Originally Posted by Vandyyy
Forgive me for the small sample size, but I'd gladly appreciate some more data. I have access to 3 Galaxy Nexus models and the one Moto X. By the end of the night I should have an iP5 sample.
My suspicion is telling me something is off on my LTE ping. The 3 GNexes all ping 50-80ms under identical network conditions on my home tower and the Moto X is seldom under 100 (same SIM, same APN, same spot in the house, all within an hour). Obviously you all will have different results since you're testing different towers on different servers, etc. Band 4 with a TMo SIM was pinging pretty low under different network conditions, but this is my only band 4 capable device so I have nothing else to test against.
The lower latency isn't a deal-breaker by any means, but if there's something I can do to fix it, I'd like to. But before beginning I wanted to see if it's more symptomatic of hardware difference in antennas. Since I only have two different makes/models so far, I was hoping one of you lovely folk would be bored enough to test. The tower I've been testing from is band 13, not 4. The Moto X is consistently reporting stronger signal across the board, which is hardly surprising given the difference in hardware age/obsolescence. The ping, however, worries me.
So, any input?
A 50-30ms ping discrepancy is nothing to worry about and well within tolerance for even multiplayer gaming. I don't know if the Moto X is slower or not but if it IS slower than it isn't slower-enough to make a difference in the user experience. You have to be pinging a web site, there is also network congestion to worry about not to mention variances in the load on the tower as just atmospheric conditions constantly changing.
In short... If you're talking about a max variance of 50ms (just for comparison that is about a tenth of the time it takes to blink an eye) it is impossible to say what exactly the cause is. If you really did want to investigate you'd want to ping your devices default gateway, so the first hop for an IP packet as it leaves for the Internet. If you COULD do that then you'd need a LOT of data, so I'd say you'd do multiple samplings from each device over a period of time. If you've got access to the ping command on the device then you could use Android Terminal and instruct ping to do this, I don't remember the exact command right now, but you'd want to ping 3 different times of day, on three different days of the week. Each time you ping, I would do 50 pings.
So each device would end up with 450 data points (50 pings at Time 1, Time 2, and Time 3 for Day 1, Day 2, and Day 3). Then you could reasonably compare them and variances might be visible between the devices. Also, we should look at one other thing, the time span you are talking about we have to take into account the speed of light because in 50ms light (or radio waves) can travel upwards of 9000 miles. This doesn't seem like you'd need to worry about it but you do, because your phone first encodes the ping in a packet for ICMP (an IP type packet). That IP packet then has to get packaged and encoded further to transmit over the LTE radio network. It gets sent to the tower, an unknown distance away. Then it gets decoded and put back into IP form (from LTE signal form) and transmitted along whatever uplink the tower has to its destination.
I suggested doing testing using your default gateway because every single step in the process adds time and adds a layer of uncertainty. If it is going to the default gateway that isn't necessarily the tower, or at the tower. If it is, great. But you don't know where it is physically located, so in 50ms your packet has to travel a lot of distance to get where it is going and then get back to you. Round trip time of 50ms means it probably takes 25ms to get where it is going and 25ms to get back - which is about 3000 miles at the speed of light, so from the middle of the US to the coast (roughly) in perfect conditions, which there never are.
I hope you see my point, though I must admit to enjoying looking up some of the details here. 50ms is not latency you're ever going to notice other than the number on your speed test. However, if you want to test it, you can and without much help either since you have 2 devices to compare. Just run as I suggested above pings to your default gateway (also note if the default gateway is the same on each device) over the cellular network. Compile your data and see what you can see, if you come back with results that show the Moto X is consistently slower than your Nexus device, I would personally love to know about it.