7 replies to this topic

[Training Instructor]

Maybe this idea is completely unfeasible, but what about having two servers, or even three, where the hosting shifts depending on what region is currently in prime time?

For example, 5pm-12am GMT the hosting server is in Euro, at which point it switches to the current North American server for their prime time, after which it switches to an Oceanic server for the Pacific rim players.

I know this would be a logistics challenge, what with their being no IT professionals outside the cities of Toronto (Current server host city) and Vancouver (PGI hq). Still though, it seems like this would improve gameplay experience for the players in these regions. People playing from different regions during these time periods could experience the joy of high-ping games.

My intent is not to split the player base, but to simply make the game more enjoyable for players outside of NA...because there are many many Euros and Pacific people who are paying customers, and it's not completely fair that they're not getting an optimal experience.

I'm not a network engineer, but I do work for an IT company that deals with a lot of international traffic, so I know there are solutions out there that would allow for PGI to have hosting servers outside of NA without a necessity for having a PGI employee on the ground in the hosting country.

[627]

This all depends on their server layout. First thing, there's not only "one" server in NA, that would be a server farm and with half a brain those would be virtual machines.

If you plan this right, you can distribute servers all over the world and have the same player base but this is very costly. furthermore, hosting a server in another region commonly means a partner who does it and with that you have all those "give me money too"-problems (like 20$ translate to 20Euro or UK guys get charged more than polish and russian).

And especially russian providers want a seperate player base so "their" players can't play international (this was done in many games afaik).

Edited by 627, 28 October 2013 - 10:56 PM.

[Training Instructor]

The costs of server distribution are continually falling though. They're already dealing with international funds transfers and conversions from over 47 European countries alone, not to mention Asia and Central/South America, so I fail to see how paying a hosting service in a foreign country would be much of an issue. Also, the most central hub for an Asian server would be Guam, as it has several submarine cables to the major Asian nations as well as Australia. Guam uses U.S. currency.

PGI has some pretty serious issues related to network performance, and their in-house solutions aren't getting the job done. Some of these problems are directly related to ping differential and home state rewind. The less ping differential you have, the less complicated the HSR calculations are, which leads to better performance.

My solution won't eliminate all the problems but it will reduce the complications related to gameplay performance. Gameplay performance is the most important metric, because it's what keeps players playing the game. A lot has been forgiven up to this point, but the game has been launched, and it's time it performed up to the standards of other international games in this genre.

[VeryVizzy]

Definitely agree. The cost factor is important but as you said server maintance costs have plummeted over the last 10 years where it really should be feasible for a company selling a product internationally that relies heavily on network performance.

The main objection I've heard to this is split player bases, which you've addressed with the servers switching for everyone based on prime-time time zones. If it were me I'd have it set up differently; I'd have it built into the matchmaker (hoooo boy) in the same way where it first tries to match players with the same Elo, then broadens it's search, you can do something similar with servers. It can ping them and the one with lowest is your 'priority' server but if no games become available within a set time frame, then it can look at all the others as well.

As a UK based player I'd love to try the game at ~50 ping rather than the 110-150 I have at the moment, and see what differences it makes.

[Training Instructor]

Playing mostly from Moscow with a 150 average ping here. PLayed from NYC with a 16-30 average ping last week, and I can say that it's definitely a different game experience.

I like your idea of building it into the matchmaker as well.

[JohnnyWayne]

Hsr was meant to counter the most urgent issue with high pings. Currently it does not though. I'm opposed to divide the playerbase in regions as long as there is no possibility to interact.

But keep one thing in mind: PGI is a small studio and solely the developement of hsr drove the release of content schedule far behind. Now think what happens if they play with the servers...

Edited by JohnnyWayne, 29 October 2013 - 03:45 AM.

[Training Instructor]

JohnnyWayne, on 29 October 2013 - 03:44 AM, said:

Hsr was meant to counter the most urgent issue with high pings. Currently it does not though. I'm opposed to divide the playerbase in regions as long as there is no possibility to interact.

But keep one thing in mind: PGI is a small studio and solely the developement of hsr drove the release of content schedule far behind. Now think what happens if they play with the servers...

I'm not saying divide the player base, I'm saying they should roll the primary server hosting to different regions in accordance with the prime-time of those regions. This way the players in those regions will be playing at the lowest possible ping during their primetime, and therefore will be having the best possible user experience.

I don't care about them being a small studio. They don't need to hire additional coders to do this, they would need to contract out to a company that specializes in server distribution and traffic routing, because their in-house team isn't all that great at it and this is obviously a task that would be far beyond them.

If they want to be innovative and actually do something that benefits the players, I see this as being a much more feasible route than trying to come up with some magic code that will conquer HSR when players with 100-300 ping are facing other players with 100-300 ping. The bigger the ping differential, the more difficult it is to accurately calculate HSR.

They're a small studio, I get that. It's no excuse for not biting the bullet and hiring outside expertise to make their game better so that they can both retain current players and attract new players. The video game industry is time sensitive. This would provide a more immediate solution while not taking any resources away from their in-house team, who could continue to work on the HSR coding to improve it.

Edited by Training Instructor, 29 October 2013 - 06:04 AM.

[JohnnyWayne]

It's by far not as easy as you pretend. You can't expect the same quality of service from different companys from aboard. It even becomes an international matter with contracts. That not enough, they also would have to write some pretty complex code that would take a ton of time.

Even with virtual Servers, you can not know how the game behaves if it is realtime migrated to another Server. What would happen to currently running games when a players ping increases? Is it even possible to proceed a live migration over these distances?
What about new Servers? Server stability?

It would be complex as hell, to set up and to manage. Personally, I prefer fixed hsr and new content.