Jump to content
  • Sign Up

It's looking like the server population reduction isn't working very well.


corwin.8356

Recommended Posts

This may be unique to our server (I doubt it) but reset night is a sh*tshow. Our entire guild has been queued out several times and since they won't just turn pips back on for EotM (or maybe even just treat it like a 5th zone) you start losing people soon after. Disappointing to say the least.

The implemented population reduction (7 per side, IIRC) won't fix the skill lag bomb issue since the problem is really based on how AWS EC2 VMs can't compensate very well for massive input flux. So when you have a map queue of a zerg prime and push it acts almost like a DDOS attack to the server. This problem is directly related to how EC2 VMs handle burst traffic. Researching the packet timestamps proves this out.

I really feel for the Devs dealing with this as you're trying to create a stopgap to a problem created by financial requirements from the parent company using the minimal tools at your disposal. Dedicated high end servers with large bandwidth would not have this issue. I'm guessing WvW is a loss-leader so the head office won't go for that.

So, what are the other options? Rolling EotM into the main WvW mix or even just having parallel rewards there is an option (if only to spread the population out). Hard capping population even lower but have overflow instances and aggregate scoring between them is another. Increasing skill cast times is possibly another, possibly better option. It is possible that an across the board 125-250msec increase to cast time could do a lot (I haven't setup something like this on my own EC2 servers; only tested burst vs latency). There are possibly other variations where we won't encounter this "teacup under a firehose" effect.

I guess I'm just asking that maybe revisit this and think a little bit out of the box. WvW is the closest thing GW2 has to an actual "Guild War". It's a shame to see it fall by the wayside.

Take care, be safe.

Link to comment
Share on other sites

@Helicity.3416 said:They should just take EVE online's time dilation solution.

Heavy load? Server slows down gameplay, ensuring every input comes out in the right order, just more slowly.

It's not perfect, but it's better than inputs being eaten and nothing working.Absolutely terrible for an action oriented game even if it works for Eve.

There is a reason people call it Excel Online.

Link to comment
Share on other sites

@Dawdler.8521 said:

@Helicity.3416 said:They should just take EVE online's time dilation solution.

Heavy load? Server slows down gameplay, ensuring every input comes out in the right order, just more slowly.

It's not perfect, but it's better than inputs being eaten and nothing working.Absolutely terrible for an action oriented game even if it works for Eve.

There is a reason people call it Excel Online.

Spoken like a true person who has never played it.

Link to comment
Share on other sites

Id love for them to use edge as a way to maintain control of the population disparity, and the lag. Id never be on any other map than Edge as its my favorite thus far~ But we do need more maps. Spread people out so that there are less crowding one map and reinforce that if one map is full... and you cant get in? Move to another.

Link to comment
Share on other sites

@"Mokk.2397" said:People brought this upon themselves by stacking servers and band wagoning instead of moving to lower populated servers . Adding 7 back won't make one bit of difference with a queue of 30 plus .

This is also why Alliances would never work and will never happen. The players would quickly find a way to stack specific "servers/teams" and it would ultimately be the same kitten-show it is now. More players than not want to be on the "winning" side whilst putting in minimal effort. The fact that most servers have massive blobs roaming the maps instead of 15-25 player squads is evidence of this.

Link to comment
Share on other sites

@Dawdler.8521 said:

@"Helicity.3416" said:They should just take EVE online's time dilation solution.

Heavy load? Server slows down gameplay, ensuring every input comes out in the right order, just more slowly.

It's not perfect, but it's better than inputs being eaten and nothing working.Absolutely terrible for an action oriented game even if it works for Eve.

There is a reason people call it Excel Online.

And still... it is a game that has managed to have an official dev-supported
GvG tournament
for 15 years and where single devs have been able to program additional supported environments for training purposes akin to the
10v10 sPvP project
or a proper OS / GH arena that actually works, so players can run their own events. They've even stepped in to
cooperate in running events
with the players.

It has also not only pushed the limits of its server architecture and performance-oriented design and/or programming year by year (pushing lag from the 50's, to 100's to 500's to 1000's even before time dilation and stuff that lets you fight at the 10.000's but fights takes multiple hours), but it has also been able to replace full backend systems dealing with physics, completely rewritten the graphics engine and done everything else that people asking for a GW2 remastered can only dream of while still having other running updates on their game. Oh, I forgot, they've also managed to release more games within their IP without abandoning, cheapening or ruining their flagship (including odd stuff like PC-console cooperative crossovers, VR and a mobile launch).

Look at this and weep at Living Story:https://en.wikipedia.org/wiki/Expansions_of_Eve_Online

To just add insult to injury, for the majority of its existance the studio was about the same size as ArenaNet with some 300 employees. They were just that much more productive, interested in making an MMO, committed to it and capable of delivery.

It must be tough for Action Anet when they are that much showed up by something called Excel Online.

Ed., just to underline it: The context here is productivity - there are companies out there who manage to do everything that Anet regard as difficult, do not manage to deliver on, will not attempt or things that players dream of from getting Dx12 support, to creating additional modes/spaces, getting community support and seeing updates to a broad variety of all the game's systems and modes piece by piece. That's not to say that everything in EVE is/was great and nothing in GW2 is good. However, the difference in productivity is nothing short of astronomical.

Link to comment
Share on other sites

To be a bit more constructive:

I'm not going to sing their praises for much longer, but I came to think about one thing more that is pretty important for the wider discussion here: CCP have also been pretty good at learning from and salvaging their failures. Their cancelled projects have had features put into the main games and tech developed have been reused for mutliple titles. For example, the scrapped second flagship and sold vampire IP lead to a complete rework of the player creator and some home-instance eye candy. That tech along with what they learnt from FPS was then also put into their more cinematic VR stuff and that what they developed for VR was then reused for additional simplified VR titles to just get something else, less ambitious, out there.

That's kind of similar to how I've been talking about what could be done with WvW where they could salvage what they have from their existing modes and submodes, experiment with what they have and build on that to run working systems (like servers and colors) side by side and being able to expand those into other things over time (like battlegroups and a risk-like mode). They don't necessarily have to create fully new modes but could just reshape existing modes and existing tech into something new or different enough. When they made the EotM arena for example, that was an example of good management like that but how old is that now? They've not taken what they may have learnt from that and iterated upon it into something else. Instead they let it rot away and it feels like they give up when something fails its test runs or as per the above even something that small seems exhaustive for them.

  • Just use the outnumbered system to fix population imbalance on scoring and nightcapping issues - it's literally db changes (raise trigger, raise impact)

  • Just put a better map on EotM and see what the players make of it (maybe it sways oppinion on rewards for it?).

  • If you have the EotM system with overflows and colors, tie a single level designer to the competetive team and have them experiment iteratively with maps and variations of them, involve the players and use the feedback/behaviour, it's a single employee who could bring an infinite amount of variation without exhausting effort, you could even involve players in it, draw from wc2/dota, your own GH JP's/racing, or do both inhouse and co-op.

  • If you make an Arena on EotM and the players still opt to play that on another spot on the same map, look at why and iterate

  • If you have timers and invulnurable objectives tech, see what you can do with that on a design level and if you could make a whole risk-like permanence mode of guild-held objectives (like EVE, like WoT, like NW) from it. Use your GH tech and see if you can tie some GH features into that. A good design could require a minimal programming effort

  • If you have systems that allow portalling to locations and setting up security measures, see what can be reused and ported over physics, maybe you could just use overflows, permanent invulnurabilities and toggle-able entires to let players use existing WvW structures to create WvW Guild Halls of sorts in which they could place an Arena. It may be crude but it could be functional. Right? You could toggle off the invul, let an opponent break a door, come in, repair the door and toggle invul back on. Now you have two GvG teams inside an objective and all they need is a reasonably sized and levelled middle space. They could even have SM, let one side kill the NPC's and fight over the claim ring. You would literally have a GvG arena by simply re-using the existing core tech in innovative ways. It wouldn't even need to involve changing structures, NPC's and whatnot, players could make it work anyway, even if some purpose-built changes over time would be welcome.

  • That's a perfect exaple of what I'm talking about - 3 differnt db values for invulnurable walls could essentially create three different WvW modes that could keep standard and replace both EotM and OS while just reskinning them to EBG's and letting players use the system. There could be standard WvW, there could be windowed sieging with overflows so different groups can cap and move maps to siege specific objectives, creating a risk-mode and there could be a mode with overflows for experimentation and toggled invulnurabilities to serve as WvW GH and GvG.

Etc, etc.

That's potentially whole new ambitious-looking modes by just re-using existing tech.

Ed., Perhaps I should give this it's own thread when I can get my own thumbs out and actually phrase and paragraph it properly.

Link to comment
Share on other sites

@Talindra.4958 said:i just realised i posted it in the wrong section.. my bad. there was a link yesterday mentioned about the server migration.. so that suppose to be posted there as part of a response to that post. totally my bad. :P

Well, it's not like we have many better things to talk about and it is at least related because stack transfers are also part of coping with the system (eg., full servers) and it impacts other systems (eg., full maps) and these things drive- and shape player behaviour (eg., two full maps and two empty maps, as people are not used to building community, because they can't and have gotten used to tack onto dwindling content). It all forms a total.

Link to comment
Share on other sites

Would be interesting if they lowered the population caps for each map, but introduced additional 1v1 (server versus server) maps for the match up. So you would have a red/green, red/blue, and green/blue map. They could be simple GvG maps with less PPT targets that focus more on promoting fights. Lag should be reduced with smaller map caps and players spread out over more maps.

Link to comment
Share on other sites

@subversiontwo.7501 said:

@Helicity.3416 said:They should just take EVE online's time dilation solution.

Heavy load? Server slows down gameplay, ensuring every input comes out in the right order, just more slowly.

It's not perfect, but it's better than inputs being eaten and nothing working.Absolutely terrible for an action oriented game even if it works for Eve.

There is a reason people call it Excel Online.

snip

All respect to EVE team and the EVE community and everything they've done to the game. But it's combat style is much more appropriate for time dilation. EVE is chess, GW is arcade with aimed skills.
Link to comment
Share on other sites

@Cuks.8241 said:

@"Helicity.3416" said:They should just take EVE online's time dilation solution.

Heavy load? Server slows down gameplay, ensuring every input comes out in the right order, just more slowly.

It's not perfect, but it's better than inputs being eaten and nothing working.Absolutely terrible for an action oriented game even if it works for Eve.

There is a reason people call it Excel Online.

snip

All respect to EVE team and the EVE community and everything they've done to the game. But it's combat style is much more appropriate for time dilation. EVE is chess, GW is arcade with aimed skills.

In all honesty, I never liked time dilation in EVE either and most of the time avoided it when I played (I did not play for very long after it was introduced and mostly tended to avoid the scale at which it appeared or only played up to the scale to where it may first begin to appear and not very deep into it).

However, nothing in my post that you qouted had anything to do with time dilation and mostly only regarded that while EVE is often seen as a slow but complicated game, its developer have managed to keep a point-on direction in terms of developing MMO, focused quite heavily on performance (more important for sleeker games), creation of "space" and "mode" (like many of the projects that seems to daunt, stunt and delay ArenaNet) and rewritten alot of backend systems (that seems to be somewhat of a pipe dream for GW2). That is the kind of stuff GW2 needs and is often brought up in discussion here though from Dx12 to Alliances to Arenas to UI mechanics to physics.

The only link to time dilation I made was pointing out that EVE didn't really start looking into time dilation until it had hit figures that are well beyond the needed performance for WvW. They worked alot more on the server and communication side of things before they began working on the design side of things. Also there are other things you can do to improve performance with design rather than time dilation (such as looking at ability/effect spam's impact on performance etc.). Funnily enough though, GW2 does seem to have some functionality akin to dilation built into how abilities get queued up and prioritized under heavy server load and lag, so the idea isn't that far fetched even for a game like GW2, but as said, I wouldn't put my attention there on GW2 if I was looking to make general improvements to WvW. A mode designed to never really scale beyond 300 on a given map shouldn't need such attention so there are plenty of other more effective options to explore if their 300-at-most scale mode is suffering from performance issues, whether we are talking server, client, design or shaping behaviour (eg., encouraging spread).

... And that the biggest difference does not lie in resources but in ArenaNet's somewhat infamous organisation, vision, direction and workflow. I've said it many times before but since HoT it has felt like they have built a company suited for adventure, action-adventure or standard RPG games and GW2 has taken said shape as a result We've gone from a vision for GW2 to a vision more akin to GW1 to where they'd seem more apt at making a Zelda game than a GW3. I think that has been easily as important as whatever market driven (committing to an idea of single-player dominating MMO content) or driven by preservation (keeping existing players or GW1 and GW2 remaining majority groups keen) factors.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...