Lets talk about NAT.

[el pescador]

Doing some research into problem i and many have with fifa and stumbled across some articles which basically says nat isnt standardised and that there will be different nat code around.

Got me thinking that this could be an issue when two players are playing over p2p (and even a server in bigger multiplayer games)

Could it be possible that one nat system is too good compared to another and gives its user a disadvantage compared to the bad nat?

 

[chadster766]

NAT is NAT no good or bad AFAIK.

Many gaming system use UPNP to open ports to bypass any NAT restrictions. Now there is definitely differences in UPNP versions. The new UPNP software versions provide the Open NAT gaming scenario everyone is looking for when it comes to Xboxes but not many routers run the newer version UPNP services or at least they didn't a few years ago but that could have change a lot since then.

 

[RMerlin]

The new UPNP software versions provide the Open NAT gaming scenario everyone is looking for when it comes to Xboxes but not many routers run the newer version UPNP services

What do you mean by that? NAT-PMP (or even PCP) versus UPNP?

I wish console manufacturers would properly document how they test things out. I see a lot of bogus "Your NAT isn't working properly" reports out there...

 

[RMerlin]

Note that there's more than one way to implement NAT.

Here's a quick primer about it: http://arstechnica.com/civis/viewtopic.php?t=678069

 

[chadster766]

What do you mean by that? NAT-PMP (or even PCP) versus UPNP?

I wish console manufacturers would properly document how they test things out. I see a lot of bogus "Your NAT isn't working properly" reports out there...

NAT and PAT are pretty much the same thing for the most part the difference being that with PAT you have more control over the ports used in the NAT translation.

AFAIK the newer versions of UPNP can quickly switch NAT destinations between multiple LAN Clients allowing for one port forward to handle communications for multiple clients creating OpenNAT for when more than one gaming console is on the network.

 

[RMerlin]

NAT and PAT are pretty much the same thing for the most part the difference being that with PAT you have more control over the ports used in the NAT translation.

Look at the difference between using MASQUERADE and SNAT in Linux's Netfilter as two potentially different implementation with a few minor differences.

AFAIK the newer versions of UPNP can quickly switch NAT destinations between multiple LAN Clients allowing for one port forward to handle communications for multiple clients creating OpenNAT for when more than one gaming console is on the network.

I don't see how this could be possible at a technical level. How can the router figure out which LAN client is intended to receive a specific packet if it got it on the same port? The only information the router has is source IPort and WAN IPort (which is the packet's destination).

(EDIT: removed the part about UPNP and IGD, as I keep mixing these two up).

 

[RMerlin]

BTW, this is why I keep UPNP/2.0 advertising disabled in my firmware:

https://github.com/miniupnp/miniupnp/issues/167

Might be interesting to retest this with newer Windows 10 builds.

 

[chadster766]

Look at the difference between using MASQUERADE and SNAT in Linux's Netfilter as two potentially different implementation with a few minor differences.



I don't see how this could be possible at a technical level. How can the router figure out which LAN client is intended to receive a specific packet if it got it on the same port? The only information the router has is source IPort and WAN IPort (which is the packet's destination).

(EDIT: removed the part about UPNP and IGD, as I keep mixing these two up).

The newer versions on UPNP Services software may just have faster reaction time and the UPNP clients change the port forwarding rule quickly enough to get there gaming communication through individually. It was a few years ago when I looked into this but the information was available on the Microsoft Xbox support site.

 

[chadster766]

BTW, this is why I keep UPNP/2.0 advertising disabled in my firmware:

https://github.com/miniupnp/miniupnp/issues/167

Might be interesting to retest this with newer Windows 10 builds.

Looks like the Windows 10 UPNP V2 issue you advise miniupnp about hasn't been updated in almost 6 months

 

[sfx2000]

Doing some research into problem i and many have with fifa and stumbled across some articles which basically says nat isnt standardised and that there will be different nat code around.

Got me thinking that this could be an issue when two players are playing over p2p (and even a server in bigger multiplayer games)

Could it be possible that one nat system is too good compared to another and gives its user a disadvantage compared to the bad nat?

NAT is very well standardized, and the uPNP specs are as well - most of the problems that I've seen with Consoles is that they're fairly non-standard in their implementation, and have been for years - was hoping with the current Generation (PS4/XBone), that more of the networking would be handled by the OS directly, but there continues to be problems here .

(and missed opportunities)

Maybe MSFT will get it right with the Win10/XBone rollout...

And we'll continue to see more NAT'ing in IPv4 - whether it's regular NAT, or Carrier Grade NAT - so this is a problem that the console OEM's need to solve. I would also add the Game ISV's, as they also contribute to the problem...

 

[System Error Message]

just read the wikipedia article about NAT, you may not understand it but once you do you will understand that NAT methodology is the same across all implementations with the implementations themselves being different. For example a few months ago a bug was found in ubiquiti's NAT involving SSL that caused 2 different hosts visiting the same SSL website that it didnt work for the 2nd person. Its not a serious bug but it shows that implementations can differ however this is not the cause of concern for gamers especially console. Just use UPNP and it will work fine, the issue however is with UPNP implementation and not NAT.

Also my advice is to stop playing FIFA. I know a lot of people like football/soccer but the game could just be updated yearly for new info rather than an entirely new game for every year which is essentially just EA monetising the game in a greedy way. Its like taking the same game code and just updating the content every year of the same kind and calling it a new game instead of an update or DLC.

If your ISP performs NAT this can be an issue as they will not implement UPNP. manual port forwarding if they support it can be a way or perhaps using a VPN that doesnt perform NAT that is close to you with as low latency as possible.

 

[sfx2000]

Carrier Grade NAT is the most insidious one, as most folks might not be aware of it... and this will cause no end of pain for gamers and folks that are looking to "phone home"...

If your WANIP is somewhere inside the 100.64.0.0/10 block - you're CGNAT'ed...

And we probably shouldn't (or should) mention pain with NAT64...

 

[el pescador]

Thanks.
Ive tried asking my isp re NAT to no avail.
Ive tried the vpn and disabled nat and it does perform better.....but i thought that was down to better routing by the vpn.

Also odd one today.
Did a connection test on the ps4 and for the first time i got NAT type 3.
Played a game online on uncharted and checked again afterwards and it came up NAT type 2.
Ive got upnp and port forwarding off at the moment.
Testing games under different settings.
TBF no difference to having upnp on with port forwarding.

Uncharted is eratic from game to game.
With fifa i get the feeling my connection is helping my opponent in some way....they seem to be faster running turning passing etc....although this has improved since i changed to a symmetrical fibre service.

Something has changed on fifa going back to 2012.....still cant figure out what it is.
Used to be rock solid.
Still is from time to time.
When my isp had to reset the voip system gaming was superb for a couple of days.

 

[RMerlin]

Looks like the Windows 10 UPNP V2 issue you advise miniupnp about hasn't been updated in almost 6 months

Yeah, not totally surprising tho since it's Microsoft we're talking about here. Their track record speaks for itself... Still, might be worth retesting this issue under a recent Insider Build (I run them on my laptop) in case it might have been silently fixed.

 

[el pescador]

Dunno if anyone can help?
Excuse my wordings lol as im not anywhere as technical as most on here.


If nat coding is different could it be more or less efficient in dealing with packets coming in?
Could there be differences between two game players over p2p is what im getting at?
Say if one player has an older nat coding which then hands onto upnp (i presume thats what happens) but does so less efectively than say the asus firmware does.
Is this a possible scenario?
Lets say im sending my packets but they are getting more resistance than my oponents packets thus im getting a delay compared to his???

im racking me brains over why even though its improved by moving to symetrical some games i still get issues.

 

[System Error Message]

Dunno if anyone can help?
Excuse my wordings lol as im not anywhere as technical as most on here.


If nat coding is different could it be more or less efficient in dealing with packets coming in?
Could there be differences between two game players over p2p is what im getting at?
Say if one player has an older nat coding which then hands onto upnp (i presume thats what happens) but does so less efectively than say the asus firmware does.
Is this a possible scenario?
Lets say im sending my packets but they are getting more resistance than my oponents packets thus im getting a delay compared to his???

im racking me brains over why even though its improved by moving to symetrical some games i still get issues.

In networking i dont think the term resistance matters really even if you are talking about the resistance of copper.

In games especially FPS games except games that cater to laggers, having low latency + no packet loss is important. Say you have below 10ms ping and you have the free starter guns, enemy has been playing a while and has weapons you have to buy with real cash which are way better than what you have but enemy has 100ms ping, you will have the advantage regardless. Latency is very important and processing speeds arent really much of the concern as link speeds are. As long as some link between you and the server isnt full your latency will be consistent. If a node in between you and server has full CPU usage, you will see packet losses. If link between you and server is damaged (like damaged cabling) than you will see packet losses.

Coding for NAT or networking or any sort of stuff doesnt matter as long as the algorithm is essentially the same. Humans being humans will make mistakes in trying to follow that algorithm and how well it is coded for performance. NAT really isnt much of the issue with gaming as with PC games if you are not hosting than you do not need to worry about UPNP but consoles on the other hand tend to have issues because of the policies of the companies that design them (microsoft and sony). Their architectures for the games produce (not platform architecture, talking about software architecture) tend to be more towards monetising and greed rather than actual technical superiority. For example consoles need UPNP because they like to use p2p to avoid any sort of central server whereas a traditional PC game uses central server with gamers being client so only central server needs port forwarding or no NAT.

 

[Nullity]

Dunno if anyone can help?
Excuse my wordings lol as im not anywhere as technical as most on here.


If nat coding is different could it be more or less efficient in dealing with packets coming in?
Could there be differences between two game players over p2p is what im getting at?
Say if one player has an older nat coding which then hands onto upnp (i presume thats what happens) but does so less efectively than say the asus firmware does.
Is this a possible scenario?
Lets say im sending my packets but they are getting more resistance than my oponents packets thus im getting a delay compared to his???

im racking me brains over why even though its improved by moving to symetrical some games i still get issues.

By "resistance" do you mean latency? If so, I would not suspect NAT as the cause of your latency spikes. If anything, NAT would cause a service to fail partially or completely, rather than increasing latency.

Delay is likely caused by bufferbloat created by uploading/downloading while gaming.

 

[el pescador]

What i mean (excuse my bad wording) is one nat router may be very efficient dealing with nat but another may be slower....less efficient.
If there is different codings out there for nat the odds are they wont act exactly the same.

 

[Nullity]

What i mean (excuse my bad wording) is one nat router may be very efficient dealing with nat but another may be slower....less efficient.
If there is different codings out there for nat the odds are they wont act exactly the same.

True, but if you are experiencing latency spikes, NAT is not a likely culprit. Look elsewhere.

NAT latency wasn't a problem back when CPUs were 10x slower...

 

[el pescador]

Its not latency spikes im experiencing.
My opponents mostly seem to be faster more agile etc on fifa.