Posted - 2007.10.23 21:13:00 - [2]

 

Simply put, something has got to be done about overpopulation. Relying on players to use judgement and simply avoid places like Jita won't work. There are forces at work that invariably cause Jitas to occur and you can't fight them. But you can provide other outlets for them with a little creativity. It's common knowledge that Jita is the universe trade hub. This simple fact alone is enough to cripple any attempt at dissuasion. And when combined with every player's inherently selfish nature Jita was born. You could try and publicly announce that X and Y systems are now trade hubs. But it won't work. If everything isn't equally available for the same price peoples imply won't go there. It's like trying to herd cats. You can only do it with something they want for themselves. You can't count on their good nature, they don't have one.

Another take on the same solution: go to your semi-local trade hub and buy/sell something and it's delivered through an intra-station jump portal to Jita. And through the magic of insert/update statements the item appears in Jita/local hub.

Simply wishing that the problem will go away is an indication of stagnation. If you don't do something about it you're implying that EVE cannot grow further. As mentioned in that chronicle, and in the real world, stagnation is death to a business. So take a page from your own book or borrow one from someone else, but fix the problem.

Giving up will hurt CCP more than it's players.

Posted - 2007.10.24 10:26:00 - [4]

 

Originally by: CCP Lingorm

---

Edited by: CCP Lingorm on 24/10/2007 09:26:47
We are looking at a number of ways to deal with the market hub 'issue'.

We are approaching it from a number of different approaches.

Code improvement = What can be improved in our current code base to make it better/faster/more scalable

Hardware = What new hardware can we get that will allow us to scale our infrastructure even more (think Actual Cluster)

New Software = What new software/code can we deploy/write that will take care of this issue


You suggestion has been brought up internally but there are a couple of issues that it does not solve.

If you link a group of systems in a market network all you do is actually drive more people to those systems and empty the other markets. If there is 1 system in every region that links all the markets, then why put your stuff for sale anywhere else? OK, so to buy stuff you would only need to travel in Region but you would HAVE to travel as more goods would be sold at those hubs and less sold else where, so this would further move all the markets together. Not exactly what we want.

We have also considered 'enabling' Interbus inside a region so that you can have you modules delivered anywhere in the region, for a price and a short delay, but this kills courier contracts and logistics, as you get everything delivered with no risk .... again not a solution.

If you have any suggestions to overcome these problems then we are more than willing to listen as we are trying to solve the problem, but when you look at the suggested solutions closely you see they have side-effects that we do not want even more than the current problem.

I did suggest that we put in code that anyone that logs off in Jita is randomly moved to a 0.0 station, but this, while funny, also does not really solve the problem.

Yes there is a problem with Jita, not with how it works, there is nothing wrong with market hubs, but that we need to be able to scale it more.

Hope this lets you see that we ARE looking into the problem, but we want to Solve it in such a way that does not create more problems.

---

Low security systems dividing up the empires...

Seriously...think about it.

Currently everyone autopilots to Jita, and then does their shopping there.

If you could ONLY get from Caldari to Gallente space by going through 2 or 3 jumps of low sec then people would have to think twice about it.

If you could only get from Caldari to Amarr space via a 30 jump roundabout route or a few jumps of low sec again people would have to think about it.


Yes you would see four market hubs spring up (one for each empire) but each would be smaller than Jita and we would see real price differences appearing between the empires.

Posted - 2007.10.24 14:21:00 - [6]

 

Originally by: CCP Lingorm

---

If you link a group of systems in a market network all you do is actually drive more people to those systems and empty the other markets. If there is 1 system in every region that links all the markets, then why put your stuff for sale anywhere else? OK, so to buy stuff you would only need to travel in Region but you would HAVE to travel as more goods would be sold at those hubs and less sold else where, so this would further move all the markets together. Not exactly what we want.

---

Personally, I think this with the drawbacks would be better than what we currently have.

What's worse, distributed markets, or players being completely unable to play the game? That really is what it comes down to. Players trying to accomplish anything there spend literally 30 seconds to 30 minutes in utter frustration being unable to do anything in EVE. You might as well add a tetris game to EVE so we at least have something to do while we wait for the client to respond to our commands.

The place already has such a horrible stigma about it that people avoid it until they need a Snake implant, or want to sell one, then they go anyway. Now imagine it with 6000 players rather than 600. Now 60,000.

Originally by: CCP Lingorm

---

We have also considered 'enabling' Interbus inside a region so that you can have you modules delivered anywhere in the region, for a price and a short delay, but this kills courier contracts and logistics, as you get everything delivered with no risk .... again not a solution.

---

It is a solution. It's not perfect, but it would fix the more severe problem. What percentage of players actually use courier contracts? What percentage of players would rather have their courier contracts than the jita problem solved? Just because it's not perfect doesn't mean it should be discounted immediately. Which is more valuable, the logistical game mechanic or being able to play the game?

I think any internal discussions on what to do about jita-problems should be held in Jita on TQ on Sundays.

Originally by: CCP Lingorm

---

If you have any suggestions to overcome these problems then we are more than willing to listen as we are trying to solve the problem, but when you look at the suggested solutions closely you see they have side-effects that we do not want even more than the current problem.

---

What about what the players want? What about the players literally stuck in jita? What about the players that can only play on the weekends and have to go to Jita for something? Inflict jita on them enough and they'll just stop playing.

Originally by: CCP Lingorm

---

Yes there is a problem with Jita, not with how it works, there is nothing wrong with market hubs, but that we need to be able to scale it more.

---

It doesn't matter that you or I think there's nothing wrong with market hubs. They're going to coalesce unless you can offer all the same benefits (or more) through other means. So we either need to offer an equally appealing alternative or implement limitless scalability. I don't know about you, but I think a little creativity is easier than harnessing infinity.

Originally by: CCP Lingorm

---

Hope this lets you see that we ARE looking into the problem, but we want to Solve it in such a way that does not create more problems.

---

I'm thankful for that much but perception is reality. And until there's a huge fundamental change that dramatically improves the performance in jita to support limitless numbers or a change in game mechanics that renders jita moot, the perception that you're doing nothing will stand and your customers will get disgruntled. You may disagree - and I know I don't speak for everyone - but you never win by arguing with your customers. They can always win by simply ceasing to be your customers.

Posted - 2007.10.25 07:31:00 - [8]

 

Originally by: Miranda Duvall

---

Any solution that doesnt allow you to "scale out" (divide over a number of systems, as opposed to scale up: making the one system bigger/better /faster) will fail. If you would double the performance of the Jita hardware, the number of inhabitants will simply double as well, bringing the problem right back!

If you would allow interbus to transport stuff just 1 jump without delay or cost, Jita plus its surrounding systems would suddenly become one big hub, which DOES scale over a number of CPU's, while leaving the courier trade intact. If the new bigger Jita still has performance issues you could always connect a few more nearby ststems to it :)

You need to build in a system preventing IB transport over several jumps by doing it one jump at a time: one solution could be: when putting up a buy or sell order, allow it to be set in any station up to 1 jump away in the UI (IB will transport). Maybe you could demand some standing towards that particular stations owner or some other thing thats easely overcome

---

I also like the Interbus solution. The easiest thing IMHO would be to reform the contellation Jita is in so that everything 1 jump from it (that is already in The Forge) is in the new constellation, and then only activate Interbus in that constellation. RP it as Interbus only getting an operating license there if you must have a rational explanation for it.

There's nothing wrong with having a special market district, and if it works well you could implement it elsewhere.

Posted - 2007.10.25 20:43:00 - [10]

 

Edited by: Sicarius on 25/10/2007 20:44:28

Originally by: Verite Rendition

---

Originally by: Miranda Duvall

---

stuff

---

stuff

---

a better (imo) solution based on that would be to turn jita and say 12 other systems into it's own region (so it's more easily recognizeable), make every one of those 13 or so systems have jump gates to each other, so they are effectively 1 jump apart, interconnect it to as many other close regions (the forge, lonetrek and more) and spread that connection to different of the 13ish systems so that say your going from two different corners of lonetrek the shortest route drops you in two different systems. then crank up the base trade skill so everyone has remote buy/sell/edit remote trade stuff so they can mess with orders in any of the core 13ish systems, and even if people have to move around those 13ish systems youd still have spread the load from 1 system to 13 and not everybody is going to the same system to get that 1trillion trit sell order, and it's scalable so if you ever have a need to make it bigger to accomodate more people you just add more systems to that region, interconnect those systems to different systems in the nearby regions.

Posted - 2007.10.26 06:21:00 - [12]

 

Originally by: J'Mkarr Soban

---

There's a solution to this that is actually a side-effect of what could happen if factional warfare is introduced in a realistic manner, which is allowing the NPC navy to engage players of enemy aligned factions. Admittedly that will only split off into two main trading hubs (Amarr/Caldari and Gallente/Minmatar), but that's still a weight off Jita.

Even if people use alts simply to trade in Jita, they will get bored very quickly of having to transport their goods all the way down to friendly space for their main that (at least) a secondary trading hub will emerge. And once that happens, because Jita won't have everything anymore, people will start bolstering more local trading hubs, until a rough equilibrium is reached.

---

1) Everybody already has a Jita trading alt
2) There *are* secondary trading hubs. Rens, Oursalert, Amarr...

Posted - 2007.10.26 09:28:00 - [14]

 

Love the idea of several jumps of low sec between the regions. Could be justified when factional warfare turns up and would force the creation of more active markets in the different empires, plus it might make people more attached to their 'home' region and might encourage a bit more RP.

Love the idea of Interbus being active within 1 jump of the trade hub (or same constellation) as it allows the hub and spoke systems to all run on separate blades helping, a little, with 'scale out' until someone re-writes the system code to run across multiple systems/CPUs. I like the RP idea that Interbus gained the license from operating the service from the commerce commission/empires/CONCORD.

Posted - 2007.10.26 17:39:00 - [16]

 

CCP Gangleri was replying to my question. My concern was not strictly regarding Jita, but also lag induced by number of people in a system.

I am a professional software developer as well. I work on financial trading systems, so I have some knowledge of the real world market. I also concede that when real money is on the line, more resources and attention is applied and I cannot fault CCP if there is a fiscal limitation impeding a solution.

Now concerning Jita, I have a few ideas:
Idea #1: Create bypasses all around Jita. Any jump to Jita would be treated as an 'automatic' docking with the Jita station @ moon 4 (combine all the stations). That would allow multiple access points to the same pseudo-station from several systems. This distributes the 'local' size to the surrounding systems. Not a perfect solution, but it would alleviate the Jita lag.

Idea #2: Allow remote purchases to be delivered like CCP Lingorm suggested, but allow a real player to acquire the 'contract' first, and if no real player wants them, use an NPC delivery.

Idea #3: Basic retailing teaches Location, Location, Location. So just like office rentals have increasing rent based on popularity, place an ever increasing broker fee proportional to the local traffic. This would disuade massive congregation to a single node.

Posted - 2007.11.05 23:26:00 - [18]

 

I personally like the idea of making the market hubs and systems 1 jump from that system; essentially sanctuaries.

For example... Jita-Perimeter and the other systems around Jita no war zones... no suicide bombing... no offensive attacks of any kind.

Essentially to make it possible for it to jump into say Perimeter... you either have to offline or unfit any offensive weapons. Once in the system you also can't online any such weapons.

Obviously you'd have to move all ratting and missions and such out of such systems. Plus most of these systems are like .8 to 1.0 anyway so rats are basically non-existent in the systems.

The change could be made simple to be done in storyline. Take Jita for example... The Caldari being mega-corporation run... the whole policy change is run by earning money. The change of making such areas into sanctuaries to increase people into the area for trade and thusly make more profit. As many people... including myself avoid jita simply because of suicide bombers and such. With removing such problems from this area... a number of more people will move into the area for trade.

Alternatively you could make such systems in a sense 1.1 sec systems. Where you cannot engage war targets.

Posted - 2007.11.06 15:36:00 - [20]

 

Edited by: Andrue on 06/11/2007 15:39:44
I use player courier contracts to get stuff out of Jita and without exception it typically gets delivered within the hour. This is in the same region and four or five jumps away.

The only times I go to Jita are for high-value or rare items (usually the same thing) where the collateral needed to compensate me for a loss can is impractical (imagine trying to determine the collateral to cover a DG Invuln. Field ).

When I first decided to try using courier contracts I didn't think they would work because it seemed to me to be too hard to search for them. Obviously that isn't the problem - there must be quite a few players making money off them.

So my suggestions are:

1.Improve the experience of creating such contracts.
Add a button to the purchase dialogs that allows a one-click contract creation. Perhaps set up something similar to the shopping basket that web sites do. When you have finished your shopping you just select the basket and say "ship to...".

The idea here is actually just to advertise the option. When I've mentioned to other players they nearly all act surprised at the idea then amazed that it actually works. ISTM that if more players knew about these contracts it would help.

2.Improve the experience of finding such contracts.
Have the system act as a clearing house. I don't know to what extent this is really needed. It depends on what proportion of trade is already being carried by players and how this would change if (1) were implemented and courier usage accelerated. Does anyone here run these courier contracts? How easy it is?

3.High-value protection.
This one is tricky. What can you do to persuade me to use a player courier to move a 400mil, Dread Guristas Invulnerability Field? Most of my contracts have a 10% reward but 40mil is a helluva lot to pay even for avoiding Jita. Worse still DGIFs aren't that common. Even if I get my Isk back I could still be stuck if there are no more on the market.

Posted - 2007.11.07 08:24:00 - [22]

 

Edited by: Andrue on 07/11/2007 08:24:14

Originally by: Taji Seed

---

Tax allocation the same way as office rentals work.

systems with higher sales/contract volumes get higher transactional taxes.

weight it by system security status too.

---

That does suggest another idea - market stalls. In order to sell something you need to have a 'stall'. Limit their availability the same way research and manufacturing slots work.

This needs a lot of careful thought and planning but I thought I'd toss the beginnings of the idea into the ring

Posted - 2007.11.07 19:22:00 - [24]

 

as said there are different ways to tackle the lagmonster

implementing the new cluster stuff will take a long time
the question is if there will be steps in between? (one year with the current lag is a very bad future)

a) reduce the blobs
this is a game design issue and titan balancing (tripple titan nerf and the blobs where back)
I wonder if there should be a point system used like in the tournaments for grids/systems

b) handle blobs better
- simulating massive fleets with a stripped down client was suggested and confirmed as working on it (status?)
- server side event filtering (turrets and effects filtered on the server = server sends and processes less data 300*25kb/s=7.5MB/s that need to be generated and transmitted!!!)
- client side module grouping with server side optimizations (activate 1,2,3,4,5,6,7 vs. activate 1-7 and the resulting grouped processing of dmg, activation broadcasting)
- game design changes (good to see new bs with less slots and dmg bonus, carrier drone amount should be reduced to 5 but keep the dps the SAME - maybe disable fighter vs. npc like fighters vs. giant cans as the can use normal drones vs. scrambling npc and adjust the light, medium heavy drone 2 sec rof!!!)
- more nodes per cpu balanced by the os to reduce the affected lagging area

Posted - 2007.11.08 20:38:00 - [26]

 

Edited by: Andrue on 08/11/2007 20:44:32

Originally by: Alice Beeblebrox

---

Another way to further separate users out between multiple nodes for the same system would be to divide the space such that every cell "created" is handled in a 'thread' (which can be moved between nodes) and each such cell holds the users which can possibly interact between eachother.

for example, if 2 users are 500km or more apart (where they can't see each other), then they can be in different cells and possibly on 2 different physical/virtual servers.

---

These are called 'grids' and Eve has always had them. I don't know the details of how CCP handles then but tbh this solution is a very obvious one to a programmer (apart from the details, lol). I'm damn' sure CCP thought of this way back when they were designing Eve. After all this time I reckon they know all the tricks where grids are concerned.

Posted - 2007.11.26 22:45:00 - [28]

 

I'm sure these solutions have already been discussed, but here goes...

(1) Institute a hard occupancy cap on crowded systems like Jita. This is absolutely necessary. If the system can't handle the load of more occupants, don't allow people into the system. Seems like a no-brainer to me. There's no reason why the game should allow me to do something that will lag me out and prevent me from logging back in. Period. That's a horrible user experience, and there's no excuse for it.

Implement waiting queues. Know the server limits for a system, and when a system is overloaded, don't let people in until traffic dies down. It's better to be frustrated by a jump gate with a waiting queue, than to be frustrated by lagging out and being unable to play the game. Period.

(1A) A clumsy but possible solution to the aggravation caused by waiting queues would be to allow the jump gate to take you to any system that you would ordinarily be able to jump to from the impacted system, effectively allowing you to bypass the impacted system if it's choking a trade route. Autopilots would automatically use this feature. You can even make "skip jumps" only available after you've been in the queue for a certain amount of time, to make sure nobody's abusing the "skip jump" to cut time off of their voyages.

(2): When a player docks into a station, transfer them to a different node in the server cluster. There's no reason why the station interface has to be handled by the same server that's handling the star system itself. There's no reason why the same server even needs to service all pilots docked in the same station, since the station interface is only a gateway to the database anyway.

This would also solve the problem of not being able to log in, since logging into a "crowded" station would not have any impact on server load if station occupants were distributed randomly over multiple servers.

(2A) Now, if solution (2) is implemented and we have distributed station occupancy, it may be a problem when a pilot tries to undock into an overloaded system. Again, the solution is don't let pilots undock if the server can't handle it. Put them into a waiting queue, and let them undock when the server can handle it. You can even tell the pilot what position they are in line and how long is the approximate wait time. You may even give such "stuck" pilots priority over pilots who are waiting to jump into the system from another system. In the meantime, by all means let them use other station services while they're waiting. Just because someone is in a queue, it doesn't mean that they have to be kept from playing at all.

(2B) If a pilot is marooned in a station, unable to dock because of traffic, then give them an option to "skip jump" to a neighboring system instead of undocking into the impacted system.

So, to summarize:

(1) Cap the # of players per system by preventing jumps into impacted systems.
(1A) Reroute jumpers through impacted systems.
(2) Separate system and station servers.
(2A) Cap the # of players per system by preventing undocks into impacted systems.
(2B) Reroute undockers through impacted systems.

These solutions would solve the "Jita problem," since it would be impossible to occupy Jita with more pilots than the server can handle. While it may be difficult to get goods into or out of Jita, if what a pilot needs to fetch or deliver is important enough, they'll happily wait their turn. Otherwise they'll use another system. This solution would also encourage satellite markets to spring up in the systems around Jita, distributing the load appropriately across those systems.

Posted - 2007.11.27 20:43:00 - [30]

 

Originally by: CCP Garthagk

---

While a solution like your #1 might work for something like Jita without too many negative consequences, it would be disastrous for elsewhere. The only way I could see such a system working is to have it dictate the player cap based on the node load and how much of the node it has allocated, i.e. Jita is 100% of a node, it has 100% of the cap, but some XYZ-01 in 0.0 is going to be some small percent of a node, and have a player cap of maybe 30.

---

Lagging out and being unable to log in is disastrous. Having a wait queue is merely inconvenient.

Quote:

---

If you don't limit it that way and instead just limit to node population, you still run into denial of service issues. Players can clog up a system with so many docked alts that they deny any ability for the enemy to jump in and fight them. That or the enemy manages to get in and they deny the defenders access to their own system.

How would this affect logging in, too? Would you not be allowed to log in while the system was at capacity? Now we're talking about denying service to people who are paying customers, and the situation becomes even more sticky.

---

See point #2 above. If you distribute station users instead of clumping them into physical "station servers," then logging in will never be a problem until the entire cluster is overloaded, which of course would be a different problem entirely.

If distributed station users is not feasible given your architecture, then at the very least you can separate individual stations within a system from the system itself.

Quote:

---

I definitely like the idea of jump queues and more natural things like that to encourage the population to not congregate (think: traffic jams), but I don't think population caps are going to work in and of themselves.

---

Whatever measure you use to "cap" a system is fine, so long as there is a cap. My point is that the game should not let players onto a server if the server can't handle the load.