|
|
|
Hi, First post here.
Recently i signed up with three unlimited. Great deal with great speeds and price. BUT...
The big but is that the connection was intermittent. Decent enough connection/signal from local mast but pages were sometime as though there were no conection.. i forget the error code.
I thought i would dig further into this starting using standard ping and tracert and then with some tools..
So far ive been using ping plotter on laptop and traceping on android phone.
What im seeing is a rather large loss when the device is teathered compared to not.
The servers where which lose the most packets are:
188.31.255.142
188.31.255.181
With the worst being: 188.31.255.186
The amount of lost varies but when teathered reaches 40% where as a test at the same time on the phone usually is less at around 5-12%. I say usually as i just this instance tested and found the loss was between 22 and 45% on ***.***.***.181 and loss now is at 14-17% (depending on server) after 740 pings..
Can anybody confirm if they have issues also?
Note generally im pinging amazon, bbc or google.co.uk
|
|
|
|
what are you doing just pinging their IP's?
What are you actually trying to check?
I'd be checking actual sites I use rather than some of their mobile broadband IP's.
|
|
|
what are you doing just pinging their IP's?
What are you actually trying to check?
I'd be checking actual sites I use rather than some of their mobile broadband IP's.
Thanks Chris. As im sure you know connection is not straight a to b as it would be if it were a vpn to your office.. it is routed through various points or across various servers.. when you tracert you get the IPs or at least some of the IP's for the servers along the way.. What these apps do is combine ping and tracert. and instead of just pinging just the destination or as you put it "the actual sites i use" it also pings all thoese in between.
If you give the apps mentioned a try you will see. You will also see that 'generally' a high percentage of the route remains the same whichever site you wish to check. Its a good way to find out where the contention/ faults lie. Give it a try.. Some bugs in the android app but it works well.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
Some routes might simply not reply. I can run the same tests on my machines to various sites and servers and get differing results throughout the day.
I'd be more concerned about the actual connection to the site/service I'm contacting rather than the hops replying inbetween. If your rtt stays low then I wouldn't be too concerned. If you get sudden random spikes then I'd pay more attention.
To add to your VPN comment a VPN connection will still need a number of hops in the route before it hits your company vpn endpoint. It'll still have to traverse the internet along standard routes, you just wont be able to easily check each hop.
For me to get onto my work VPN it takes 11 hops before it hits one of the VPN endpoints then everything else travels there.
Extra edit, are your pings successful to the endpoints or are they too missing replies?
Edited by chriscdotcodotuk (Tue 28-May-19 17:01:41)
|
|
|
|
.
im sure you know what 40% packet loss can do to a connection and it sucks. if it hadn't why on earth would i be spending time looking at it.
anyway as said the purpose is to gather data from other users not get support or opinions.
i'd be grateful if you share your results.
|
|
|
im sure you know what 40% packet loss can do to a connection and it sucks. if it hadn't why on earth would i be spending time looking at it.
Packet loss only counts to the very end. If a device (router) in the middle of the route is ignoring pings because it is handling traffic well, that is not a problem.
Many people misread ping plotting tools.
plusnet 80/20 (2/jun/14) at 470m - Sync history highest: 64/9 (Sep/17), 54/6 (Jan/19), 51/6 (Mar/19)
20 years of broadband from 1999's ntl:cable modem trial - Live BQM
|
|
|
Testing with the same app I get really poor results for the sites you listed from my 4G connection (not tethered) and from my Virgin connection.
Funnily enough when I run network tools on any of my machines at home I dont get the same losses.
I personally do not trust the app.
Traceroute is fine for knowing a route at that time.
Ping to the actual service is better.
Tethered to my laptop just now and google was 38/40 successful pings. Lowest was 37ms, highest was 62ms.
Edited by chriscdotcodotuk (Tue 28-May-19 18:43:57)
|
|
|
Packet loss only counts to the very end. If a device (router) in the middle of the route is ignoring pings because it is handling traffic well, that is not a problem.
Many people misread ping plotting tools.
of course some servers wont accept icmp echo request packets. but that will be rather self evident as there will be 100% loss on that server or more to the point it will just show as asterixis with no ip revealed.. however the packets will be forwarded to the next hop.
if you're on three could you share your results
Edited by deleted (Tue 28-May-19 19:46:33)
|
|
|
Testing with the same app I get really poor results for the sites you listed from my 4G connection (not tethered) and from my Virgin connection.
Tethered to my laptop just now and google was 38/40 successful pings. Lowest was 37ms, highest was 62ms.
thanks for the results. if ping destination is best.. and you know 2 packets were lost. which isnt bad at all. without a similar app, script or manually going through.. how will you know the most likely point that is causing this loss?
when you said the sites mentioned... did you refer to google and bbc or the hutchinson in ips
|
|
|
|
I tried google.co.uk / bbc.co.uk and amazon.co.uk
All three gave me poor results using my phones 4g connection AND my virgin connection.
I tested google.co.uk on my laptop and got 199/200 successful on the virgin connection ranging from 11-22ms and then when tethered 38/40. the times were longer and I got impatient.
I would hazard a guess it'd be more reliable to ping individual hops one at a time rather than trying to ping them all at the same time.
40% packet loss would be terrible and sure you'd want to complain about it but if you spoke to Three and told them that you was seeing dropped ICMP Packets on a congested line they'd A) not understand a word you was saying or if they was more technical B) they'd tell you that many routers give ICMP Packets a low priority and will drop them in preference of higher priority packets. That is standard network behavior so to assume you've got a problem because a few of the hops are dropping your packets regularly seems a bit wrong.
|
|
|
I tried google.co.uk / bbc.co.uk and amazon.co.uk
All three gave me poor results using my phones 4g connection AND my virgin connection.
I tested google.co.uk on my laptop and got 199/200 successful on the virgin connection ranging from 11-22ms and then when tethered 38/40. the times were longer and I got impatient.
Thanks for that.. 5% loss teathered.. interesting..
40% packet loss would be terrible
IS terrible
many routers give ICMP Packets a low priority and will drop them in preference of higher priority packets.
Given traffic shaping method which vary depending upon setup... . I would expect higher latency not outright dropped packets but i will take it on board as a possibility and will try some other packet types..
That is standard network behavior so to assume you've got a problem because a few of the hops are dropping your packets regularly seems a bit wrong.
.
Not sure what point you are trying to make here or even why you would wish to even imply the same ..
1) 40% is far from being a few.
2) The only assumption i read is your own and it is that i base the poor connectivity upon the packet loss.
As mentioned twice previously.. making this a third time.... It is the poor performance (examples previously given) that made me look further into it.
Also. The very fact that there is a discrepancy between tethered and untethered raises suspicions on how traffic is shaped. Given your posts, Im sure you are aware there was an offcom investigation into the potentially illegal traffic shaping Hutchinson and Vodaphone are suspected of . and although the investigation is closed it is not conclusive and the case is its not a closed book.
Anyway.. its unfortunate that there is so much chat and lack of data in this post. I guess i'll have to start a new one hopefully people will respond with data
I do appreciate the point regards the traffic shaping and will dig a little further into that shortly.
|
|
|
From my fixed line (FTTP) connection:
| Text | 1
23
45
67
89
1011
1213
1415
1617
1819
2021
2223
2425
2627
2829
3031
32 | C:\>tracert 188.31.255.186
Tracing route to 188.31.255.186.threembb.co.uk [188.31.255.186]
over a maximum of 30 hops:
1 1 ms 1 ms 1 ms Linksys01421 [192.168.1.1] 2 19 ms 18 ms 18 ms SOV-D-LNS-ASR-ED.fluidata.co.uk [77.89.165.5]
3 18 ms 18 ms 18 ms 77-89-165-33.fluidata.co.uk [77.89.165.33] 4 19 ms 19 ms 19 ms 77-89-177-17.fluidata.co.uk [77.89.177.17]
5 20 ms 20 ms 20 ms 78.152.52.113 6 21 ms 20 ms 20 ms ae5-100.cr11-lon1.ip4.gtt.net [46.33.94.5]
7 29 ms 20 ms 20 ms xe-9-1-6.cr0-lon1.ip4.gtt.net [89.149.137.117] 8 21 ms 21 ms 21 ms ip4.gtt.net [46.33.78.6]
9 23 ms 22 ms 21 ms 188.31.255.169.threembb.co.uk [188.31.255.169] 10 24 ms 22 ms 22 ms 188.31.255.158.threembb.co.uk [188.31.255.158]
11 23 ms 22 ms 22 ms 188.31.255.125.threembb.co.uk [188.31.255.125] 12 21 ms 21 ms 21 ms 188.31.255.186.threembb.co.uk [188.31.255.186]
Trace complete.
C:\>ping 188.31.255.186
Pinging 188.31.255.186 with 32 bytes of data:
Reply from 188.31.255.186: bytes=32 time=22ms TTL=247Reply from 188.31.255.186: bytes=32 time=21ms TTL=247
Reply from 188.31.255.186: bytes=32 time=21ms TTL=247Reply from 188.31.255.186: bytes=32 time=23ms TTL=247
Ping statistics for 188.31.255.186:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),Approximate round trip times in milli-seconds:
Minimum = 21ms, Maximum = 23ms, Average = 21ms |
From a Three sim (unlimited everything plan) tethered via an iPhone 8:
| Text | 1
23
45
67
89
1011
1213
1415
1617
1819
2021
2223
2425
2627
2829
| C:\>tracert 188.31.255.186
Tracing route to 188.31.255.186.threembb.co.uk [188.31.255.186]
over a maximum of 30 hops:
1 5 ms 2 ms 2 ms 172.20.10.1 2 * * * Request timed out.
3 63 ms 45 ms 58 ms 172.23.128.221 4 59 ms 45 ms 62 ms 172.23.162.5
5 62 ms 49 ms 42 ms 172.23.175.3 6 * * * Request timed out.
7 74 ms 46 ms 61 ms 188.31.255.142.threembb.co.uk [188.31.255.142] 8 75 ms 49 ms 58 ms 188.31.255.181.threembb.co.uk [188.31.255.181]
9 64 ms 54 ms 59 ms 188.31.255.186.threembb.co.uk [188.31.255.186]
Trace complete.
C:\>ping 188.31.255.186
Pinging 188.31.255.186 with 32 bytes of data:Reply from 188.31.255.186: bytes=32 time=235ms TTL=248
Reply from 188.31.255.186: bytes=32 time=53ms TTL=248Reply from 188.31.255.186: bytes=32 time=49ms TTL=248
Reply from 188.31.255.186: bytes=32 time=56ms TTL=248
Ping statistics for 188.31.255.186: Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds: Minimum = 49ms, Maximum = 235ms, Average = 98ms |
Edited by deleted (Tue 28-May-19 23:57:15)
|
|
|
What phone are you trying to tether from? Android, IOS?
Forgive me. I forgot you mentioned it in your original post.
I used to use an app named "WI-FI tether router" It allowed me to tether unlimited data using my Three sim (when maximum tether allowance was 30gb) It spoofs the network into thinking your using your mobile. There is a catch and requires your android device to be rooted.
If you are into that sort of stuff, then it's worth a shot to see if Three are doing something on their end with tethering/mobile hotspot. Doubtful. But this will allow you to test that theory.
Regards
Edited by deleted (Wed 29-May-19 00:39:59)
|
|
|
On the intermediate hops it is not always a 100% ignore, since the ignore or delay in reply long enough to be seen as ignored or lost is dependent on what other activity the intermediate device is doing.
So if hop 5 (for example) is showing 40% packet loss but subsequent hops are showing no loss then there is no problem.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
Pages in this thread: 
Print Thread
