Hello Racers,

After last night's successful Brazilian host test we have now expanded that system.

You can now start a host in Netherlands, USA, Japan, Brazil, Australia, Hong Kong, Turkey.

You need Test Patch W59 to do that.

We can't guarantee that all these locations will remain available, but we would like your feedback on how well they do work for you.

Quote from Scawen :

Hello Racers,

After last night's successful Brazilian host test we have now expanded that system.

You can now start a host in Netherlands, USA, Japan, Brazil, Australia, Hong Kong, Turkey.

We can't guarantee that all these locations will remain available, but we would like your feedback on how well they do work for you.

Around 10-40ms in Istanbul using servers in Turkey.

Here you see the pings from Turkey server, all drivers joined from Turkey and everything looking fine, no problems with ping.

Thanks for the test results!

Yesterday we hosted a race (spontaneous and with no previous organization) using the Brazilian server, and the results were pretty good! People from different countries has joined, such as Argentina, Brazil, Venezuela and even Finland.
Ping for SA users where always below 30ms and for the Finland user, around 200ms. Also no one complained about hitlag or other issues related to the server connection, and for me personally, everything works just fine.

I'll keep hosting races this week using the Brazilian server and seeing if there is any issue. Thanks for the implementation

Although i think its not going to have great usage of those servers (i can be totally wrong), i liked it very much and i bet it will bring even more players out from the cracked game.

Quote from PedroXDBR :

*snip*
Although i think its not going to have great usage of those servers (i can be totally wrong)...
*snip*

Well, if those from the dark side would join, that would literally mean like 50-70k players joining (without exaggeration, not constantly online player count of course) = would definitely make those Brazil servers run on the limit

Scawen, I've did some more research. All address are from https://www.i3d.net/resources/tools/myip/. With China Unicom, the good ping situation will more often with Unicom oversea server available.

For i3D India, Mumbai 213.179.210.53, I see there is a Unicom Singapore to TATA Singapore, then TATA India so that is good ping with 120ms(but only with China Unicom)

For i3D Japan, first we saw Unicom Shanghai then straight forward NTT Japan. In this case oversea between two providers will cause signaficant congression, packet loss and high latency will occur here(still only with China Unicom).

i3D Hong Kong and Singapore are also bypass NTT Japan so it only become worse. However somehow the insim IP shows that Hong Kong server is providing by oneprovider, so it bypass Unicom USA, causing 320-380ms which is too weird. People using China Telecom reports no such issue and ping is around 110-150ms there. And BTW i3d Los Angeles has a ping of 250ms, also via Unicom USA.

Maybe we need something like brazil local or a more nearby place like Korea, or Hong Kong but a different provider. From my knowledge there are some GFW bypass service also directly using same "provider oversea outlet to bypass block/congress" method.

Looks like now we have a chance let LFS master server create a new server that using Unicom Japan to Japan server, which should improving connection quality.

About ping wise, what latency is minimum requirement can have some nice races?

Quote from kagurazakayukari :

Looks like now we have a chance let LFS master server create a new server that using Unicom Japan to Japan server, which should improving connection quality.

I'm not sure what you mean - has something changed?

I would be so happy if the connection from China to our server in Tokyo was improved.

Quote from kagurazakayukari :

About ping wise, what latency is minimum requirement can have some nice races?

I can't answer - It's not so simple as one number. The lower the number, the less warping there will be and you can get closer to the other cars with less risk of a bad collision.

Quote from kagurazakayukari :

About ping wise, what latency is minimum requirement can have some nice races?

It depends on the racing type and speed, if the prediction between packets is fine or not, like agressive trajectory on curbs will need lower ping to calm down glitches...
Anyway usually 500ms is usually the maximum you can tolerate, if everybody is around 500ms it will be complicated. 300ms is usually fine. 150-100ms is crazy good butter smooth racing, not a lot of huge fake high ping crashes. under 100ms is nice to have but almost overkill.

Quote from loopingz :

Anyway usually 500ms is usually the maximum you can tolerate, if everybody is around 500ms it will be complicated. 300ms is usually fine. 150-100ms is crazy good butter smooth racing, not a lot of huge fake high ping crashes. under 100ms is nice to have but almost overkill.

In my experience 300ms is terrible and will throw my car off the track if anybody joins the server or pits using buttons. Close racing is impossible.
100 is drive-able (not butter-smooth), but quite acceptable, 150 is starting to get into glitch territory. 40-80 is very good.

Quote from Scawen :

I'm not sure what you mean - has something changed?

The testing ip have Shanghai Unicom to Japan NTT, my server somehow added a node from Japan Unicom ,so we have Shanghai Unicom-Japan Unicom-Japan NTT so it’s kind of bypass Shanghai Unicom-Japan NTT congress situation. Not a big change but less packet shall losses.

We kind of have 80-190ms floating latency depend on if it‘s traffic jam. And it can't really improve more. Our community manager did some test and said it still not quite acceptable to make races due to latency, only use for city driving is okay. Maybe we still need local server.