Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Weaver]

I recently did a speed test over 4G on my iPad and got 200ms round trip time, as opposed to the usual 50ms for DSL. What do you get? This very inconsistent though; today, on mere 3G, I measured 77ms RTT; 6 Mbps downstream/ 1.39 Mbps upstream again on http://speedtester2.aa.net.uk

I note that that is far better speed than the miserable 2Mbps downstream that I get with my 3G USB ‘dongle’ NIC.

Btw is ‘latency’ round-trip time or one-way ?  It might be difficult to know the times of the outward versus return legs, I suppose. But I take it that no one is doubling or halving any numbers for presentation.

[burakkucat]

I have always assumed that latency, if not otherwise stated, is the round trip value.

[Weaver]

I couldn’t work out why the 200ms figure was so rubbish, in the window. Perhaps the iPad had made a bad choice; struggling to do 4G but badly when it could have alternatively done 3G but well. Is there even such a thing ? I don’t know how the criteria work nor how such a system makes the decision, whether on performance or reliability or certain hard stat-thresholds associated with the two protocol sets for 3G vs LTE.

Anyway, if such a thing does exist, then I thought that perhaps it was using 4G that was unreliable when it would have done better to give up, drop to 3G if that is more reliable and get fewer Ln re-txs and is more speed, since I would assume all the possible retx operations would be hurting the rtt. Is that possible? Lower basic rate but you get the data there first try.

[Alex Atkin UK]

AFAIK if a device can see 4G it will ALWAYS pick 4G, the auto switching is only so that if you lose a 4G signal it will switch to the next newest technology it can find.

I manually forced my phone to 3G recently and also found it got better performance for this reason, as far fewer people are using it.

Also like with DSL, there are different versions which is potentially a reason the USB stick is slower (though it just having worse reception is also a high possibility), its probably using an older version of 3G.

[Weaver]

It may be that this is a lightly loaded cell due to the low population. The poor signal (two bars out of five this second) is due to my suitable position in bed with a six foot thick stone double gable-end wall casting a shadow; I would need to move sideways to get into the line through the bedroom window towards the basestation.

[re0]

As Alex said, a device will always try to use 4G, and only 4G, when it is set to 4G as opposed to auto. It would be rather counter-intuitive for it to do otherwise.

I do not know how good the iPad's modem is, nor which LTE Band it is using. Which iPad Gen. is it even?

I have personally found latency to actually be worse and deviate more on 4G LTE as opposed to on 3G, and my results can be found here. Of course, the difference is small in comparison to what you are seeing so perhaps there is another reason - like what you mentioned with retransmission impact.

Perhaps you know of some network diagnostic tools you can use on the iPad?

[Weaver]

It’s an iPad Pro v1 which would make it hw V7 overall I suppose.