Pages: [1] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 1 post(s) |
Klandi
Science and Trade Institute
|
Posted - 2010.08.18 13:11:00 -
[1]
This is primarily for the ppl at CCP (in particular Tanis whose blog I just read) concerning testing and sorting the basis of lag.
Firstly the methodology.
Any network related problem has the minmum of 4 parts - client, infrastructure, server and application. You should be capturing as many parts as possible. Making observations or recommendations on one part in semi-isolation leads to erroneous conclusions.
Secondly is the platform.
Sorry, imo if you don't do this on TQ then you will extend the fix cycle, and you can by developing a test team (players that will sign-up to a special program - which would be rewarded with additional game time) and will allow a specially configured client-side patch to be applied to their machines.
Then - you will be alot closer to sorting the basis of the problem.
There are alot more parts to the process, but you'll have to pay for those.
That is all - let the flaming begin ....
|
Stick Cult
Unspoken Autonomy.
|
Posted - 2010.08.18 13:13:00 -
[2]
tbh you'd get better (maybe even CCP) responses if you posted this in the comments section of the blog
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Draconyx
|
Posted - 2010.08.18 13:20:00 -
[3]
Dear backseat driver.
As stated above there is a forum for blog responses.
|
Jarjar
Deep Core Mining Inc.
|
Posted - 2010.08.18 13:20:00 -
[4]
Originally by: Klandi
Sorry, imo if you don't do this on TQ then you will extend the fix cycle, and you can by developing a test team (players that will sign-up to a special program - which would be rewarded with additional game time) and will allow a specially configured client-side patch to be applied to their machines.
That doesn't take care of the server-side logging. You can't add full-on debugging to TQ or it'd lag out before 10k players were on.
|
Stick Cult
Unspoken Autonomy.
|
Posted - 2010.08.18 13:21:00 -
[5]
Originally by: Jarjar
Originally by: Klandi
Sorry, imo if you don't do this on TQ then you will extend the fix cycle, and you can by developing a test team (players that will sign-up to a special program - which would be rewarded with additional game time) and will allow a specially configured client-side patch to be applied to their machines.
That doesn't take care of the server-side logging. You can't add full-on debugging to TQ or it'd lag out before 10k players were on.
Considering they almost broke the test server with all the logging they did, with only a few hundred players on...
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
JitaPriceChecker2
|
Posted - 2010.08.18 13:42:00 -
[6]
Edited by: JitaPriceChecker2 on 18/08/2010 13:41:45 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
|
Stick Cult
Unspoken Autonomy.
|
Posted - 2010.08.18 13:44:00 -
[7]
Originally by: JitaPriceChecker2 Edited by: JitaPriceChecker2 on 18/08/2010 13:41:45 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
Go read the dev blog comments if you want the answer, Oveur answers it nicely.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Tippia
Sunshine and Lollipops
|
Posted - 2010.08.18 13:45:00 -
[8]
Originally by: JitaPriceChecker2 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
Yes. ùùù ôIf you're not willing to fight for what you have in ≡v≡à you don't deserve it, and you will lose it.ö ù Karath Piki |
Professor Tarantula
Hedion University
|
Posted - 2010.08.18 13:49:00 -
[9]
Originally by: JitaPriceChecker2 Edited by: JitaPriceChecker2 on 18/08/2010 13:41:45 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
Yes, there was no lag at all for many years until Dominion. Not even a little bit.
My deepest sympathies. Prof. Tarantula, Esq. |
Darek Castigatus
Immortalis Inc. Shadow Cartel
|
Posted - 2010.08.18 18:22:00 -
[10]
Originally by: Stick Cult
Originally by: JitaPriceChecker2 Edited by: JitaPriceChecker2 on 18/08/2010 13:41:45 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
Go read the dev blog comments if you want the answer, Oveur answers it nicely.
But then he wont be able to troll like a tard anymore.
Oh wait......just go read the comments already
http://desusig.crumplecorn.com/sig.php |
|
Exploding Tukey
Gallente Capital Construction Research Pioneer Alliance
|
Posted - 2010.08.18 18:44:00 -
[11]
Originally by: JitaPriceChecker2 Edited by: JitaPriceChecker2 on 18/08/2010 13:41:45 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
Before EVE had 10,000 players = no lag After EVE had over 100,000 players = lag.
RE: Client side patch - This may seem really complicated, but the data coming out of your eve client is the same as the data going into the eve server, which can be monitored almost as easily. also if CCP wanted to introduce measurement tools into the client, they dont need to bribe us with game time, simply a matter of patch or dont play.
<Signature Type="Clever" Width=100% Height=5 Edible="No" /> |
BinaryIdiot
|
Posted - 2010.08.18 18:47:00 -
[12]
Originally by: Klandi Any network related problem has the minmum of 4 parts - client, infrastructure, server and application. You should be capturing as many parts as possible. Making observations or recommendations on one part in semi-isolation leads to erroneous conclusions.
I think you're just trying to act intelligent but what you said makes no sense.
Network related problems have a ton of parts. You only said 2.5 parts (the client is the application, the server is mostly the infrastructure).
You don't even know if it's related to network issues. Lag could easily be an issue stemming from inefficient code or maybe just poor resource management / division of tasks. Don't speculate and throw out random things. It doesn't help.
Originally by: Klandi Sorry, imo if you don't do this on TQ then you will extend the fix cycle, and you can by developing a test team (players that will sign-up to a special program - which would be rewarded with additional game time) and will allow a specially configured client-side patch to be applied to their machines.
Testing on a production server is downright stupid. Sure, they would have a large amount of users but they'd still need to coordinate everyone setting up a massive fleet and then, do you really want to deal with all of the folks who 'helped' asking for their ships back, ISK or SP compensation?
Let's not forget the amount of tracing that would need to be enabled which would slow down the entire game for everyone.
I don't even understand why this is such a big issue; setup some massive tests on singularity, capture everything that the pilots did then constantly replay it. Simple. You can even get more complex by trying to do this (and ONLY this) type of data capture on TQ.
When you have hundreds or thousands of scenarios it should be incredibly simple to test.
|
Stick Cult
Unspoken Autonomy.
|
Posted - 2010.08.18 18:48:00 -
[13]
Originally by: Exploding Tukey
Originally by: JitaPriceChecker2 Edited by: JitaPriceChecker2 on 18/08/2010 13:41:45 Before Dominion = no lag
Dominion rolled out = lag
Complicated ?
Before EVE had 10,000 players = no lag After EVE had over 100,000 players = lag.
RE: Client side patch - This may seem really complicated, but the data coming out of your eve client is the same as the data going into the eve server, which can be monitored almost as easily. also if CCP wanted to introduce measurement tools into the client, they dont need to bribe us with game time, simply a matter of patch or dont play.
The issue isn't the data coming out of your client/into the server, the issue is the server itself and how it processes things. Duuuuh...
Also, they like you to run the logserver during mass tests, too.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Syn Callibri
Minmatar Blacklight Incorporated
|
Posted - 2010.08.18 18:49:00 -
[14]
Originally by: Klandi This is primarily for the ppl at CCP (in particular Tanis whose blog I just read) concerning testing and sorting the basis of lag.
Firstly the methodology.
Any network related problem has the minmum of 4 parts - client, infrastructure, server and application. You should be capturing as many parts as possible. Making observations or recommendations on one part in semi-isolation leads to erroneous conclusions.
Secondly is the platform.
Sorry, imo if you don't do this on TQ then you will extend the fix cycle, and you can by developing a test team (players that will sign-up to a special program - which would be rewarded with additional game time) and will allow a specially configured client-side patch to be applied to their machines.
Then - you will be alot closer to sorting the basis of the problem.
There are alot more parts to the process, but you'll have to pay for those.
That is all - let the flaming begin ....
OHBOY...another, I have the solution to all of the problems thread.
Syn Callibri Ilharess of the Scorpion Clan
|
Stick Cult
Unspoken Autonomy.
|
Posted - 2010.08.18 18:53:00 -
[15]
Originally by: Syn Callibri
Originally by: Klandi ~~~~stuff~~~~
OHBOY...another, I have the solution to all of the problems thread.
For once, Callibri, I agree with you.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Klandi
Science and Trade Institute
|
Posted - 2010.08.18 19:46:00 -
[16]
Originally by: BinaryIdiot Rambling of an inconsequence and incompetent fool
Living up to your name I see, but I will bite
The client RUNS the application, the infrastructure DELIVERS the bits down the wire and the n-tiered server(s) SERVES the application to the client. Four distinct areas you B...Idiot
|
|
CCP Adida
C C P C C P Alliance
|
Posted - 2010.08.18 20:20:00 -
[17]
As mentioned before we do have an official thread going in regards of the most recent blog. You may want to post in there.
Adida Community Rep CCP Hf, EVE Online
|
|
|
|
|
Pages: [1] :: one page |
First page | Previous page | Next page | Last page |