These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE General Discussion

 
  • Topic is locked indefinitely.
 

A Permanent fix to a long standing problem: Node crashing

First post
Author
Dinsdale Pirannha
Pirannha Corp
#121 - 2013-12-16 22:03:13 UTC
Gizznitt Malikite wrote:

Truthfully, I think TiDi generally goes the wrong direction.

When time slows down in game, it gives pilots tons of RL time to organize more people and move them to the fight. This creates 1000v1000 battles, which is good publicity, but will always be abused and pushed to the point that the system breaks (node crashes, slow-ass game play, etc).

Instead, if it is possible, speed up time instead of slow it down. As time speeds up, the time to escalate a fight diminishes. This rewards players pre-positioned to escallate. This also is self-correcting, as faster time means more people die which means less pilots on the field which results in less demand on the server.

Unfortunately, this would be extremely non-trivial to optimize. Speeding up the server adds load on the server now with the intention that the server's load will lesson as things come to a conclusion faster. There also might be some exploits this creates (suddenly you can increase your fleets dps by bringing so many people). I don't know if this is at all possible, but it is a conter-intuitive approach that might result in better overall gameplay.



How about this:

If the node crashes, there is a percentage chance that any ships in that node are totally destroyed, and pilots podded.
Suddenly there is risk involved with crashing a node.

Of course, that means that people will just flood a system with T1 frigs if there are any supercaps in that node.

There is simply no solution for this blob-sec problem.
But it is just that, a blob-sec problem, that only affects a tiny percent of the overall subscription base.

And CCP should just stop throwing any resources at it, and focus those resources on the majority of the sub base.
Pinky Hops
Caldari Provisions
Caldari State
#122 - 2013-12-16 22:03:28 UTC
The reason node crashes happen in the end is simple.

The gameplay mechanics encourage it to happen. Nothing more complicated than that.

If ownership of a system is coming down to a single timing window, then the system is going to be mobbed during that window.

If timing windows weren't so incredibly important for sov, I don't think node crashes would happen anywhere near as often. They would still happen as blobbing is inherently strong -- but there would be a lot less of it.
Tippia
Sunshine and Lollipops
#123 - 2013-12-16 22:08:56 UTC  |  Edited by: Tippia
Mr Sniggle-Worth Onzo wrote:
On any given day there are more ships blown up in Low sec / High sec.
So? How is that relevant to… well… anything, really? And what do you base that claim on? Or are you retracting your previous nonsense about how it's a nullsec problem since apparently stuff blow up everywhere?

Quote:
To say you can't crash a node is pretty silly tbh. It can happen pretty easy.
It happens, but it requires a fair amount of work under the right circumstances to do. The problem is that doing so means everyone loses, so it's not something that benefits anyone.

Quote:
Null is broken
In what way?

Quote:
When do you think that you will see Titans used again in a large scale battle ? w/o the node crashing.
When will you see 200+ Carriers vs 200+ Dreads w/o the node crashing ?
Oh pretty much all the time. It's not like there's a fixed limit on the ships and numbers that are allowed into the system before it automatically crashes.

So, again:
• In what way is nullsec broken?
• How is it relevant where ships die?
• Do you have any kind of numbers to support your assertions about where ships die?
• How does giving large coalitions an unassailable stranglehold on all of null open up more of null?
• How is giving one fleet an unassailable advantage supposed to generate more fights and more willingness to fight?
• How do you propose to ensure that null is opened up without giving massive advantages to the established players?
• Why are you proposing a solution that lets fewer people play in null?
• What makes you think that anyone is benefitting from node crashes (where everyone loses) but won't benefit from poplocks (where one side automatically wins)?
Vald Tegor
Empyrean Guard
Tactical Narcotics Team
#124 - 2013-12-16 22:13:53 UTC
Mr Sniggle-Worth Onzo wrote:

Ok whats to stop the next fight from crashing again?

Why do you care?

You do realize these fights get put on their own node? When they crash, only that system is affected. You can choose not to be there.

According to your claims, more ships are lost in high and low sec in countless small fights every day than there are in null. You are free to participate in them with no TIDI or crashing. So why do you care about what happens in null? There are greener pastures in New Eden for the kind of gameplay you are looking for. Leave null to those of us who enjoy participating in Homeworld Online, as slow and soul crushing as it gets.

What seems to be your problem with renting? Yesterday's node crashing fight saw about 300 billion in ships destroyed. Where do you think the replacements are coming from? What exactly do you think happens with rental income?
Mr Sniggle-Worth Onzo
Caldari Provisions
Caldari State
#125 - 2013-12-16 22:25:13 UTC
Its simple really!

By limiting the number of people to a system ( like in Jita ) CCP remains in control, not the players.

Limiting the number of people in a system would allow for many more people to come into Null and have the good fights and maybe take them some SoV.

How much Sov is empty ??

How many people in Null really do enjoy tidi fights ?

How many fights in the past year have ended in a node crashing ?

There is a reason why so many people refuse to even consider the idea of limiting the number of people to a system.
they would no longer be in control the game would have a new mechanic that would affect only a very small number of players.


The only people that would be affected by limiting the number of people in local would be very small


Who is it that benefits from the tidi fights ? Node/ system crash fights ? Has there been any Sov lost or gained from having the most people in a system ?


And for the people that don't know most of the large fights in Null are won before they start.
Tippia
Sunshine and Lollipops
#126 - 2013-12-16 22:29:31 UTC
Mr Sniggle-Worth Onzo wrote:
Its simple really!
Yes its: you can't come up with any kind of explanation of what the problem is or how your supposed solution is going to fix that problem.

So, again:
• In what way is nullsec broken?
• How is it relevant where ships die?
• Do you have any kind of numbers to support your assertions about where ships die?
• How does giving large coalitions an unassailable stranglehold on all of null open up more of null?
• How is giving one fleet an unassailable advantage supposed to generate more fights and more willingness to fight?
• How do you propose to ensure that null is opened up without giving massive advantages to the established players?
• Why are you proposing a solution that lets fewer people play in null?
• What makes you think that anyone is benefitting from node crashes (where everyone loses) but won't benefit from poplocks (where one side automatically wins)?
Tyberius Franklin
Federal Navy Academy
Gallente Federation
#127 - 2013-12-16 22:29:35 UTC
Gizznitt Malikite wrote:

Truthfully, I think TiDi generally goes the wrong direction.

When time slows down in game, it gives pilots tons of RL time to organize more people and move them to the fight. This creates 1000v1000 battles, which is good publicity, but will always be abused and pushed to the point that the system breaks (node crashes, slow-ass game play, etc).

Instead, if it is possible, speed up time instead of slow it down. As time speeds up, the time to escalate a fight diminishes. This rewards players pre-positioned to escallate. This also is self-correcting, as faster time means more people die which means less pilots on the field which results in less demand on the server.

Unfortunately, this would be extremely non-trivial to optimize. Speeding up the server adds load on the server now with the intention that the server's load will lesson as things come to a conclusion faster. There also might be some exploits this creates (suddenly you can increase your fleets dps by bringing so many people). I don't know if this is at all possible, but it is a conter-intuitive approach that might result in better overall gameplay.

Time isn't slowing down as a cosmetic feature. It's actually the core mechanism by which Tidi works as advertised, allowing more real time for the server to keep up with the activity going on in the slowed game time. Sped up Tidi only makes sense if you thought pre-tidi node crashes and selective unplayability didn't happen fast enough. It's not so much a solution as an attempt to turn a few hundred people in a system into a node death sentence.
Jill Chastot
Black Water Oasis
#128 - 2013-12-16 22:35:42 UTC
Mr Sniggle-Worth Onzo wrote:
Not only will limiting the number of pilots to the system increase the strategic planning for fights but will open up Null sec to many more low sec/ high sec players that wish not to be apart of the null carebear Zergs. Null is broken / large fleet fights are broken this will fix that.


New meta, Alt filled systems. Good job OP.

You win EVE

(obligatory biomass comment here)

https://forums.eveonline.com/default.aspx?g=posts&t=298596&find=unread OATHS wants you. Come to the WH "Safety in eve is the greatest fallacy you will ever encounter. Once you accept this you will truely enjoy this game."

James Amril-Kesh
Viziam
Amarr Empire
#129 - 2013-12-16 22:41:19 UTC
There should be a rule about "mind-blowingly stupid ideas" so threads like these would get nipped in the bud.

Enjoying the rain today? ;)

Mr Sniggle-Worth Onzo
Caldari Provisions
Caldari State
#130 - 2013-12-16 22:42:27 UTC  |  Edited by: Mr Sniggle-Worth Onzo
New game Mechanic ( you have my Titian pointed ) log in fast fast jump Crash ships saved..... Insert witty comment here ?

Limiting the number of people in local is nothing new ? As the game is now you can't have X number of people in a system w/o the node crashing so whats the difference by putting a number on it ?

Put the Number it takes to crash a node out in the open. So we all know what the Cap of players is on that system.


Tidi fights are fun. ? Node crashing alliances/corps at there will is ok ?



The system is already in place with every tidi fight there is a cap.

With every node crash guess what you have reached the player limit on that node/system.
James Amril-Kesh
Viziam
Amarr Empire
#131 - 2013-12-16 22:43:24 UTC
Dinsdale Pirannha wrote:
How about this:

If the node crashes, there is a percentage chance that any ships in that node are totally destroyed, and pilots podded.
Suddenly there is risk involved with crashing a node.

Of course, that means that people will just flood a system with T1 frigs if there are any supercaps in that node.

There is simply no solution for this blob-sec problem.
But it is just that, a blob-sec problem, that only affects a tiny percent of the overall subscription base.

And CCP should just stop throwing any resources at it, and focus those resources on the majority of the sub base.

This idea is just as bad as the OP.

Enjoying the rain today? ;)

Qweasdy
Dreddit
Test Alliance Please Ignore
#132 - 2013-12-16 22:49:20 UTC
He's trolling, stop feeding the trolls, 12 day old character, probably approaching the end of his 14 day free trial, decided not to subscribe or perhaps just a market spying alt about to expire etc

This is a terrible thread. As such, it's locked. - CCP Falcon

Gizznitt Malikite
Agony Unleashed
Agony Empire
#133 - 2013-12-16 22:49:25 UTC
Tyberius Franklin wrote:
Gizznitt Malikite wrote:

Truthfully, I think TiDi generally goes the wrong direction.

When time slows down in game, it gives pilots tons of RL time to organize more people and move them to the fight. This creates 1000v1000 battles, which is good publicity, but will always be abused and pushed to the point that the system breaks (node crashes, slow-ass game play, etc).

Instead, if it is possible, speed up time instead of slow it down. As time speeds up, the time to escalate a fight diminishes. This rewards players pre-positioned to escallate. This also is self-correcting, as faster time means more people die which means less pilots on the field which results in less demand on the server.

Unfortunately, this would be extremely non-trivial to optimize. Speeding up the server adds load on the server now with the intention that the server's load will lesson as things come to a conclusion faster. There also might be some exploits this creates (suddenly you can increase your fleets dps by bringing so many people). I don't know if this is at all possible, but it is a conter-intuitive approach that might result in better overall gameplay.

Time isn't slowing down as a cosmetic feature. It's actually the core mechanism by which Tidi works as advertised, allowing more real time for the server to keep up with the activity going on in the slowed game time. Sped up Tidi only makes sense if you thought pre-tidi node crashes and selective unplayability didn't happen fast enough. It's not so much a solution as an attempt to turn a few hundred people in a system into a node death sentence.


I'm aware of why TiDi was implemented: The server can hold only so much load, and slowing down game time eases the load on the server. This is brilliant, especially since it helped alleviate some of the major lag issues that plagued fleet battles. At the same time, it has a very adverse affect: Players will always push the server to maximum given the opportunity, and the slowing down of game time results in more opportunity for a response. This positive feedback loop results in more pilots which mean more load. The reason people keep proposing really stupid ideas like system limits, is to deny pilots the ability to overload the server. I'm not sure speeding up time is at all pragmatic, but it is a feedback system that limits the "time-frame" of most engagements.

Tippia
Sunshine and Lollipops
#134 - 2013-12-16 22:49:27 UTC  |  Edited by: Tippia
Mr Sniggle-Worth Onzo wrote:
Limiting the number of people in local is nothing new ?
No. It's been going on in Jita for a while since Jita isn't a system where that kind of mechanic breaks anything.

Quote:
As the game is now you can't have X number of people in a system w/o the node crashing so whats the difference by putting a number on it ?
It would be the introduction of a limitation that doesn't currently exist. The reason it doesn't currently exist is because such a limitation would break the game.

Quote:
Node crashing alliances/corps at there will is ok ?
Probably not, but if that ever happens, you can just report them to the GMs.

So, again:
• In what way is nullsec broken?
• How is it relevant where ships die?
• Do you have any kind of numbers to support your assertions about where ships die?
• How does giving large coalitions an unassailable stranglehold on all of null open up more of null?
• How is giving one fleet an unassailable advantage supposed to generate more fights and more willingness to fight?
• How do you propose to ensure that null is opened up without giving massive advantages to the established players?
• Why are you proposing a solution that lets fewer people play in null?
• What makes you think that anyone is benefitting from node crashes (where everyone loses) but won't benefit from poplocks (where one side automatically wins)?
James Amril-Kesh
Viziam
Amarr Empire
#135 - 2013-12-16 22:50:40 UTC
Mr Sniggle-Worth Onzo wrote:
New game Mechanic ( you have my Titian pointed ) log in fast fast jump Crash ships saved..... Insert witty comment here ?

Limiting the number of people in local is nothing new ? As the game is now you can't have X number of people in a system w/o the node crashing so whats the difference by putting a number on it ?

Put the Number it takes to crash a node out in the open. So we all know what the Cap of players is on that system.


Tidi fights are fun. ? Node crashing alliances/corps at there will is ok ?



The system is already in place with every tidi fight there is a cap.

With every node crash guess what you have reached the player limit on that node/system.

There IS no fixed number. It's not "x number of people in system crashes the node." It depends on SEVERAL factors.
Node crashing at will is not ok. Prove to me that it's happened.

Enjoying the rain today? ;)

Vald Tegor
Empyrean Guard
Tactical Narcotics Team
#136 - 2013-12-16 22:54:41 UTC
Mr Sniggle-Worth Onzo wrote:
Its simple really!

By limiting the number of people to a system ( like in Jita ) CCP remains in control, not the players.


No, it does not.

It gives players the control of denying other players access to a portion of the game. Which is the most broken thing you can do from a design perspective in a game that ensures even a -10 sec status player can enter a 1.0 security system.

Mr Sniggle-Worth Onzo wrote:

How many people in Null really do enjoy tidi fights ?

Goons systematically cancel their accounts during peace time and sign back up for sov wars. As you said yourself there are sufficient numbers of people showing up to these fights on a regular basis for it to become an issue for the server to handle. This in itself indicates the people participating must enjoy it to some degree. Otherwise they would not systematically spend their entertainment time participating, let alone be getting up in the middle of the night to do so.

Mr Sniggle-Worth Onzo wrote:
How many fights in the past year have ended in a node crashing ?

Are you asking us to do your homework for you or something?
Do you have any research to back up your claims of this being a problem, other than some news articles about servers crashing over the last couple fights following a new expansion to the game?

If you want to use this as any kind of backup to your argument, you're going to have to go and tally up every single fight over the last 12 months that involved over 1,000 pilots. Then figure out how many of those crashed. Because a simple number of how many crashed is completely irrelevant and you don't even have that ready to back up your claim. TIDI is also, for all intents and purposes, irrelevant. It facilitates the node handling a larger scale fight and playing out large scale strategic maneuvers that some people enjoy.


Mr Sniggle-Worth Onzo wrote:
copy paste the same crap for the 20th time

You need to stop that and actually participate in a conversation, otherwise we might as well go and have ISD lock this for spamming
Mr Sniggle-Worth Onzo
Caldari Provisions
Caldari State
#137 - 2013-12-16 22:55:05 UTC
People are afraid of change.

A new game mechanic such as limiting the number of people to a system would really just be stating in no or less terms what we already have in game but with out the crashes.


How many large fleet fights this year ended with a crash ?

Do you think that as the game moves forward that crashing a node will become less of a issue or more ?

I bet more


Th
Vald Tegor
Empyrean Guard
Tactical Narcotics Team
#138 - 2013-12-16 23:04:23 UTC
Mr Sniggle-Worth Onzo wrote:

How many large fleet fights this year ended with a crash ?

Do tell us.

You must obviously know. You keep bringing up how big of a problem it is. So you must have some concrete evidence of it being a significant issue.
Tyberius Franklin
Federal Navy Academy
Gallente Federation
#139 - 2013-12-16 23:04:55 UTC
Gizznitt Malikite wrote:
I'm aware of why TiDi was implemented: The server can hold only so much load, and slowing down game time eases the load on the server. This is brilliant, especially since it helped alleviate some of the major lag issues that plagued fleet battles. At the same time, it has a very adverse affect: Players will always push the server to maximum given the opportunity, and the slowing down of game time results in more opportunity for a response. This positive feedback loop results in more pilots which mean more load. The reason people keep proposing really stupid ideas like system limits, is to deny pilots the ability to overload the server. I'm not sure speeding up time is at all pragmatic, but it is a feedback system that limits the "time-frame" of most engagements.
This solution doesn't keep players from escalating load, but does have the node itself guarantee an increase in load. It's a gamble between a situation finishing naturally and the node killing itself trying to make that natural end come sooner. Depending on what would have been the real time length of the engagement, accelerated time may not be enough to prevent sufficient escalation responses from reaching the system affected. In the case of large battles where this would be invoked a successful conclusion is even less likely (recall we'd be dealing with scenarios where the server was going to have issues in normal time and are now making it worse). It's a bad solution because it replaces either of the systems we've had with a mechanism to increase performance degradation yet somehow expect these more degraded battles to conclude sooner if at all.
Vald Tegor
Empyrean Guard
Tactical Narcotics Team
#140 - 2013-12-16 23:13:46 UTC
Accelerated time would do hilarious things to piloting frigates that already do 4k+ m/s. Doubly so when you get to ideas like locking and blapping targets and guns that cycle several times per server tick in a 1Hz environment.