Lets talk about NAT.

[sfx2000]

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...

NAT shouldn't add any latency - but I think OP's problem is likely upstream - and the game he's playing - there are some operators that do a much better job than others...

One can control the latency inside one's own network, and some minor QoS tweaking perhaps, but many of the problems might be upstream with the ISP specifically, esp. if they've oversold capacity... and bufferbloat and latency spikes are a good indication of congestion at the ISP (or even their peer links to the internet backbone)...

 

[Nullity]

or maybe everyone else is better at the game...

 

[sfx2000]

or maybe everyone else is better at the game...

Perhaps - but it's common knowledge that there are some "advantages" with regards to certain regional ISP's - mostly in the quality of the WAN link...

It also goes to locality of the game servers - and where they are in relation to the player...

These are things that I do care about, the WAN side - and some do have more visibility that others on what's going on there... not saying I do have that visibility...

 

[Nullity]

Perhaps - but it's common knowledge that there are some "advantages" with regards to certain regional ISP's - mostly in the quality of the WAN link...

It also goes to locality of the game servers - and where they are in relation to the player...

These are things that I do care about, the WAN side - and some do have more visibility that others on what's going on there... not saying I do have that visibility...

There are also those who just kick butt and take names regardless...

boom
headshot

 

[sfx2000]

boom - mic drop

See what happens with a Modem Upgrade/Swap - the advantages of a quantified network... packet loss and latency - which translates into "twitch time"...

 

[el pescador]

So its more likely to be the isp than differences in nat?
Makes sense as back in 2012 i won bucket loads of div 1 titles and then changed to a faster fibre.
Fifa 13 came out about the same time and i noticed the issue then.
Went back tot he previous fifa to test and same problem.
Moved country and obviously isp since and they have the same issue.

re WAN.
Yesterday i was playing and noticed a slight degrading of play so switched off a few things on the router.
Turned off port fowarding and upnp.
Didnt reboot the router but the wan ip changed...could be coincidental.
5 mins later turned them back on and it changed again.
Is that normal?
Can understand if i rebooted the router.

 

[sfx2000]

Try different times of the day - residential access is usually busiest in the evening - quiet time is typically between 1AM to around 6AM, which is why ISP's do their maintenance work in that time frame...

 

[System Error Message]

It would be your ISP or your link between you and server/ other player.

Differences in NAT dont matter. As i said in my previous post if the CPU of a node between you and server/player is fully loaded you get dropped packets rather than increased latency.

I despise consoles and many games for this reason in the sense that they use p2p as a player to server link is better than a player to player link due to some ISP routing limitations especially when it crosses backbones and international borders. Player to player links require more upload as well essentially this is called p2p which is despised by many for the wrong reasons because VOIP is also p2p but torrents are also p2p so an ISP restricting p2p can hamper your console gaming. Games like console games and GTA 5 use p2p with the server taking stats and managing sessions.

Why not play other games rather than FIFA, i know many play it but to me i consider it an unethical game because of the way its made and all the politics involved in the game (like who gets a cut and so on).

In the traditional player to server link you can have to players on the opposite side of the country and server in the middle and the gameplay will be better or less laggy than if 2 players were directly connected to each other. The other thing is p2p incurs latency panelties due to the way internet is routed, something to do with paths. This doesnt affect all ISPs but its prevalent with greedy ISPs such as with cable.

 

[el pescador]

Agree but they use p2p and even had the cheek to say p2p works well in a vid last year....admittedly the speaker was doing hsi best to keep a straight face on that....

Problem is i like football so will probably be stupid enough to buy it again.
I am moving isp in sept when my contract is up.

I have heard orange use a secret qos within their network and can confirm some gamers with orange who even have to use the movistar infrastructure have far lower pings.....bizarre.

 

[el pescador]

This is off topic but anyone care to explain hwo this may affect p2p gaming?

https://en.wikipedia.org/wiki/P2P_caching

Came across something similar form a company offering a service related to this during my research.
The timeframe on the other article was around the time when fifa became bad for me.
Putting 2 and 2 together.........

My thinking would be that if your using p2p the isp will check every packet header to see fi they have the content in their own cache to save on transit costs?
Am i off the mark?

 

[el pescador]

Doing more research...stumbled across this.

P2P communication without a third party host
Some of the problems with UDP hole punching is that a third party server is always required to relay traffic or to establish UDP states between the machines. This increases the load on the third party server. There has been research to determine whether it is possible to establish P2P connections between systems without a third party host. Researchers Andreas Muller, Nathan Evans, Christian Grothoff and Samy Kamkar have discussed some theoretical methods in their paper “Autonomous NAT Traversal” which could help in achieving this goal. Their approach uses fake replies to establish connection between peers. As always, we have the same four cases.

Case 1: The victim and the client are both behind separate NATs.

Case 2: The victim and the client are both behind same NATs.

Case 3: The victim is behind a NAT whereas the client isn’t and has a global IP address.

Case 4: The client is behind a NAT whereas the victim isn’t and has a global IP address.

According to their paper, the technique proposed by them works quite well for cases 3 and 4 but doesn’t work quite as well for the others.

question is does 3 and 4 still connect and if so what are the possible problems?

http://resources.infosecinstitute.com/udp-hole-punching/

 

[Nullity]

Before wasting time researching NAT, please verify that NAT is causing your latency problems.

 

[el pescador]

How do i verify its nat causing latency?

 

[System Error Message]

Nat isnt the cause of latency issues. As i mentioned before if the CPU is fully loaded you get dropped packets. If the link is fully loaded you get increased latency. So NAT doesnt affect latency.

The issue you are having isnt NAT related, it is ISP p2p profiling. If you want to test try using VOIP like teamspeak for example and look for latency and packet drops.

 

[el pescador]

thanks.
havent tried teamspeak but did test on this website.

https://www.voipreview.org/speedtest

the results showed a big difference in download and uplaod consistency.
98% for download and on one test 63% for upload (second test 53%)

bufferbloat was visible in both downa and up...jitter was more prevalent in upload.

not good

 

[el pescador]

BTW isp isnt profiling p2p.
I have checked on a site which says they arent.

Thinking its a simple issue with upload and the quality of it.

 

[System Error Message]

and it is quite common for ISPs to restrict upload. Bufferbloat is a big problem.

 

[el pescador]

An update re fifa.
Isp changed the network they use to exit Spain.Telia.
Running pingplotter the nodes in that network change quite frequently.
At first i was puzzled but i presume its to route via the best available route for icmp packets (and i presume udp) if a node gets busy?
The game feels a lot better for this change.
get the odd laggy game but very little of what i was experiencing before...ie.my opponent able to run faster turn faster etc.

Re nat...my isp dont do ipv6 yet but am looking forward to using it avoiding nat my end all together.

 

[RMerlin]

Re nat...my isp dont do ipv6 yet but am looking forward to using it avoiding nat my end all together.

This is something I admit never crossed my mind, but IPv6 deployment will actually be a boon to online gamers whose game rely on P2P. Just need to open (for REAL, this time - those NAT guides will finally be technically accurate LOL!) the required ports in the firewall.

 

[sfx2000]

This is something I admit never crossed my mind, but IPv6 deployment will actually be a boon to online gamers whose game rely on P2P. Just need to open (for REAL, this time - those NAT guides will finally be technically accurate LOL!) the required ports in the firewall.

It should - but knowing the past - the console vendors will find a way to screw that up as well