Netcode

[This Forum Sucks]

Of particular importance: "Latency is an unavoidable fact of online games, caused by not only network latency, which is largely out of a game's control, but also latency inherent in the way game simulations are run"

As I've said multiple times, there are issues that go beyond simple lag, which is obviously an unavoidable fact of internet gaming.

I don't want the last word - but wouldn't mind a simple "that makes sense".

Are you really so desperate for validation that you need me to pat you on the head for pointing out obvious truths, even though they're irrelevant to the topic at hand? I mean, yeah, if someone bought a business or professional grade router instead of a consumer model, they'd be "rated for higher traffic" even if that would make zero difference to 99% of home setups. And yes, obviously connections over the internet suffer lag from time to time but again, lots of games manage to deal with that unavoidable fact without allowing every client connected to an instance to see their own distinct version of events. So well done mate, have a biscuit.

And again - I've never once said there aren't or can't be issues with the Cobra Engine's networking components - I've not had a reason to do any traffic analysis, so I can only speculate that it makes heavy use of UDP, so it very well could have some problems with error control, flow control or datagram routing. Though I, like so many others, believe if this were the case, the issues resulting from this would be far more widespread, that is, everyone would have them, most or all of the time.

Well, if you and so many others are correct and there are no problems with the engine that aren't down to user error/bad equipment, why do we see some of these bugs fixed in patches? I have a very hard time believing you or any of the other professed IT wizards in this thread are very good at your day jobs when you're so closed to the possibility of intermittent or hard to replicate bugs that only show up for a subset of users.

"But this other game runs perfectly..."

And is probably less demanding, at least in terms of network traffic. Less precision required, less information to update per tick, see the above articles.

You keep saying this but I'm yet to see any proof of it. Admittedly a player in a spaceship has six degrees of freedom, which is more than most games but most games have more players than the average Elite instance, plus on a client/server model, updates would need to be sent out for any and all NPCs involved (based on the fact that a multicrew instance can desync entirely without the game becoming unplayable, it's a reasonably safe bet that Elite isn't sending any such data and just relying on each client to determine NPC positions for itself).

 

[Farmboy]

Well, if you and so many others are correct and there are no problems with the engine that aren't down to user error/bad equipment, why do we see some of these bugs fixed in patches? I have a very hard time believing you or any of the other professed IT wizards in this thread are very good at your day jobs when you're so closed to the possibility of intermittent or hard to replicate bugs that only show up for a subset of users.

No-one has said this. It is just that other causes are more probable. What you can't do is have a default assumption that 'this is a bug' because frequently it is not. In our experience. Which some of have quite a lot of.

You keep saying this but I'm yet to see any proof of it. Admittedly a player in a spaceship has six degrees of freedom, which is more than most games but most games have more players than the average Elite instance, plus on a client/server model, updates would need to be sent out for any and all NPCs involved (based on the fact that a multicrew instance can desync entirely without the game becoming unplayable, it's a reasonably safe bet that Elite isn't sending any such data and just relying on each client to determine NPC positions for itself).

You aren't doing anything but saying 'I don't believe you' here. Some games use more network data and/or are more latency sensitive, some use less and are less sensitive. This is self evident - what proof are you looking to see?

If you have a specific game which performs perfectly on your client, where ED is performing badly, then you can gather evidence - in the form of network logs/wireshark traces - and you can analyse these to see the number and size of packets going backwards and forward and get an idea of whether there is a correlation between performance and network load. Trying to guess what that might be from first principles based on what you see in the game is like trying to read a book across a room with a telescope. You might be able to do it, but crossing the room to get closer to the source of information is much more likely to be effective.

 

[This Forum Sucks]

No-one has said this.

Plenty of people are implying just that. Or is the constant harping about how terrible end users are and how they all have tons of spyware and jam in their computers whereas all of the enlightened professionals can play for hours a day, for days on end without having a single issue just small talk as far as you're concerned?

You aren't doing anything but saying 'I don't believe you' here. Some games use more network data and/or are more latency sensitive, some use less and are less sensitive. This is self evident - what proof are you looking to see?

I don't know, maybe some proof of what he's repeatedly claiming? You're absolutely right to say I don't believe him, I think he's mistakenly assuming that because only one of us has felt the need to boast about how much we know that the other must know nothing, so he reckons he can bluff his way through by sounding authoritative. As I pointed out, it would appear that the game is trusting each client in an instance to determine NPC locations for itself (which would leave a relatively small amount of data to be transferred) but I'm perfectly happy to be put right if that's not the case.

If you have a specific game which performs perfectly on your client, where ED is performing badly, then you can gather evidence - in the form of network logs/wireshark traces - and you can analyse these to see the number and size of packets going backwards and forward and get an idea of whether there is a correlation between performance and network load. Trying to guess what that might be from first principles based on what you see in the game is like trying to read a book across a room with a telescope. You might be able to do it, but crossing the room to get closer to the source of information is much more likely to be effective.

I think you're also mistakenly assuming that you can throw around a bit of jargon and everyone will be so impressed they'll give up and concede that maybe it's all down to user error. I'm pretty sure that if you ran into a recurring issue during gameplay, you would report it as a bug without first doing a full research paper on network traffic, so don't act like that's a reasonable standard to hold anyone else to.

 

[thirstyfish]

Why are two different sets of people turning this thread into a contest of measuring roosters?

Seriously unhelpful and unproductive.

 

[shadragon]

I've heard good things about Ubiquiti gear.

We use a couple of dozen Ubiquiti devices across a ten acre campus with 400+ people and it works brilliantly with easy central management. However, I would not recommend WiFi for playing Elite as it is fifty times slower than an Ethernet cable connection. Save yourself a lot of issues and plug in that laptop or PC.

I've got one of these for Wifi. I have my TV, Rasp Pi, tablets and phones all connected to it via Wifi for watching videos from my Home NAS. My PC is plugged into the back of it with a CAT 6 cable.

 

[Dbrn47]

We use a couple of dozen Ubiquiti devices across a ten acre campus with 400+ people and it works brilliantly with easy central management. However, I would not recommend WiFi for playing Elite as it is fifty times slower than an Ethernet cable connection. Save yourself a lot of issues and plug in that laptop or PC.

I've got one of these for Wifi. I have my TV, Rasp Pi, tablets and phones all connected to it via Wifi for watching videos from my Home NAS. My PC is plugged into the back of it with a CAT 6 cable.

Haven't had any issues with being on WiFi. Latency is a tad higher but speed is the same.

 

[shadragon]

Well, if you and so many others are correct and there are no problems with the engine that aren't down to user error/bad equipment, why do we see some of these bugs fixed in patches? I have a very hard time believing you or any of the other professed IT wizards in this thread are very good at your day jobs when you're so closed to the possibility of intermittent or hard to replicate bugs that only show up for a subset of users.

Issues that affect everyone are dead simple to fix because they are obvious. No one can get on the Internet, then it has to be the firewall or a specific switch. An issue with an individual or sub-set of people, if you like, is harder to find because it only affects them. Right away, it has to be something local to them. Their OS, their drivers, their connection to the switch, their operating environment, etc, but which? Today alone, our helpdesk handled a corrupt print driver screwing up only three people, a PDF file that wouldn't print for one person, but did for someone across the room and a Mac user locked out of our Windows file server.

Dealing with sub-sets of people with issues is what IT does. We logically track down the commonalities of the issue, if a group, or find the differences, if an individual.

If 95% of Elite players are not having issues with a given feature, but 5% are, then what is the cause of that problem? Drivers, OS, BIOS, hardware, type of networking gear in use or some other factor? It's harder because A) the devs have to take descriptions of the problem from users who don't give enough info, may report unneeded things or not represent the issue accurately. How do you fault find a connection when the sum total of the bug report is "Networking is broken again. Fix it"? B) They don't have physical access to the system and must fault find it remotely through game log files (if they are sent them) that don't give the whole picture. C) Time differences. I once had to fault find through error logs on two servers that had clocks out by twelve seconds. A nightmare. The devs deal with global time zones. D) Elite is a complex beast and how it interacts with a specific version of firmware on a Taiwan NIC card put out five years ago by a company that no longer exists, can cause issues.

Yes, they can fix some things in patches, but not all. They can't fix hardware that was set up improperly, or installed with the wrong driver version, or has a corrupt Windows DLL, or DDR RAM that has the wrong timings, or malware infestation, and on and on and on.

From personal experience, 80% of computer issues I run into are avoidable. They're caused by people abusing systems or networks and doing dumb things. Add to that the folks who don't maintain their computers properly. You wonder if they run their cars continually without changing oil, checking tire pressures, and ignoring warning lights. Frontier can't fix stupid.

You don't believe it, because you don't do it for a living. I do. The boffins behind the scenes at Frontier are exceptional at what they do. You should see some of the train wreck companies out there I've encountered and how they operate.

- - - Updated - - -

Haven't had any issues with being on WiFi. Latency is a tad higher but speed is the same.

Wifi is half-duplex and cannot send / receive at the same time. Ethernet is full duplex and is a more efficient medium, not subject to EMF or sunspots either.
You can use Wifi, but Ethernet will give you less trouble long term, especially if playing games that depend on low latency timings like real time space shooters.

 

[IndigoWyrd]

I'd like to thank those who "get it". Those who refuse it, well.. I've done all I can do.

Are there bugs, glitches and issues in Elite? Sure.
Is EVERYTHING that goes wrong a bug, glitch or issue with Elite? Not in the slightest.

If it looks like a bug, then report it. If they ask you to enable logging and submit your logs, then do so.
And if they tell you "it must be you, because it's not us." - Well, you can argue all you like at that point.

Yes, some bugs, glitches and issues can be extremely difficult to track down - especially those that are intermittent and/or only manifest under certain uncommon conditions - and even more so when those conditions are related to the client-side.

 

[This Forum Sucks]

Right away, it has to be something local to them.

You mean it could be something local to them. I love that just because I've not boasted about my computing knowledge or joined in with the "end users are idiots" camaraderie your default approach is to talk to me like you were explaining fusion to a dog, it's very revealing.

You don't believe it, because you don't do it for a living.

Quite the opposite actually, I don't believe any of you are good at your jobs precisely because I've done it for a living and I know "if it's only affecting a few people it must be their fault" is as likely to be wrong as right. And yes, people aren't good at accurately describing the problems they're having and getting them to upload logs can be like pulling teeth and they have the temerity to use their PCs instead of treating them like sacred temples and so on and so on but that's just how things are and probably always will be. I'm pretty sure Frontier (and most other devs) aren't writing their game to only run on a perfectly maintained OS running on state of the art hardware - if nothing else that's a very small target market.

And if they tell you "it must be you, because it's not us."

Luckily for everyone concerned, Frontier support are a bit more professional than that.

 

[Dbrn47]

Wifi is half-duplex and cannot send / receive at the same time. Ethernet is full duplex and is a more efficient medium, not subject to EMF or sunspots either.
You can use Wifi, but Ethernet will give you less trouble long term, especially if playing games that depend on low latency timings like real time space shooters.

Yes, but I'm saying that in the context of Elite, my WiFi setup is working fine.

 

[xondk]

Measuring the pass-by and hitreg difference with frames. I record with 60fps and can tell per frame how long it takes for the game to recognize the ram or impact. With weapons it is quite easy. With rams I take the closest location of both ships and count how many frames it takes till the damage is applied (even though we never rammed on my end in the first place). Rearrange 1000ms to 60Hz and you get the difference. In this case it took 65 frames so slightly over 1000ms. With Rinzler who is located on the other end of the globe it takes up to 3000ms.

I'd say I would love to see the video.

But you also should realise it also does depend on your connection, and 3000ms generally should not happen if your connection is working properly.

My maximum direct ping world wide reaches a maximum of 200ms. If there's problems with the route it can increase, but tracing often reveals it is on receivers end not my network or even country network. Or server is simply under heavy load from whatever it is running.

And that is really the problem, and it is out of Frontier's hands, even if they had dedicated servers, this would be a problem, because a possible US player woulds till need to play with one from australia, and server could be located either at a midpoint or on either end, lowering potential quality.

 

[IndigoWyrd]

You mean it could be something local to them. I love that just because I've not boasted about my computing knowledge or joined in with the "end users are idiots" camaraderie your default approach is to talk to me like you were explaining fusion to a dog, it's very revealing.

Oh thank the Powers that Be. Finally. And explaining fusion to a dog might be less frustrating.
Two or three pages of this would never have happened had we gotten to this right off the bat.

Quite the opposite actually, I don't believe any of you are good at your jobs precisely because I've done it for a living and I know "if it's only affecting a few people it must be their fault" is as likely to be wrong as right. And yes, people aren't good at accurately describing the problems they're having and getting them to upload logs can be like pulling teeth and they have the temerity to use their PCs instead of treating them like sacred temples and so on and so on but that's just how things are and probably always will be. I'm pretty sure Frontier (and most other devs) aren't writing their game to only run on a perfectly maintained OS running on state of the art hardware - if nothing else that's a very small target market.

Doesn't mean it's a user-issue - but could, and often is, something that particular group of users have in common - a shared switch, network segment - any commonality between them. And it can be equally as difficult to determine what to even ask to locate those commonalities. And no, you can bet the development team isn't designing this to be run on ideal configurations, under ideal circumstances - though I'd be quite willing to wager they're designing this ON ideal, or close to ideal configurations and close to ideal circumstances. Most software developers DO work IN these kinds of conditions. This is also one of the reasons for out-of-network public beta tests as well - to see if things perform the same in a more "realistic" environment, as compared to a "lab" environment.

Luckily for everyone concerned, Frontier support are a bit more professional than that.

Well, they certainly phrase things more professionally, to be sure - or at the very least, would say "we're looking into this" or "have escalated this", which can mean "there's nothing we can do for this, but thanks anyways."

 

[Asp Explorer]

"Why does my network die when I tie up all the cables neat and tidy, wrapped around pencils, and plug the loose ends into all those spare holes on the rooter so they are all used up, to make sure no performance is wasted, and I get the fastest wifi possible using the shiny and drop-dead-gorgeous eithernot hub the cute guy at Asda just sold me?"

"Why does my laptop go faster when it's on top of the fridge?"

"Why is it called a wireless extender when I have to plug bloody wires into it!"

"My 100/10 is faster than your pathetic 50/50!"

Oh the list goes on. It's what lulzbuckets are filled with

 

[This Forum Sucks]

Oh thank the Powers that Be. Finally. And explaining fusion to a dog might be less frustrating.
Two or three pages of this would never have happened had we gotten to this right off the bat.

Haha, you're as bad at reading comprehension as you are at your day job it would seem.

Doesn't mean it's a user-issue - but could, and often is, something that particular group of users have in common - a shared switch, network segment - any commonality between them. And it can be equally as difficult to determine what to even ask to locate those commonalities. And no, you can bet the development team isn't designing this to be run on ideal configurations, under ideal circumstances - though I'd be quite willing to wager they're designing this ON ideal, or close to ideal configurations and close to ideal circumstances. Most software developers DO work IN these kinds of conditions. This is also one of the reasons for out-of-network public beta tests as well - to see if things perform the same in a more "realistic" environment, as compared to a "lab" environment.

"It may be the user's fault or it may not" is a far cry from your initial position of "there is nothing wrong with the game at all and anyone who suggests there might be is an idiot who doesn't have as much money as me", so I'm glad I've managed to talk you around.

Well, they certainly phrase things more professionally, to be sure - or at the very least, would say "we're looking into this" or "have escalated this", which can mean "there's nothing we can do for this, but thanks anyways."

Here's the thing mate, not everybody who works in support is as bad at their job as you are. Just because you'd be lazy enough to lie about escalating a problem to save yourself from doing any work, doesn't make it standard practice.

Oh the list goes on. It's what lulzbuckets are filled with

The list of exaggerated half-truths and outright lies you made up because you hold the people you're employed to help in open contempt in the mistaken belief that your knowledge makes you better than them? Is that the list you mean?

 

[shadragon]

Yes, but I'm saying that in the context of Elite, my WiFi setup is working fine.

I get that. Let me use an analogy.

In a BMW 7 series, a 4.4l V8 producing 473 bhp will work fine, where a 6.0l V12 544 bhp engine in the same car will work better.

Wifi is made for convenience, not performance.

 

[This Forum Sucks]

I get that. Let me use an analogy.

In a BMW 7 series, a 4.4l V8 producing 473 bhp will work fine, where a 6.0l V12 544 bhp engine in the same car will work better.

And I would assume they're both more than capable of reaching the maximum speed limit.

 

[Bunkerkind Anni]

you can argue about client side equipment and settings all you want,

its a fact that peer 2 peer is the least practical protocoll to use when it comes to multiplayer gaming with more then two player,
as the data to be transfered and confirmed by each client goes up expenentially.


has onyone discussing here followed the development of a game named "Factorio"?
a small indie game that started with a p2p netcode that didn't allow to have more then 3 player in one game for a longer time
they have rewritten the whole thing to client2server code and now you can have 400+ player in one game.

also,
things like "combat logging" cannot be prevented in a reasonable way when the data for the ships in an instance is handled by the clients...

 

[Soopyyy]

They have to move to dedicated servers, if they ever want a feasible, stable, multiplayer experience.

 

[Dbrn47]

I get that. Let me use an analogy.

In a BMW 7 series, a 4.4l V8 producing 473 bhp will work fine, where a 6.0l V12 544 bhp engine in the same car will work better.

Wifi is made for convenience, not performance.

I'm aware.

- - - Updated - - -

And I would assume they're both more than capable of reaching the maximum speed limit.

I think that sometimes, people just get caught up in the beating of the "WiFi is bad" drum. Arguing that twisted pair is better because of current wireless limitations was beside my point. I'd love to have fiber myself, but you can't always get what you want and Elite doesn't require it.