Sponsored

Latency with IDNet

BigAlbert

Regular Member
I'm after some idea of latency with IDNet mainly for gaming purposes could someone using IDNet post the results from the below commands ?

tracert 185.60.112.157
tracert 149.28.112.244
tracert 155.133.247.155
 

baby_frogmella

Regular Member
The results will be pretty much meaningless - unless someone connected to the same exchange as you provides the tracert results.
 
Last edited:

BigAlbert

Regular Member
baby_frogmella no they won't be meaningless, I can subtract the time of the first hops until the exit from the core network to get a time for exit from the core network to the destination IP address i.e. how long the peering arrangement is taking. That will translate to the same FTTC / exchange latency I have, which I already know plus the peering time. I then do a simple comparison with a trace route with my current provider i.e. same calculation of removing the core network hops. Of course it could turn out that my core network is faster therefore making up for a slower peering setup - a trace route will tell me all this / what I need to know...
 

vcommsmind

Pro Member
1 <redacted> (192.168.100.1) 0.848 ms 0.587 ms 0.613 ms
2 <redacted> (192.168.100.5) 0.669 ms 0.607 ms 0.592 ms
3 <redacted> (<redacted>) 8.279 ms 8.679 ms 8.399 ms
4 81.25.207.41 (81.25.207.41) 8.918 ms 8.716 ms 9.026 ms
5 ae-3.r24.londen12.uk.bb.gin.ntt.net (129.250.4.23) 8.811 ms 8.918 ms 9.919 ms
6 ae-7.r20.nwrknj03.us.bb.gin.ntt.net (129.250.6.147) 77.616 ms * *
7 * * *
8 ae-1.r20.chcgil09.us.bb.gin.ntt.net (129.250.2.166) 93.520 ms 118.933 ms 95.751 ms
9 ae-12.r08.chcgil09.us.bb.gin.ntt.net (129.250.2.190) 94.011 ms 92.538 ms 92.528 ms
10 ae-1.a01.chcgil09.us.bb.gin.ntt.net (129.250.5.94) 93.171 ms
ae-0.a01.chcgil09.us.bb.gin.ntt.net (129.250.4.218) 122.299 ms
ae-31.a00.chcgil09.us.bb.gin.ntt.net (129.250.3.159) 114.872 ms
11 * * *
12 * * *
13 * * *
14 149.28.112.244.vultr.com (149.28.112.244) 96.394 ms !Z 94.605 ms !Z 97.169 ms !Z


1 <redacted> (192.168.100.1) 0.848 ms 0.587 ms 0.613 ms
2 <redacted> (192.168.100.5) 0.669 ms 0.607 ms 0.592 ms
3 <redacted> (<redacted>) 8.279 ms 8.679 ms 8.399 ms
4 81.25.207.41 (81.25.207.41) 8.918 ms 8.716 ms 9.026 ms
5 ae-3.r24.londen12.uk.bb.gin.ntt.net (129.250.4.23) 8.811 ms 8.918 ms 9.919 ms
6 ae-7.r20.nwrknj03.us.bb.gin.ntt.net (129.250.6.147) 77.616 ms * *
7 * * *
8 ae-1.r20.chcgil09.us.bb.gin.ntt.net (129.250.2.166) 93.520 ms 118.933 ms 95.751 ms
9 ae-12.r08.chcgil09.us.bb.gin.ntt.net (129.250.2.190) 94.011 ms 92.538 ms 92.528 ms
10 ae-1.a01.chcgil09.us.bb.gin.ntt.net (129.250.5.94) 93.171 ms
ae-0.a01.chcgil09.us.bb.gin.ntt.net (129.250.4.218) 122.299 ms
ae-31.a00.chcgil09.us.bb.gin.ntt.net (129.250.3.159) 114.872 ms
11 * * *
12 * * *
13 * * *
14 149.28.112.244.vultr.com (149.28.112.244) 96.394 ms !Z 94.605 ms !Z 97.169 ms !Z
 

vcommsmind

Pro Member
But ping times would be more useful for your stated goal...

PING 185.60.112.157 (185.60.112.157): 56 data bytes
64 bytes from 185.60.112.157: icmp_seq=0 ttl=248 time=16.594 ms
64 bytes from 185.60.112.157: icmp_seq=1 ttl=248 time=16.696 ms
64 bytes from 185.60.112.157: icmp_seq=2 ttl=248 time=15.135 ms
64 bytes from 185.60.112.157: icmp_seq=3 ttl=248 time=15.145 ms
64 bytes from 185.60.112.157: icmp_seq=4 ttl=248 time=15.354 ms
64 bytes from 185.60.112.157: icmp_seq=5 ttl=248 time=15.421 ms
64 bytes from 185.60.112.157: icmp_seq=6 ttl=248 time=15.447 ms
64 bytes from 185.60.112.157: icmp_seq=7 ttl=248 time=15.324 ms
64 bytes from 185.60.112.157: icmp_seq=8 ttl=248 time=15.672 ms

PING 149.28.112.244 (149.28.112.244): 56 data bytes
64 bytes from 149.28.112.244: icmp_seq=0 ttl=55 time=94.681 ms
64 bytes from 149.28.112.244: icmp_seq=1 ttl=55 time=94.434 ms
64 bytes from 149.28.112.244: icmp_seq=2 ttl=55 time=94.589 ms
64 bytes from 149.28.112.244: icmp_seq=3 ttl=55 time=94.247 ms
64 bytes from 149.28.112.244: icmp_seq=4 ttl=55 time=94.474 ms
64 bytes from 149.28.112.244: icmp_seq=5 ttl=55 time=94.439 ms
64 bytes from 149.28.112.244: icmp_seq=6 ttl=55 time=94.316 ms
64 bytes from 149.28.112.244: icmp_seq=7 ttl=55 time=94.459 ms
64 bytes from 149.28.112.244: icmp_seq=8 ttl=55 time=94.246 ms

PING 155.133.247.155 (155.133.247.155): 56 data bytes
64 bytes from 155.133.247.155: icmp_seq=0 ttl=56 time=32.241 ms
64 bytes from 155.133.247.155: icmp_seq=1 ttl=56 time=32.044 ms
64 bytes from 155.133.247.155: icmp_seq=2 ttl=56 time=50.813 ms
64 bytes from 155.133.247.155: icmp_seq=3 ttl=56 time=32.133 ms
64 bytes from 155.133.247.155: icmp_seq=4 ttl=56 time=32.050 ms
64 bytes from 155.133.247.155: icmp_seq=5 ttl=56 time=31.987 ms
64 bytes from 155.133.247.155: icmp_seq=6 ttl=56 time=31.999 ms
64 bytes from 155.133.247.155: icmp_seq=7 ttl=56 time=32.163 ms
64 bytes from 155.133.247.155: icmp_seq=8 ttl=56 time=32.399 ms
 

BigAlbert

Regular Member
vcommsmind thanks so much that's really useful, you must be close to London ?

My goal is this, I have 14ms into my current ISP, this is due to as far as I'm aware the hand over from BT wholesale to the ISP and the L2PT tunnel used.

I'm in the midlands so distance is a factor as well but when I was with BT I had 7ms to their core network, so I've had a 7 to 8 ms increase.

What I was hoping for is if IDNet have fast peering and worst case I have the same time into the IDNet core network of lets say 14ms the fast peering will offset this time.

I think though it's becoming clear that I'm not seeing any ISP with a noticeable peering advantage, you have roughly 8ms to reach IDNet and your pings are about 6 to 7 ms faster than mine at the moment, so adding on the 6ms gets me to the same ping results I currently have.

In other words as I was hoping to see faster peering that doesn't seem to be the case.

I have no ISP to switch to that will give me a faster time to the core network and that is fit for gaming, sad times.
 

vcommsmind

Pro Member
vcommsmind thanks so much that's really useful, you must be close to London ?

My goal is this, I have 14ms into my current ISP, this is due to as far as I'm aware the hand over from BT wholesale to the ISP and the L2PT tunnel used.

I'm in the midlands so distance is a factor as well but when I was with BT I had 7ms to their core network, so I've had a 7 to 8 ms increase.

What I was hoping for is if IDNet have fast peering and worst case I have the same time into the IDNet core network of lets say 14ms the fast peering will offset this time.

I think though it's becoming clear that I'm not seeing any ISP with a noticeable peering advantage, you have roughly 8ms to reach IDNet and your pings are about 6 to 7 ms faster than mine at the moment, so adding on the 6ms gets me to the same ping results I currently have.

In other words as I was hoping to see faster peering that doesn't seem to be the case.

I have no ISP to switch to that will give me a faster time to the core network and that is fit for gaming, sad times.
Well given some of your examples they would be transit, not peering, and there aren't that many choices of transit providers for a given long distance destination so no... physics mostly dictates the time.
 

BigAlbert

Regular Member
So vcommsmind can you confirm you're close ish to London ?

I'm still hanging onto some most likely false hope I have an option of lowering my ping other than switching back to mr BT bouncy ping land.
 

ManOfMeans

ULTIMATE Member
Ping (ICMP) is almost marked as the lowest priority as well so it isn't the definitive answer to any question.
 

vcommsmind

Pro Member
Ping (ICMP) is almost marked as the lowest priority as well so it isn't the definitive answer to any question.
Yes but a series of pings would give an indication of typical performance, as although lower priority, it shouldn't be that low if the router isn't oversubscribed.

A single trace route doesn't tell you much about typical latency, just a one off snapshot of one event.
 

BigAlbert

Regular Member
vcommsmind can you give me some idea where you are then location wise ? seeing as you've got a reasonable ping into IDNet core network, which is the problem I'm trying to solve. Are you by any chance on a LLU connection such as Talk Talk business ?
 

Captain_Cretin

ULTIMATE Member
The only real way of lowering your ping is to change how your PC sends packets.
By default, it saves up packets to send in bursts, change it to send as soon as the packet is ready to go knocks 10-20mS off.

Leatrix Latency Fix - Leatrix.com
 

BigAlbert

Regular Member
Captain_Cretin I've tried about every tune and tweak to the Windows network stack and network card I can find and I have improved network responsiveness by some amount. I'd be surprised though if you can shave 10ms off a ping, it's one packet, I can ping my router in less than 1ms so I don't think I'm seeing any problem with my PC sending packets.

I'm still really interested in where vcommsmind is located - how far away from London ?
 

Captain_Cretin

ULTIMATE Member
I am talking real packets in a real online game over the real internet, before instigating the previously mentioned little hack, my average game ping to the servers in the USA was 100-120mS; after the hack 85-95mS.
I got the hack from the WOW forums originally, and I realise the size of the reduction depends on how many packets the PC and game server are swapping; the game I play is probably the LOWEST bandwidth game still going; a typical up speed is in the 1kbit range, and down speed rarely goes into two digits.
An hours play uses about 10MB.
 

BigAlbert

Regular Member
Captain_Cretin I've used 'Leatrix Latency Fix' a while ago and I don't think it does anything that special I think it's just the TCP ACK delay change, perhaps a few other tweaks ? I'm sure it might help if your game uses TCP but most games use UDP.

There's lots of tweaks for game networking and I've most likely tried them all, I'd certainly start with your network card settings and change your interrupt moderation to disabled, that makes a big difference, perhaps not with ping but with game responsiveness. In fact a lot of NIC offloading / default NIC settings need changing for gaming such as LSO etc. Indeed many tweaks won't as you say have a major effect on ping but will have an effect how fast you push packets out or in or how fast they get processed by the game. Have you tried TCP optimizer ? that's another 'goto' starting point with tweaking.

Also major places to tweak in the registry are below, certainly AFD (Ancillary Function Driver) is a place where buffering happens and I've made significant changes here.
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\AFD\Parameters

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters

As always though tweaking the registry values can have a detrimental effect if you're not careful.

Setting QoS / tagging packets in and out with a DSCP value of 46 can improve things as well, in effect making your packets look like a VOIP call i.e. Windows group policy editor and router changes for QoS.

Other practical things such as disabling Windows 10 updates, disabling not needed services like all the xbox junk, windows search etc. try 'windows 10 shutup', another good tool for helping, The whole tweaking scene is massive with lots to consider and try, even your anti Virus provider and settings can have a detrimental effect on gaming, I could probably write a book on tweaking :)
 

BigAlbert

Regular Member
Here's another one I've tried, there's a free version and a paid one, I'd probably buy it if it wasn't so expensive, maybe when I've got some free cash I will do


I don't currently have it installed but at the time I could perhaps detect a minor improvement with gaming - well online gaming that is.
 

BigAlbert

Regular Member
oh and another interesting option is process lasso from https://bitsum.com/

This has process optimization and IO priority although I've recently uninstalled it, plus all the other gaming optimizers such as Razer Cortex, anyhow I'll stop rambling on tweaking, I could go on and on from routers to high refresh rate monitors to gaming mice and so on :)

The big tweak I'm after is the ISP, the 'right' ISP will have a big impact, in fact I could feel a major improvement just switching away from BT and that's with a slightly higher ping. I could well be with the best ISP for gaming but I just need my ping to be a little lower, time will tell...
 
Top
Promotion
Cheapest Superfast ISPs
  • Hyperoptic £19.95 (*22.00)
    Avg. Speed 50Mbps, Unlimited
    Gift: Promo Code: FEB2020
  • Post Office £20.90 (*37.00)
    Avg. Speed 38Mbps, Unlimited
    Gift: None
  • Vodafone £21.95
    Avg. Speed 63Mbps, Unlimited
    Gift: None
  • SSE £22.00
    Avg. Speed 35Mbps, Unlimited (FUP)
    Gift: None
  • xln telecom £22.74 (*47.94)
    Avg. Speed 66Mbps, Unlimited (FUP)
    Gift: None
Prices inc. Line Rental | View All
Helpful ISP Guides and Tips
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
Promotion
The Top 20 Category Tags
  1. BT (2616)
  2. FTTP (2399)
  3. FTTC (1711)
  4. Building Digital UK (1649)
  5. Politics (1505)
  6. Openreach (1485)
  7. Business (1302)
  8. FTTH (1190)
  9. Statistics (1140)
  10. Mobile Broadband (1102)
  11. Fibre Optic (1005)
  12. Ofcom Regulation (959)
  13. 4G (953)
  14. Wireless Internet (951)
  15. Virgin Media (907)
  16. EE (630)
  17. Sky Broadband (623)
  18. TalkTalk (607)
  19. Vodafone (573)
  20. 3G (431)
Sponsored

Copyright © 1999 to Present - ISPreview.co.uk - All Rights Reserved - Terms  ,  Privacy and Cookie Policy  ,  Links  ,  Website Rules