Jump to content
LOTROCommunity

Lying about Lag and Performance: business as usual for Sapience


nosam9
 Share

Recommended Posts

Once again, Sapience blames the internet for any lag players experience. Note the hall-of-fame fanboy CDQ is quick to tell players it's just network problems and HE never has any lag.

 

Business as usual. Blaming performance issues on the internet and on players.

https://www.lotro.com/forums/showthread.php?538963-Lag-and-playability-have-improved-after-server-restart&p=7075938#post7075938

 

Why not just be honest, Sapience? Since most people know the game lags, it is ok to admit it. Why not say this is an issue that Turbine knows and cares about?

 

Seriously? You are going to pretend lag in Lotro is due to this:

 

Network issues outside of our control are often responsible for sudden spikes or other strange bursts of lag that seem to come out of no where. Often times in researching these we discover that a section of the world, or a portion of the Us is experiencing very heavy traffic issues, high packet loss, or routing issues that impact a number of services. They can last minutes or days. They're hard to track down and very much outside of our ability to control or correct. So this would be consistent with what you're saying.

 

Funny how the rest of the world and other games are able to function on the internet without constant performance issues and lag. But no, it could not be the game client or servers.

Link to comment
Share on other sites

And I'd be willing to bet if the lag disappeared (due to some external fix that Turbine had no part in), they would be quick to claim it as their own success. 'Our team has been working tirelessly over the weekend just for you!' And then banning/infracting, for 'trolling/provoking', anyone who dared point out that they clearly had no part in it.

Link to comment
Share on other sites

I wonder if this is because of their choice of backbone and their use of UDP over TCP?

 

I noticed something monitoring temps. LoTRO really works my CPU but not so much my GPU, just the opposite of nearly everything else I play.  

 

 

or they are just scaling back resources.  that turbine server farm is host to more than just Lotro after all.

 

me thinks if the LoTRO player demand exceeds resources allocated, scaling to demand doesn't happen past a certain point & the "network traffic"[/sic] issues increase.

Link to comment
Share on other sites

I wonder if this is because of their choice of backbone and their use of UDP over TCP?

 

I noticed something monitoring temps. LoTRO really works my CPU but not so much my GPU, just the opposite of nearly everything else I play.

No, UDP is the only way to go unless you want to go really crazy with complexity. A single TCP stream would be death.

I maintain that to me many of the lag instances specifically introduced with U6 look like they introduced a bug in how they react to exceptional UDP packets (missing, out of order, duplicate, late). Then the only hacker who knows how this is supposed to work was fired and there you go. That explanation fits very well with the pattern of who is seeing the problems when.

Link to comment
Share on other sites

so that wouldn't be packet priorities but just a poor UDP layer?

 

I know GW2 went TCP for sake of time constraints and don't seem to have near the problems, but Patrick W did say it was not what he wanted, just what he could implement succesfully and come back to later as needed. It worked well enough so they left it to concentrate on more serious issues (like their trading post, lol).

Link to comment
Share on other sites

I actually logged in my Guard near Harwick Tuesday evening and the lag was horrible. I don't remember ever seeing it so bad in that area in the past. Sap infracted me last year for speaking out about the lag and rebuking the Turbine cheerleaders. heh

Link to comment
Share on other sites

From what I have seen the last days the lags might be a problem that is more prominent on the old European servers. It was strange how Belegaer was lagging like hell but Landroval at the same time running quite smooth.

 

That seems to kill the internet hypothesis and points probably more at server resources and/or database problems for the servers migrated from Codemasters.

 

On the other hand a few players with high latency might affect the server in another way than all players having high latency. In the end it is surely not an Internet problem alone. Also the problem seems to be a lot less noticable after server reboots. This all just does not add together to result in mostly internet or client related problems.

Link to comment
Share on other sites

so that wouldn't be packet priorities but just a poor UDP layer?

 

I know GW2 went TCP for sake of time constraints and don't seem to have near the problems, but Patrick W did say it was not what he wanted, just what he could implement succesfully and come back to later as needed. It worked well enough so they left it to concentrate on more serious issues (like their trading post, lol).

Does GW2 use multiple TCP connections?

It is true that TCP is generally much more usable than it was. ISPs still drop packets but in a manner that is dealt with efficiently with state of the art TCP implementations. And Microsoft did clean up some of the TCP issues in windows.

My speculation about Turbine's UDP handling comes from people observing clearly that lag happens when there is no change in ping response time to the same IP address. But we don't know what is happening behind that. If you use UDP you have to build your reactions to bad packets (missing, out of order, duplicate) yourself, and you usually do that in asynchronous code. That is hard to get right. And very expensive to test.

So you find a lot of code in the wild that uses UDP because it doesn't like TCP's recovery behavior but then screws up doing the same recovery when re-implementing half of TCP on top of UDP. Typically this isn't broken from the beginning. In the beginning they have a wizard who can get this right, then he fuck-this-shits it and leaves. The next person pokes around in the code until it does what is wanted and still works for the normal use case - having broken some more or less exotic sequence of UDP packets arriving in the wild.

Link to comment
Share on other sites

I'm not sure if Sapience just doesn't get it or if he does and is desperate enough to try this anyway.

 

It almost doesn't matter if the problem is with Turbine or 'the internet'.  If people continually can't play LOTRO due to lag, but they can play RIFT or SWTOR or WoW or other games, regardless of whose fault it is, they'll end up playing the playable game.  There's a level of dishonesty and insincerity in Sapience's actions that makes me think maybe he is convinced he can convince others that they should keep playing LOTRO so long as the fault of its unplayability lies elsewhere.  Or maybe he knows better but has no other options.  He has, though, shown himself to not be above using scapegoats in his public communications.

Link to comment
Share on other sites

I'm not sure if Sapience just doesn't get it or if he does and is desperate enough to try this anyway.

 

It almost doesn't matter if the problem is with Turbine or 'the internet'.  If people continually can't play LOTRO due to lag, but they can play RIFT or SWTOR or WoW or other games, regardless of whose fault it is, they'll end up playing the playable game.  There's a level of dishonesty and insincerity in Sapience's actions that makes me think maybe he is convinced he can convince others that they should keep playing LOTRO so long as the fault of its unplayability lies elsewhere.  Or maybe he knows better but has no other options.  He has, though, shown himself to not be above using scapegoats in his public communications.

 

I have a relatively high end system. LOTRO takes an age to load on my computer, has low frame rates and lots of lag. I just recently tried to get into Rift again. It runs like a dream and the store opens almost immediately, whereas LOTRO's store can take up to 30 seconds to load on my system. I also play GW2, The Secret World and Final Fantasy Online and have also had Star Trek Online, SWTOR, ESO running on my computer too - all with no lag or problems. Yet they have the temerity to blame it on the Internet. Turbine are a joke. Do they have any decency?

Link to comment
Share on other sites

Does GW2 use multiple TCP connections?

 

I don't know, if I get things to work i am good. But the way their store and in-game wiki link works I would bet they use separate ones. Probably for the login too, in fact it has to be as you can establish a login and connect to/modify your account while patching.

 

There were some changes a few months ago for reducing lag in WvW so its possible they moved to UDP for gameplay, but I heard a comment from someone that the text issues in the store are due to a windowed TCP browser.

Link to comment
Share on other sites

I don't know, if I get things to work i am good. But the way their store and in-game wiki link works I would bet they use separate ones. Probably for the login too, in fact it has to be as you can establish a login and connect to/modify your account while patching.

 

There were some changes a few months ago for reducing lag in WvW so its possible they moved to UDP for gameplay, but I heard a comment from someone that the text issues in the store are due to a windowed TCP browser.

The big question is really whether they have separate TCP streams for different things like the store.

But theoretically in an environment that has plenty throughput for low data volume but is potentially lossy you could just have more than one TCP connection for the same thing (gameplay here) and duplicate everything. Then if network condition manage to outfox whatever TCP might be in your Windows you can just ignore the stalled one and use of the others. Of course now you also have to deal with discarding duplicate traffic. But it would be easier than reimplementing your own lame version of TCP on top of UDP and the cost is bandwidth which might be available.

Link to comment
Share on other sites

I assume they lucked out and the current network for most of their playerbase has good ping to Washington, and TCP just works overall.

 

If you build a UDP scheme to fire off duplicate packets at a set rate until it gets a response, didn't you just duplicate part of what TCP does for you? That's an honest question for you, not a meant to be snarky. I'd like to learn more.

Link to comment
Share on other sites

If you build a UDP scheme to fire off duplicate packets at a set rate until it gets a response, didn't you just duplicate part of what TCP does for you? That's an honest question for you, not a meant to be snarky. I'd like to learn more.

No, the most common reason why games use UDP is that you have control of when and how to retransmit data that didn't make it. If you use TCP the OS kernel is doing it, and the OS kernel will not skip anything. It will mercilessly delay later traffic in favor of getting ahold of earlier traffic. In a video game such as a mutiplayer OS you will just skip and go on with the game, ignoring some delayed data that has no meaning anymore. You can't tell TCP that a certain piece from the past is useless to you now. So with UDP you just continue doing new data while a TCP connection would still be stalled. It really didn't help that Windows TCP sucked for a long time and isn't exactly up to BSD standards today either.

In practice there are many grey zones and many systems built on UDP rapidly approach the complexity of a lame, buggy, untested version of TCP. Plus it is almost always fragile code, which is one reason why I suspect that Turbine's lag has been caused by whoever did their UDP juggling in the first place left and no nobody can mess with the code anymore. Plus Turbine is in Boston which makes it double difficult.

Link to comment
Share on other sites

No, the most common reason why games use UDP is that you have control of when and how to retransmit data that didn't make it. If you use TCP the OS kernel is doing it, and the OS kernel will not skip anything. It will mercilessly delay later traffic in favor of getting ahold of earlier traffic. In a video game such as a mutiplayer OS you will just skip and go on with the game, ignoring some delayed data that has no meaning anymore. You can't tell TCP that a certain piece from the past is useless to you now. So with UDP you just continue doing new data while a TCP connection would still be stalled. It really didn't help that Windows TCP sucked for a long time and isn't exactly up to BSD standards today either.

In practice there are many grey zones and many systems built on UDP rapidly approach the complexity of a lame, buggy, untested version of TCP. Plus it is almost always fragile code, which is one reason why I suspect that Turbine's lag has been caused by whoever did their UDP juggling in the first place left and no nobody can mess with the code anymore. Plus Turbine is in Boston which makes it double difficult.

 

That explained it all in one short sentence, thanks! It makes the client sit and twiddle its thumbs while it waits for the OS to sort it out, which doesn't have a clue or a wastebasket. Is that a government metaphor? LOL

Link to comment
Share on other sites

Remember, a power outage can knock out a server that runs the lotteries and it can take them months to fix it.

 

If that's true, then they either don't know what they are doing or they're really running low on money.

Hard to comprehend that they don't have insurance on their hardware, either.

The power outage thing always seemed a bit suspicious to me. Like the server farm/data center is in somebody's garage, or something...

Link to comment
Share on other sites

I'm not sure if Sapience just doesn't get it or if he does and is desperate enough to try this anyway.

Sapience knows very well the game has performance issues that cause lag and have nothing to do with the internet. He is intentionally lying. It is his strategy to try to keep players in the game, to get more players to try the game, and to get players to spend money in the store. Since they can't (or wont spend the money) to fix these problems, they instead lie about them to try to hide them from potential players/store purchasers.

 

If the truth came out clearly about how badly Lotro runs for some players, many people would stay away. The game seems to be badly coded and optimized. It will not run well on some good gaming rigs, and will run decently on much worse machines.

 

It seems the game is playable for most people, but it still can lag and have some big issues like long loading times. When Sapience blames this on the internet and refuses to admit the game had problems, he is being intentionally misleading. I don't think I would be happy in a job where I had to lie over and over again.

Link to comment
Share on other sites

LOTRO ran fine and was relatively lag free until about halfway through RoI(I think Darmokk stated possibly Update 6 - sounds about right).

All this tripe they are spewing about it being caused by issues beyond their control is BS. Of course there are times when backbones will suffer performance wise(contention, faults, you name it), but this wouldn't be the root cause of Turbines issues in this case.

Does anybody know what their other titles are like performance wise(lag)?

Link to comment
Share on other sites

Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...