These forums have been archived and are now read-only.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

 
  • Topic is locked indefinitely.
 

Dev Blog: Building a Balanced Universe

First post First post
Author
Berluth Luthian
Sebiestor Tribe
Minmatar Republic
#41 - 2013-12-03 16:35:21 UTC
So could people do node-fu like they do grid-fu and warp around to try and map a node in order to use that information for more...nefarious...purposes?
GeeShizzle MacCloud
#42 - 2013-12-03 16:36:20 UTC
Prism is it possible to put hard barriers for the pre-loader to stop connecting low and null sec systems together on one node? does this happen already?

when remapping systems from an overloaded node to another can a low and null system end up on the same node after tomorrow?
Starbuck05
Abiding Ormolus
#43 - 2013-12-03 16:37:14 UTC
my brain is hurting from trying to decypher all of that ... i got nothin

Just because i am blond does not make me stoopid !

Berluth Luthian
Sebiestor Tribe
Minmatar Republic
#44 - 2013-12-03 16:39:01 UTC
CCP Explorer wrote:
Aquila Sagitta wrote:
How is this being applied to wh systems?
It isn't. Solarsystems that are not geographically connected (WH systems) and non-solarsystem services are loadbalanced differently.


So does this mean that there are systems that are non-wh, non-solar systems? This further satisfies my theory that our new systems could be in the folds between known space in the dangerous pockets of spacetime like "The Vapor Sea".
Dom Roland
Butlerian Crusade
#45 - 2013-12-03 16:39:04 UTC
This is cool stuff and its good to hear about the technical details. I hope your next job is some performance mods so that tidi doesn't kick in as much!
Vincent Athena
Photosynth
#46 - 2013-12-03 16:39:24 UTC
This all looks good. It will be interesting to see what happens to TQ when you turn this on. (I'll set a long skill).

Where is dynamic re-mapping effort these days? Where are other efforts to reduce server load by improved algorithms, allowing a node to get the same results with less CPU?

Know a Frozen fan? Check this out

Frozen fanfiction

Vincent Athena
Photosynth
#47 - 2013-12-03 16:43:01 UTC
Berluth Luthian wrote:
CCP Explorer wrote:
Aquila Sagitta wrote:
How is this being applied to wh systems?
It isn't. Solarsystems that are not geographically connected (WH systems) and non-solarsystem services are loadbalanced differently.


So does this mean that there are systems that are non-wh, non-solar systems? This further satisfies my theory that our new systems could be in the folds between known space in the dangerous pockets of spacetime like "The Vapor Sea".

Non-solarsystem services are things like portrait distribution, character sheet, in-station stuff (CQ is not run on the same node as space) and so on.

Know a Frozen fan? Check this out

Frozen fanfiction

CCP Explorer
C C P
C C P Alliance
#48 - 2013-12-03 16:43:19 UTC
GeeShizzle MacCloud wrote:
mynnna wrote:
Grarr Dexx wrote:
Looks like a whole load of baloney! Us lowsec residents have been suffering from utterly ridiculous amounts of tidi, in sub-100 man fights or even moving thirty people ten gates. If you did do this rebalancing act, you sure seemed to have forgotten the systems between 0.5 and 0.0.


So I'm going to quote you so you can't edit out the part where you look silly, and then quote the devblog that explains why you look silly.

Quote:
We're hoping to have this code out by tomorrow Wednesday, December 4th.


also just saying that the phrase used being 'empire systems' would most likely refer to both high and lowsec systems, as im fairly certain all of the 4 main factions have lowsec areas. if you consider the area u fly in as being non-empire i suggest looking at the logos on the station toolbar next time you're docked up.
Yes, empire refers to high and low security space together.

Erlendur S. Thorsteinsson | Senior Development Director | EVE Online // CCP Games | @CCP_Explorer

Artimis Scout
Wormhole Cartography
#49 - 2013-12-03 16:44:17 UTC
That Dev Blog scared and confused me. I don't know if I should bow down before you or burn you guys for witchcraft.
CCP Explorer
C C P
C C P Alliance
#50 - 2013-12-03 16:44:29 UTC
Berluth Luthian wrote:
CCP Explorer wrote:
Aquila Sagitta wrote:
How is this being applied to wh systems?
It isn't. Solarsystems that are not geographically connected (WH systems) and non-solarsystem services are loadbalanced differently.
So does this mean that there are systems that are non-wh, non-solar systems?
Polaris is such a system.

Erlendur S. Thorsteinsson | Senior Development Director | EVE Online // CCP Games | @CCP_Explorer

CCP Prism X
C C P
C C P Alliance
#51 - 2013-12-03 16:52:00 UTC
GeeShizzle MacCloud wrote:
Prism is it possible to put hard barriers for the pre-loader to stop connecting low and null sec systems together on one node? does this happen already?

when remapping systems from an overloaded node to another can a low and null system end up on the same node after tomorrow?


Currently Nullsec is split off from Empire Space (High and Low). That's a change we did a long time ago and was covered in my first draft of this blog. That was however about three times the size of this one, without pictures. So yeah we could so that, but we don't.

There's probably a case to be made for their load fingerprint being different due to different player behaviour. But as it stands they get lumped in with high-sec.
Jessica Danikov
Network Danikov
#52 - 2013-12-03 17:04:05 UTC
Dev Blogs like this make me reconsider my current job... sounds like you guys have fun problems to work on.
BigSako
Aliastra
Gallente Federation
#53 - 2013-12-03 17:14:11 UTC
First of all: Great Dev Blog!

Second: I'm sure you probably heard that 1000 times now, but here it is:

Staging systems should be on separate nodes.
E.g: NCdot staging in GXK should be on a DEDICATED node (doesnt have to be a super reinforced node, but a dedicated node with nothing else on it), same for N3 Staging I-N, PL staging BPK, SOLAR Staging, DD Staging, Sendaya, Doril, etc...

You can
a) contact the alliance leaders or give them a way to identify their staging systems so they are NEVER ever in bad shape
b) do it with an algorithm that checks for average pilots docked in the last 24 hours


Right now I have seen our staging system linked together with another staging system and we constantly had 10% tidi in both systems (all we do is undock and warp to titan...) - so THAT is completely stupid.

A lot of fights happen in staging systems too, so tidi is to expected.

Third:
Not sure if that is possible, BUT every system with a reinforcement timer (IHUB, Station, POS) should be identified and CCP should have a team looking at these systems and sorting the nodes for possible fights.
E.g: Most timers in Immensea and Catch will cause fights and TIDI because THAT is where the fight is on at the moment.



Fourth:
Lowsec nodes are laggy. Very laggy. Probably because of Crime Watch. Not sure how to fix that, but I wanted to raise the issue again.
Ashterothi
The Order of Thelemic Ascension
Convocation of Empyreans
#54 - 2013-12-03 17:17:17 UTC
So does that mean that the live test event was success after all?
Altrue
Exploration Frontier inc
Tactical-Retreat
#55 - 2013-12-03 17:17:57 UTC
If it works, just... good job !!

Signature Tanking Best Tanking

[Ex-F] CEO - Eve-guides.fr

Ultimate Citadel Guide - 2016 EVE Career Chart

Bienator II
madmen of the skies
#56 - 2013-12-03 17:20:44 UTC
i would have tried to somehow solve it with voronoi maps but binary trees are probably easier to maintain since everybody knows them.

Long term Eve scalability will depend on the worst case scenario: Lots of people on one node. Or: lots of people moving between two nodes at the same time.

The first part of the dev blog basically tries to create a static load balancer which mitigates the second issue by splitting the bottle neck of a jump to two nodes (not placing neighboring systems on the same node). The first bottleneck however can only be truly solved by having real parallelism, e.g. running one system, if required, on multiple nodes. If this would be solved you wouldn't have to care about static load balancing since it could be all handled on demand.

how to fix eve: 1) remove ECM 2) rename dampeners to ECM 3) add new anti-drone ewar for caldari 4) give offgrid boosters ongrid combat value

Weaselior
GoonWaffe
Goonswarm Federation
#57 - 2013-12-03 17:28:48 UTC
CCP Prism X wrote:

Currently Nullsec is split off from Empire Space (High and Low). That's a change we did a long time ago and was covered in my first draft of this blog. That was however about three times the size of this one, without pictures. So yeah we could so that, but we don't.

What's the reasoning behind this?

Head of the Goonswarm Economic Warfare Cabal Pubbie Management and Exploitation Division.

BigSako
Aliastra
Gallente Federation
#58 - 2013-12-03 17:31:40 UTC
Weaselior wrote:
CCP Prism X wrote:

Currently Nullsec is split off from Empire Space (High and Low). That's a change we did a long time ago and was covered in my first draft of this blog. That was however about three times the size of this one, without pictures. So yeah we could so that, but we don't.

What's the reasoning behind this?


So huge fights in Nullsec don't affect ANY nodes in highsec which would potentially drive people out of the game?
And vice versa - you can't **** up a node in highsec to influence fights in nullsec.
jonnykefka
Adhocracy Incorporated
Adhocracy
#59 - 2013-12-03 17:32:06 UTC
1) This is one of the best devblogs we've had in months, and you should feel good about that. Communicative, clear, interesting, explains why we've been having issues and what you're doing about it in just enough detail that we know someone's thinking about it. Well done.

2) So let's take a few steps back to that big ol' live event just before rubicon launched that went so badly awry. One of the major issues was the tidi slowdowns from gathering a huge fleet in a central Empire system and then going many, many gate-jumps to nullsec. I know the LE team is going to find different solutions for that in the future, but broadly speaking how much can this help with fleet movement? Do you need more warning to reinforce things along the way or distribute the pathway in a particular way? As people have noted, there are going to be issues when big fleets (or multiple fleets) move through several system on the same node. They may be heading to a fleet-fight node, but the route there is going to be REALLY screwed sometimes.

3) Does your algorithm factor in jump bridges (the POS kind, obviously the titan kind can go anywhere)? That seems like it would alter which systems get jump-loads in some sections of nullsec.

4) And as for W-space, can you comment on where we get TiDi from? We do get it from time to time, and I presume your logs will show when that happens.

5) Final curiosity: Does your mapping algorithm for W-space account for static wormholes? I know that the w-space "regions", or in some cases constellations, are divided up by static connections. You might not know which C3 a given hole will connect to, but you know that it will always connect to a C3...
Abdiel Kavash
Deep Core Mining Inc.
Caldari State
#60 - 2013-12-03 17:33:22 UTC  |  Edited by: Abdiel Kavash
Weaselior wrote:
CCP Prism X wrote:

Currently Nullsec is split off from Empire Space (High and Low). That's a change we did a long time ago and was covered in my first draft of this blog. That was however about three times the size of this one, without pictures. So yeah we could so that, but we don't.

What's the reasoning behind this?

Crimewatch. Which is very expensive to run, can't be easily decoupled from the node code, and is not needed in nullsec systems at all. Therefore it's easy to split empire and nullsec onto separate nodes, and completely turn off crimewatch for nullsec nodes to save heaps of computing power.

Above based on a CCP post some time ago, I think it was from Veritas.