|
|
|
I was looking at making a house move in the coming year and was just curious how far would I need to live from Docklands (or would it be Slough) to get 1ms or lower latency on London gaming servers for Steam / BattleNet games and or Google servers.
I know it might sound silly but I was just curious as to how these things work and how far out one can go from these main internet exchanges before going past the 1ms mark.
Thanks
|
|
|
If you've a leased line with a really direct route Reading at least is fine.
If you've Virgin Media cable you could be next door to their equipment in Poplar and you're still not getting 1 ms. All about how you're getting onto the provider's network and how direct it is. Hyperoptic are very good with their FTTB, most full fibre you'd need to be pretty close as just to get to the 'exchange' is a little over a millisecond round trip: not much time to get to the server left before you're above 2.
If you're inside the M25 and on the right provider you're probably going to make it as long as the app rounds down, not to the actual nearest number.
----------
Exceptionalism diminishes, cooperation enhances.
Edited by XGS_Is_On (Sun 30-Apr-23 14:36:28)
|
|
|
|
Yeah I reckon you’ll push to get under 1ms. More than likely 2ms. Not so much distance in London but network configurations, router hops and peering config that will drive the delay.
Are you really going to notice the difference between 1ms and 3ms in the real world? You’re gaming not high frequency trading at the end of the day.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
Yeah I reckon you’ll push to get under 1ms. More than likely 2ms. Not so much distance in London but network configurations, router hops and peering config that will drive the delay.
Very true and part of why I mentioned Hyperoptic's FTTB - it's not a shared network at the access layer so no need to wait.
Even over XGSPON the first hop is over a millisecond for obvious reasons - takes a little while to get a timeslot to send. This is to a router next to the OLT that serves me, less than 2 km away straight line and about 3 by road and likely fibre route.
Packets: sent=4, rcvd=4, error=0, lost=0 (0.0% loss) in 1.509285 sec
RTTs in ms: min/avg/max/dev: 1.071 / 1.191 / 1.327 / 0.096
Bandwidth in kbytes/sec: sent=0.159, rcvd=0.159
----------
Exceptionalism diminishes, cooperation enhances.
Edited by XGS_Is_On (Sun 30-Apr-23 14:46:19)
|
|
|
before going past the 1ms mark.
Thanks
Whilst having the smallest latency as possible helps [alot in games] and you can reduce that by mouse/ram/gfx/keyboard/internet/monitor. Sometimes the server is [censored] or the game could be badly optimised.
Or you have a bad game plan. but the worst, is when you are unwell.
The above is about the fact sometimes its a bit of everything and don't chase the poster graphics of 1ms
|
|
|
I stand corrected. With a voice call and a game running it drops notably as the voice and game are taking care of reserving the bandwidth.
From 45.92.47.33: bytes=60 seq=0001 TTL=254 ID=d239 time=1.005ms
From 45.92.47.33: bytes=60 seq=0002 TTL=254 ID=d23a time=1.001ms
From 45.92.47.33: bytes=60 seq=0003 TTL=254 ID=d23b time=0.814ms
From 45.92.47.33: bytes=60 seq=0004 TTL=254 ID=d23c time=0.836ms
Packets: sent=4, rcvd=4, error=0, lost=0 (0.0% loss) in 1.507531 sec
RTTs in ms: min/avg/max/dev: 0.814 / 0.914 / 1.005 / 0.089
----------
Exceptionalism diminishes, cooperation enhances.
|
|
|
Whilst having the smallest latency as possible helps [alot in games] and you can reduce that by mouse/ram/gfx/keyboard/internet/monitor. Sometimes the server is [censored] or the game could be badly optimised.
Should we also be starting to factor in that a Human Eye blinks at 0.1-0.4 sec each time and we blink around 20 times a min....so maybe us mere humans by design are also part of latency.
Martin Pitt
Company Founder
Aquiss Limited
https://www.aquiss.net
FTTC, FTTP, GEA, EFM, Leased Lines, Telecoms and Hosting
|
|
The above post has been made by an ISP REPRESENTATIVE (although not necessarily the ISP being discussed in the post).
|
|
|
|
There really are more important things in life and certainly better places to live.
|
|
|
Whilst having the smallest latency as possible helps [alot in games] and you can reduce that by mouse/ram/gfx/keyboard/internet/monitor. Sometimes the server is [censored] or the game could be badly optimised.
Should we also be starting to factor in that a Human Eye blinks at 0.1-0.4 sec each time and we blink around 20 times a min....so maybe us mere humans by design are also part of latency.
i think we should or do so. Tbf, i was annoyed that it was about the 1ms thing, and whilst the OP is valid one it it was like the headline thing you hear. And in my head i was like "well theres other issues here too".
|
|
|
There really are more important things in life and certainly better places to live.
Lifestyle choice eh. I understand the OP was in Knutsford based on previous threads. Seems a pretty nice place overall, but who knows.
|
|
|
Lifestyle choice eh. I understand the OP was in Knutsford based on previous threads. Seems a pretty nice place overall, but who knows. Not sure anyone would want to swap Knutsford with Manchester Road on the Isle of Dogs to get 1ms pings.
|
|
|
Lifestyle choice eh. I understand the OP was in Knutsford based on previous threads. Seems a pretty nice place overall, but who knows. Not sure anyone would want to swap Knutsford with Manchester Road on the Isle of Dogs to get 1ms pings.
I'm with you on that one. But then again I can't think that most people in Knutsford would worry overly much about sub 1ms pings.
|
|
|
Lifestyle choice eh. I understand the OP was in Knutsford based on previous threads. Seems a pretty nice place overall, but who knows. Not sure anyone would want to swap Knutsford with Manchester Road on the Isle of Dogs to get 1ms pings.
…. what are the City Limits like round there ??
|
|
|
|
I hear it is a 25mph speed limit and they don't allow motorbikes.
|
|
|
|
I ping 3ms to the first hop (Lit Fibre XGS-PON), I presume the first hop is in a datacentre somewhere - traceroute would suggest it's in Telehouse. I am 20 miles in a straight line from that location.
|
|
|
|
|
|
|
Don't forget the store on Friday.
|
|
|
Don't forget the store on Friday.
 We could all go on a road trip to visit
|
|
|
Oh ofcourse that's why I prefaced by saying its a silly quesiton  .
I was just curious to know what the limits are to latency in the London area. I'm moving for a project for 6-12 months in Canary Wharf so that's why I was curious as to how far I could be and still have the 1ms latency that I see a lot of people posting on their community fibre / hyperoptic speed tests.
|
|
|
|
Hyperoptic depending on the specific London location could deliver some impressive ping times as their traditional FTTB network variously used direct EAD/leased circuits which could be faster than other networks that have a local aggregation / PON and then a backhaul to Docklands/Slough etc.
Although in reality there wouldn’t be that much in it, a ms perhaps. Community Fibre are a great provider and solid network.
Depends really on where you are intending to move to in London and looking at what’s available. There’s still plenty of places that only have a Virgin HFC cabled monstrosity as a realistic optic as they are DSL only or not yet served by any of the AltNets.
So that’s what I’d be checking first and foremost and “distance” later. You’ll be fine in zones 1-4 👍
|
|
|
You’ll be fine in zones 1-4 👍 They are train zones for the less well travelled
|
|
|
|
In east london I could get pings of 0.3ms to various London datacenters on hyperoptic!
|
|
|
Hyperoptic will likely be the lowest latency option available in London short of a leased line. I had a connection from them in Manchester several years ago and would frequently see game servers reporting my ping as 1ms.
Community Fibre are solid but the XGS-PON hop definitely bumps the latency up a little compared to Hyperoptic. It's superior to Openreach FTTP both for upload speeds and latency though. Some pings below from an address in Zone 3, with Plusnet FTTP for comparison at the same location.
CF 1.1.1.1
sent=5 received=5 packet-loss=0% min-rtt=2ms316us avg-rtt=2ms415us
max-rtt=2ms485us
CF 8.8.8.8
sent=5 received=5 packet-loss=0% min-rtt=1ms740us avg-rtt=1ms828us
max-rtt=1ms884us
CF 9.9.9.9
sent=5 received=5 packet-loss=0% min-rtt=1ms175us avg-rtt=1ms727us
max-rtt=1ms993us
CF bbc.co.uk
sent=5 received=5 packet-loss=0% min-rtt=1ms256us avg-rtt=1ms705us
max-rtt=1ms845us
CF first hop
sent=5 received=5 packet-loss=0% min-rtt=831us avg-rtt=1ms376us
max-rtt=1ms628us
CF l2tp.aa.net.uk
sent=5 received=5 packet-loss=0% min-rtt=1ms744us avg-rtt=1ms826us
max-rtt=1ms953us
PN l2tp.aa.net.uk
sent=5 received=5 packet-loss=0% min-rtt=2ms351us avg-rtt=2ms707us
max-rtt=3ms137us
PN 8.8.8.8
sent=5 received=5 packet-loss=0% min-rtt=2ms728us avg-rtt=3ms656us
max-rtt=4ms522us
PN 1.1.1.1
sent=5 received=5 packet-loss=0% min-rtt=2ms706us avg-rtt=3ms691us
max-rtt=4ms430us
PN 9.9.9.9
sent=5 received=5 packet-loss=0% min-rtt=2ms546us avg-rtt=2ms870us
max-rtt=3ms981us
PN bbc.co.uk
sent=5 received=5 packet-loss=0% min-rtt=2ms354us avg-rtt=2ms506us
max-rtt=2ms704us
Edited by ft247 (Thu 04-May-23 06:33:38)
|
|
|
Just to provide some context - I see 0.5 to 0.6ms between servers in datacentres in different buildings.
Similar to a peer over LINX, so that doesn't seem to add much.
I'm seeing 0.9 to 1ms over to LINX website which is in another building (over LINX but not actually LINX network) probably 2-3 miles away fibre wise.
Docklands to City across networks is 0.8ms-1ms++ (few blips to 1.5ms)
Docklands to City on same network 0.8ms-1ms
I'm also seeing 1.5ms on another circuit but could be the router on the end of it.
...my point is more general that it's not just the fibre path but the bits in between that could be the issue, so you may struggle with consistency. When you're talking about 1ms resolution you're going to struggle with quite a few risk factors.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
...my point is more general that it's not just the fibre path but the bits in between that could be the issue, so you may struggle with consistency. When you're talking about 1ms resolution you're going to struggle with quite a few risk factors.
exactly and its likely his input and out devices that will be the bottlenecks ..... rather than extra 0.001s or 0.006s
|
|
|
|
OP could shorten his commute and rent a flat on one of the very many new highrises on the Isle of Dogs and potentially have a sub ms ping with an apartment on HyperO for all that extra gaming he’ll get in by not having to take the tube across London twice a day. Just the small issue of rents. 😅
|
|
|
OP could shorten his commute and rent a flat on one of the very many new highrises on the Isle of Dogs and potentially have a sub ms ping with an apartment on HyperO for all that extra gaming he’ll get in by not having to take the tube across London twice a day. Just the small issue of rents. 😅
Note that a flat on the Isle of Dogs may not go back to the datacenter directly. It may end up somewhere else for aggregation. You'd need to really chat to the provider in question.. you might find some will use aggregators based in Slough so it goes Isle of Dogs -> Sloug -> Docklands... Also the servers might move.. you can't be 100% sure on these kinds of things .. and what is there today may change tomorrow. And you might not find all the above out from a traceroute.
Unless you're in the financial trading space where tiny latency causes massive issues (and the you think about which datacenter you're in and which switch fabric has lowest latency in switching) this is all something you can optimise but you can't be perfect.
I've had server clients with very specific requirements which we've accommodated (server to server rather than home). They wanted to be as close as possible to some other destinations. You could get a fibre to a DC (either Openreach EAD which goes via exchange or even from fibre operators) which gives you a fixed path to a known place.. but again it begs the question whether it's worth it.. Also 10G may have better performance than 1G.. do you go 10G?
Or just rent a rack in a datacenter and go play games there
seb
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
Or just rent a rack in a datacenter and go play games there  Might be a tight squeeze to get the family and furniture into the rack 🤣
|
|
|
Or just rent a rack in a datacenter and go play games there  Might be a tight squeeze to get the family and furniture into the rack 🤣
Might be a squeeze getting yourself into the rack too.
..and the rack would also cost more than the rental of your flat
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
Might be a tight squeeze to get the family and furniture into the rack
Whilst not one to make rash generalisations, I wonder if the family will hang around long …. corking ping times isn’t really the basis of happy family life.
|
|
|
|
Strikes me as a young single guy. Could be completely wrong though 🤣
|
|
|
|
|
|
|
Strikes me as a young single guy. Could be completely wrong though 🤣 With a 1ms ping the world is his oyster (card) 🤣
|
|
|
|
🤣🤣🤣
|
|
|
|
If want low ping times forget fibre, microwave is quicker link. There is a reasion why banks (High-Frequency Traders) use microwaves links where nanoseconds costs money.
If want data tranfers rate (ie upload/download rate MB/GB/PB) fibre wins every time.
The bigger issue is the gaming server. To make simple I use use term Tick this is to refer to work unit or gaming instruction. You can only send 1 instruction per tick that will be processed. This is done this way to manage work units so the server deltaTime stays fixed and fair for everyone.
The server is set up as have 30 ticks (33ms-34ms) or 60 ticks (16ms-17ms) a second. No matter what what your game system FPS it is irrelevant. It only 1 instuction recived/send for each server tick.
Let say computer has FPS of 200 so deltaTime is 5ms.
If server set at 60 ticks second nothing will change until at least 16ms so until you have displayed at least 4 frames. If at server set at 30 ticks second nothing until displayed at least 7 frames.
Anyway to get display that has deltaTime of 1ms you need at least 1000 FPS.
PS When you findout how long it take what see going throw our optic nerve is, you will see how realy pointless this is. There are lots of animals which have lot quicker then our version of the monkey brain.
|
|
|
PS When you findout how long it take what see going throw our optic nerve is, you will see how realy pointless this is. There are lots of animals which have lot quicker then our version of the monkey brain.
Yep its like saying "i can get you top place at google"......nearly pointless
|
|
|
|
Very interesting.... so you're telling me all the fights I have been losing have nothing to do with me not having 1ms latency and its a L2P issue?
|
|
|
The research says it about 13ms to identify images.
Also there 50 milliseconds it takes along the optic nerve.
https://news.mit.edu/2014/in-the-blink-of-an-eye-0116
|