirmed beyond just ping but by endpoint connections measured via other means. I wanted to keep it simple though with a method that I confirmed can easily duplicate it in a manner that anyone with little networking experience could do.

It also brings into account jitter, jumping up and down between 3 and 6 just locally adds up with internet jitter too, which can also be noticeable on real-time streaming.

This is all just high level general still. We could probably both go into far more depth but that gets OT.

I just want to know what other people are seeing with what equipment and settings to see if we can identify a possible root cause is all as I have limited access to other clients to do this Smiley Happy

I do appreciate your response though and understand why you posted it. It just didn't apply to me in this case.
Message 3 of 6
Highlighted
Master

Re: Latency increase vs. R7000

Hi,
OP observed it right. R7500 is slower in ping rsponse compared to R7000.
I only use one server for ping test about 200 Km away from my home. R7000
I had always gave better numbers compared to R7500 I have now.
Message 4 of 6
Highlighted

Re: Latency increase vs. R7000

I to have experienced extra latency since going from a R7000 to a R7500, hope this can be sorted through firmware.

As the R7000 felt quicker in real world despite the lower spec.
Message 5 of 6
Highlighted
Virtuoso

Re: Latency increase vs. R7000

Did you experience this latency issue with the latest firmware?
Message 6 of 6
Top Contributors
Discussion stats
  • 5 replies
  • 8713 views
  • 0 kudos
  • 5 in conversation
Announcements