I'll reiterate that I've gotten MUCH, MUCH less lag since switching to a unity with far fewer people. I went from being 4-6 second lag 95% of the time to having 0-2 second lag 95% of the time. I've done dozens of runs before and about a dozen after, so I don't think it's just luck.
It's not a perfect situation, but it was quite helpful for me.
Gaol: Unplayable Content Because Of Lag
Gaol: Unplayable content because of lag
RadialArcana said: »
The problem is:
No the problem is they probably dont even know its such a large scale problem for western population and how big it actually is. They think several people has some lag and its probably our connection, since no JP players are experiencing this. I would like them to actually see gameplay of someone with 6 sec delay and understand it happens to a lot of western players.
[+]
Asura.Sechs said: »
Wish it was 5-6 seconds for me! I could probably play with 5-6 seconds of lag! :-D
But on WHM that kind of lag is a no-go. I had our party call out a WHM trust and I was basically useless the one boss I was playing WHM.
SimonSes said: »
RadialArcana said: »
The problem is:
No the problem is they probably dont even know its such a large scale problem for western population and how big it actually is. They think several people has some lag and its probably our connection, since no JP players are experiencing this. I would like them to actually see gameplay of someone with 6 sec delay and understand it happens to a lot of western players.
Someone posted a video on pg 5
YouTube Video Placeholder
Jeez, there are delays literally in excess of 10 seconds in there. How anyone thought this was latency is beyond me.
[+]
RadialArcana said: »
SimonSes said: »
RadialArcana said: »
The problem is:
No the problem is they probably dont even know its such a large scale problem for western population and how big it actually is. They think several people has some lag and its probably our connection, since no JP players are experiencing this. I would like them to actually see gameplay of someone with 6 sec delay and understand it happens to a lot of western players.
Someone posted a video on pg 5
YouTube Video Placeholder
We were lucky to see his Unity chat as well here, showing he is in Jakoh. Geriond swears by low population Unities, and I absolutely believe it works for him. I, however, tried 4 NM fights on Jakoh(about 30ish people in Unity) and saw no difference in lag from being in Apururu Unity.
I posted earlier about ExpressVPN being night and day for lag in my personal experience. I did some testing yesterday both on and off VPN, and actually haven't been able to reproduce the lag that I initially had. I cleared an entire NM battle with 5 Trusts while OFF the VPN and experienced some lag and gearswap mishaps, but nothing like the 5sec I used to see. I've also sat in the lobby room while OFF the VPN and had no lag.
Being unable to reproduce the lag really stumps me. For visibility here is a video of me testing lag while ON ExpressVPN connected to their Tokyo server. I will also include my ping and tracert results while ON and OFF VPN:
Connected to ExpressVPN Tokyo Server:
Not connected to ExpressVPN Tokyo Server:
This morning's lobby lag test while connected to ExpressVPN Tokyo Server:
a 4 ping sample is not anything, please run it for longer before posting numbers(if you kick out the 328ms outliar, you're comparing 160ms to 189ms instead of 224ms.. you need more samples to see if the 328ms occurance is common or a one off)
also, still no possible explanation where the ping is the root cause, only several where it can be a contributing factor(either way, SE needs to fix their code)
also, still no possible explanation where the ping is the root cause, only several where it can be a contributing factor(either way, SE needs to fix their code)
Asura.Geriond said: »
I'll reiterate that I've gotten MUCH, MUCH less lag since switching to a unity with far fewer people. I went from being 4-6 second lag 95% of the time to having 0-2 second lag 95% of the time. I've done dozens of runs before and about a dozen after, so I don't think it's just luck.
It's not a perfect situation, but it was quite helpful for me.
It's not a perfect situation, but it was quite helpful for me.
Asura.Wotasu said: »
Asura.Geriond said: »
I'll reiterate that I've gotten MUCH, MUCH less lag since switching to a unity with far fewer people. I went from being 4-6 second lag 95% of the time to having 0-2 second lag 95% of the time. I've done dozens of runs before and about a dozen after, so I don't think it's just luck.
It's not a perfect situation, but it was quite helpful for me.
It's not a perfect situation, but it was quite helpful for me.
I did the same with my alt, made no difference to lag what so ever.
I can confirm it is possible to enter and have inventory load instantly and have 0 input lag even if it normally takes 3-4mins and 8-10 second input lag. Just happened on one of my chars randomly (same unity as other 2) and it was glorious.
This char has a few less trusts and only a few 99 and Master jobs but doubt that was it, noone there had lag.
This char has a few less trusts and only a few 99 and Master jobs but doubt that was it, noone there had lag.
You can enter this zone in different ways and have zero lag (lillith and a-b-c), then you go into this boss area and you get the lag.
The only difference is the trust and job lockout system in gaol, the trust lockout system sends data constantly since it's overlaying the limited list of trusts over the top of your real list and constantly checking it.
I would guess they are simply overloading the server connection limits with too much data (ffxi is capped at 56k right?), since some things seem to help the answer possibly is just to limit how much data you're sending/getting from the server.
So take off your linkshell, remove RoE quests, be on a small unity, put some more chat filters on etc. Try everything you can think of to limit how much data is being sent outside of what you need.
The only difference is the trust and job lockout system in gaol, the trust lockout system sends data constantly since it's overlaying the limited list of trusts over the top of your real list and constantly checking it.
I would guess they are simply overloading the server connection limits with too much data (ffxi is capped at 56k right?), since some things seem to help the answer possibly is just to limit how much data you're sending/getting from the server.
So take off your linkshell, remove RoE quests, be on a small unity, put some more chat filters on etc. Try everything you can think of to limit how much data is being sent outside of what you need.
I've been experiencing weird, random lag too. It's not my ISP, PC or anything like that.
If it's really just that, can't they simply remove the stupid limitation for trusts not being resummonable and make so you can't summon trusts while you have enmity to compensate?
Hopefully they do just that, we will see I guess.
Last night I did get one run where the Moglophone II escape item didn't load after several minutes, but my inventory bag did. I had to use Warp Ring to get out. It even caused a trade window to not appear on my screen, despite a party member telling me I had "cancelled the request"; the game just said "unable to process the trade" on my end. That was the worst lag I had experienced to date. Strangely enough, earlier in the same set of runs (like 30min earlier); I had zero lag on one of the staging areas/zones after warping down. I could do JAs and use items instantly. And as soon as the NM walked over, the 5+ second lag started. So yeah, it can be wildly random for no apparent reason.
Lakshmi.Buukki said: »
Last night I did get one run where the Moglophone II escape item didn't load
It NEVER loads for me lol
Wonder what is the additional variable that seems to be making this worse for me than it is for everybody else.
XD
That was the only run out of dozens that never loaded. Weekends are the worst.
I’m away from home with crappy internet at the moment and have been getting pretty annoyed with 5-10 second delays in gaol. Today I went in with my alt whm and it was drastically improved, like 1 second or so. Same internet connection, computer etc. My main is pretty decked out but my alt doesn’t have a lot of augmented gear or trusts.. same unity for the record though.
Tried to replicate the issue in Gaol, so I connected to FF11 via UK based VPN connection. My latency rose 200ms~ above the usual I get from direct, and I felt absolutely no difference in areas outside Gaol.
Once I entered Gaol, the scenario changed completely, 5~10+ second delays with spells, gearswap was a mess, and it was just a horrible experience.
It seems now that latencies within or above a certain threshold trigger this server side anamoly. Definitely needs to be addressed to Square Enix, as this can also be correlated with poor coding in Divergence zones, except the lag there is not as severe as Gaol.
I posted a video with my findings. I personally don't have lag issues when connecting direct (UAE > Japan), but I decided to test and share this for more insight.
At around 6:30 into the video, I exit to Rabao, at that point the lag is gone, and the spells cast without delay.
Once I entered Gaol, the scenario changed completely, 5~10+ second delays with spells, gearswap was a mess, and it was just a horrible experience.
It seems now that latencies within or above a certain threshold trigger this server side anamoly. Definitely needs to be addressed to Square Enix, as this can also be correlated with poor coding in Divergence zones, except the lag there is not as severe as Gaol.
I posted a video with my findings. I personally don't have lag issues when connecting direct (UAE > Japan), but I decided to test and share this for more insight.
YouTube Video Placeholder
At around 6:30 into the video, I exit to Rabao, at that point the lag is gone, and the spells cast without delay.
[+]
I've looked at the UDP packets the game sends and receives in Wireshark. There is always one incoming packet followed by an outgoing one. These are regularly sent back and forth 2-3 times per second (this number is possibly dependent on my ping), even if nothing is happening in-game like sitting in the mog house.
The packets that windower addons see, which I analyzed before, are grouped into these UDP packets. So when you enter Ody, all you see in Wireshark is the UDP packet size increasing, not more individual packets received.
In a typical laggy Ody situation the maximum UDP packet size I observed was roughly 1200 bytes. The packet size itself might be capped, it is generally good practice to keep UDP packet sizes low in case they get lost. 2-3 packets at 1200 bytes size equals roughly the R3000 bytes received cap I'm seeing in-game.
So what does this all mean? What if the game's network protocol just sends a packet, waits for the answer and then sends the next. With lower ping, you would observe more UDP packets per second. If JP players get more updates per second, they basically have a higher bandwidth to the server and all the useless trust list updates are no problem for them.
Maybe someone with a way lower ping than my 250, maybe on the US west coast, can confirm that they are seeing an R cap above 3000ish inside Odyssey?
Of course all of this is an educated guess, as we cannot look at the server code.
The packets that windower addons see, which I analyzed before, are grouped into these UDP packets. So when you enter Ody, all you see in Wireshark is the UDP packet size increasing, not more individual packets received.
In a typical laggy Ody situation the maximum UDP packet size I observed was roughly 1200 bytes. The packet size itself might be capped, it is generally good practice to keep UDP packet sizes low in case they get lost. 2-3 packets at 1200 bytes size equals roughly the R3000 bytes received cap I'm seeing in-game.
So what does this all mean? What if the game's network protocol just sends a packet, waits for the answer and then sends the next. With lower ping, you would observe more UDP packets per second. If JP players get more updates per second, they basically have a higher bandwidth to the server and all the useless trust list updates are no problem for them.
Maybe someone with a way lower ping than my 250, maybe on the US west coast, can confirm that they are seeing an R cap above 3000ish inside Odyssey?
Of course all of this is an educated guess, as we cannot look at the server code.
[+]
Packet size is definitely capped, but beyond that they also cap individual components of each packet. I did an experiment that supports this looking at inventory here:
https://www.ffxiah.com/forum/topic/55065/july-2020-version-update/10/#3529435
The basic takeaway is that each packet can only hold so many inventory updates, even though the actual packet could be much larger. Presumably, they do this so inventory updates don't choke out player updates and other things needed. Note that packets are encrypted with blowfish and compressed with zlib, so if you're looking at both wireshark and an ingame monitor, the ingame monitor can have as much as 3x more volume of data.
I would point out that while that is a very interesting theory, it doesn't quite explain why a VPN will cause such a noticable difference. While the packet is received by the VPN much faster, it still has to travel from the VPN to the client, and from the client back to the VPN, to trigger the next packet. Perhaps it is a backlog of packets getting queued up that ends up choking out other packets, and the small difference between VPN and not pushes that over a threshold such that you do not accumulate a backlog with VPN while you do without.
However, since it seems people experience this lag even immediately after zoning in when they are going to experience it(see the video of 8 second lag casting solo).. that still doesn't explain why latency is such a strong contributor. If the trust-related updates were taking up the majority of a packet's size, you'd see exponential growth in backlog once things start happening in the instance.
It's certainly an interesting problem to theorize about, but the bottom line is still the same, if you want real action you need to pressure SE and make it clear how bad things are. Even if we are able to identify why latency is such a strong component, it's clear the cause is related to the volume of traffic generated by the server, likely primarily due to the spell updates.
https://www.ffxiah.com/forum/topic/55065/july-2020-version-update/10/#3529435
The basic takeaway is that each packet can only hold so many inventory updates, even though the actual packet could be much larger. Presumably, they do this so inventory updates don't choke out player updates and other things needed. Note that packets are encrypted with blowfish and compressed with zlib, so if you're looking at both wireshark and an ingame monitor, the ingame monitor can have as much as 3x more volume of data.
I would point out that while that is a very interesting theory, it doesn't quite explain why a VPN will cause such a noticable difference. While the packet is received by the VPN much faster, it still has to travel from the VPN to the client, and from the client back to the VPN, to trigger the next packet. Perhaps it is a backlog of packets getting queued up that ends up choking out other packets, and the small difference between VPN and not pushes that over a threshold such that you do not accumulate a backlog with VPN while you do without.
However, since it seems people experience this lag even immediately after zoning in when they are going to experience it(see the video of 8 second lag casting solo).. that still doesn't explain why latency is such a strong contributor. If the trust-related updates were taking up the majority of a packet's size, you'd see exponential growth in backlog once things start happening in the instance.
It's certainly an interesting problem to theorize about, but the bottom line is still the same, if you want real action you need to pressure SE and make it clear how bad things are. Even if we are able to identify why latency is such a strong component, it's clear the cause is related to the volume of traffic generated by the server, likely primarily due to the spell updates.
[+]
Hopefully they just disable the trust lockout system with the next patch, it would be easy to do and wouldn't really impact difficulty much.
What difference does it really make to not be able to use healer 1 for the next 2 fights and instead have to use healer 2 and then 3. Needless farting about.
The job lockout system is already fine in terms of forcing players to change things up in the run.
What difference does it really make to not be able to use healer 1 for the next 2 fights and instead have to use healer 2 and then 3. Needless farting about.
The job lockout system is already fine in terms of forcing players to change things up in the run.
[+]
Shiva.Thorny said: »
Note that packets are encrypted with blowfish and compressed with zlib, so if you're looking at both wireshark and an ingame monitor, the ingame monitor can have as much as 3x more volume of data.
Shiva.Thorny said: »
If the trust-related updates were taking up the majority of a packet's size, you'd see exponential growth in backlog once things start happening in the instance.
Shiva.Thorny said: »
I would point out that while that is a very interesting theory, it doesn't quite explain why a VPN will cause such a noticable difference.
Shiva.Thorny said: »
It's certainly an interesting problem to theorize about, but the bottom line is still the same, if you want real action you need to pressure SE and make it clear how bad things are.
I mean, you could, ya know, not do the content.
Nothing says this is broken, quite like not putting up with it.
Batshit crazy idea, I know.
Nothing says this is broken, quite like not putting up with it.
Batshit crazy idea, I know.
Asura.Eiryl said: »
I mean, you could, ya know, not do the content.
Nothing says this is broken, quite like not putting up with it.
Batshit crazy idea, I know.
Nothing says this is broken, quite like not putting up with it.
Batshit crazy idea, I know.
I've cleared the NMs once and just farm segments now and wait until they fix it since you can spam it nonstop afterwards as long as you have segments to buy the mooglephone2s.... I wouldn't bother with V5+ for augmenting this ***until it's playable.
[+]
The lag is bad but the queue on Asura is even worse at times. Dumping 100k segments here is like is Lilith V 2.0 lol
If you do /jump it fires off instantly, if you type /wave it has the long delay.
All FFXI content and images © 2002-2024 SQUARE ENIX CO., LTD. FINAL
FANTASY is a registered trademark of Square Enix Co., Ltd.