Game Design, Programming and running a one-man games business…

When is it fair for the game servers to switch off?

cliffski

Here is a question for you. For how long after the release of a game do you think there should still be support for online features to that game?

I’m not talking MMO games here, obviously that should continue all the time any sort of subscription fee is being charged. I’m talking about games where you buy it once, and you own it, but part of the game is online. Like Guild Wars (although that’s all online) or, to a smaller extent: Gratuitous Space Battles.

Gamers are rightly angry when they suspect the servers for a game get shutdown just to encourage you to buy this years copy. EA are great at that, as I recall. The thing is, we can probably agree that if a game hasn’t sold a single copy in ten years, turning off the server is just fine. It’s silly to suggest otherwise. We also all know that most evil EULAs indicate the servers can be turned off when they fancy it, so it’s not a legal issue either.

What we lack, as an industry, is any sort of expectations or standards for this. If I buy Guild Wars today, and the server is turned off in 6 months time, is that just tough for me arriving late? Or is it understandable?

We could argue that as long as one person has bought the game in the last six months, the game should still be running, but what if that person bought the game in a steam sale or pay what you want deal for just $0.10. Still reasonable?

I guess to me, one benchmark would be profitability. If the server costs $100 a month to rent, and the game isn’t bringing in $100, it seems fair to axe it, although even then, what if 8 people bought it yesterday for $10 each?  It seems right now a long way off for me to worry about that. GSB made much more than $100 yesterday, let alone this month, and as a fraction of the server costs, it probably doesn’t cost $100 a month to run anyway. Also, the server is busy:

New challenges posted in last 24 hours: 48
Challenge victories in the last 24 hours: 143
new survival mode scores in the last 24 hours: 6
challengevictory: 170

This will not last forever though. When sounds reasonable to you as a gamer?

****Note: I can’t see the GSB server being shutdown deliberately before 2015 at this rate, so don’t panic :D****

 

27 thoughts on When is it fair for the game servers to switch off?

  1. It’s an interesting question.
    If I remember rightly, your next game is going to be more focussed on the online portion than gsb, and so if the game doesn’t sell well, how long should you be keeping it alive before pulling the plug?

    With GSB I can’t see this ever being a problem, at any time that you decide enough is enough, you could add a patch that includes a large amount of fleets and let people run these locally. Other games aren’t so lucky.

    I suppose it depends on how much of the game is online, or rather how many of the bullet-point features require the servers. If anyone bought a game touting online multiplayer, I think it reasonable to expect that this would be a feature for the life of the product. If, as GSB, it is tacked on to a largely offline product, I think you can get away with being a little more quick to kill it.

  2. To me, the “right” way to handle this is to open source the server (or at the very least release binaries) about a year or so before you plan to decommission it, give your users plenty of notice, and help the community get a server going if anyone is interested in doing so. Don’t say “nobody’s playing anymore so we’re pulling the plug”, put it more like “wow this game has been more popular than we imagined and we can’t believe people still want to play it after all these years so we’re giving away the server stuff so the community can run with it!”

  4. Honestly, I think that any time is right as long as you open the servers. What I mean is that if a game has a fanbase but running the servers is costly it could be wise to let them run them themselves. Of course, there are more things to consider, but that probably feels good for everyone.

    Also, tell everyone with a more than 2 months warning, Also, the server shutdown is a great time for the people who played the game long ago to return for some nostalgia-induced gaming.

  5. Private servers are always a popular feature for a game. If nothing else, it lets people run modded servers for custom content or restricted competitions. Releasing the binaries (preferably the source) for the server software at least a year before pulling the plug goes a long way to keep the userbase happy.

    I actually have two games whose servers were shut down before I even purchased the games. I didn’t know that at the time :-( One is just gone, no fallback functionality, and one continues to provide a limited set of functionality which is enough to keep most people happy.

  6. “How about never? Use the server for newer games, but keep the old service available.”

    That’s not practical in a whole range of circumstances. The software needs maintaining, it’s eating system resources (not much, but the games will add up over time), migrations to new physical servers or operating systems might render the old code incompatible (requiring even more work to keep it running), and so on. Plus the costs in keeping people around who know how to work with the system. It’s not as simple as “just leave it running”.

  7. If the multiplayer aspects is essential, then the only acceptable date for me is: “any time after the users are able to host servers themselves, if they so wish, and it makes finantial sense”.

    This means releasing at least one executable server, as well as allowing the game to “change the server IP” on the options menu.

    Extra kudos if the server sourcecode is also released.

    And before that? Never!

  8. I purchased a game recently that had the server shut down a week before I bought it. Of course, it was a used games (video store going out of business), but it got me to thinking. I agree with javoc that, when you do decide to shut down, encourage an open-source or private servers. That way, people who pick up your game in 5, 10 years after it was shut down can still play it. abandonia and the other abandoned games site points out, there is some real nostalgia that goes on with games.

    As for when to stop the official site? I’d say when it doesn’t become profitable, including the time spent to maintain it. Even with older games, you have to do patching of someone finds a security hole or you wrote it for the 32-bit operating systems and the rest of the world is moving to 128-bit architectures (I give it 15 years). So, give your time as money and time to host, shut it down when it doesn’t remain profitable.

  9. I would say the answer isn’t so much a specific timeframe, as much as a metric, ie, how long has it been since the last login, how many regular players has the plateau dropped to, how long since the last new player signed on, etc. I think the Free Allegiance population is a useful metric to start from, as they’ve got what, about 400 members in their community, and that seems to be enough to be totally self sufficient, from patching the game code, to operating game servers, or at least the last time I looked.

    So the first half is determining when to launch your exit strategy, and the second half is determining just what your exit strategy will consist of, ranging from the one extreme of opening up absolutely everything, from client to server, to the other extreme of just closing up shop completely, taking all your balls, and going home, and in between, you have your various licensing and transfer options.

  10. Oh an addendum:

    Your exit strategy should at the least consist of patching the client to allow user configurable servers, and releasing the server as a binary with associated runtime license. But that’s at a minimum. I agree with those who feel that opening the sources is far better.

  11. I like the public release of the server software idea. Let the community run with it. Make a final patch that allows the server ip to be modified in-game, and you’re good. Releasing the server source is even better.

  12. “I guess to me, one benchmark would be profitability. If the server costs $100 a month to rent, and the game isn’t bringing in $100, it seems fair to axe it, although even then, what if 8 people bought it yesterday for $10 each?”

    I think it’s reasonable to only maintain a server while it’s profitable. Once it reaches the point where it’s not profitable there should be some reasonable warning (3 or months maybe?) that the servers will be going down.

    If a game requires a server to be played then the game should immediately stop being sold. if the online component is optional, then all new purchasers should be warned that the online part will not continue to be available.

    Alternately, if the game has still has a devoted following, offering some sort of subscription to the online service seems reasonable. Using the example numbers above, with 1,000 active players, they’d only need to pay $0.10/month which is pretty hard to complain about.

  13. I will support the notion of: “Shut it down when it becomes unprofitable” plus “Release server binaries so people can run their own”.

    At the point where your game sells less than one copy per month, you might also think about releasing the source. That might give you good publicity and advertising, and you only give out something you cannot sell anymore. It shows your new customers that you care about them.

  14. As someone who’s hosting an online game via a builtin game sever I would say never… of to infinity… and by that I would say…
    When the game dies, just release the server as part of your game as well and let people start their own servers, if you’re game is dead what would you lose?
    You dont even need to release you game server code, just an implementation, evven better you can reduce your server to hold only a couple of games so its done in LAN only or small environment…

  15. Yes closing server after a few years seems right to me, as long as users can run some sort of server themselves, or have a convenient way to exchange challenges.

  17. It is reasonable to drop the server when it is no longer profitable from a business point of view.

    However, doesn’t doing so feel like a punch in the gut? Letting go of the thing you’ve created. Your time, your effort and love got poured into the game. If you only use a business metric to understand how well the game is doing you’ll miss something else which is really important: How much the players love your creation.
    How much they’re invested in the game you’ve created.

    When the server has become a financial burden, give the community the opportunity to take that burden off your hands.

  18. give notice on shutdown linked to profitability and for current users issue reductions on the game that nearest replaces it (keeping your market share)

  19. In this age of cloud services and VPS, it should be pretty simple to scale back a game that no longer needs the same level of service that it did until it’s cost effective again.

    I think that so long as some people are active then, if it’s possible and cost effecitve to downgrade the system by migrating to a lesser server even one sharing with other games, then it’s probably worth doing.

    Of course that gets harder with shard databases which by definition are independent so merging them becomes problematic, but it’s not impossible to overcome and if it’s thought about at design time then shouldn’t be a real problem.

    The problem with open sourcing your server software for smaller developers is that the same code will often be used again for later games if only in part, so it’s more of a risk imho for a small dev to release their server source.

  20. I agree with the crowd for the release the server idea. Releasing the source for the server can be problematic if you are re-using a large portion of it for a future game. I think a better approach would be to release it as an API, expose methods and include documentation so that they can interact with your server, mod it, but still keep the internal state a black-box. But once you are no longer using the server code for any game, no reason not to open source it.

  21. I think there should be a final patch. An end game plan for when the servers are eventually turned off that allows the game to be played offline, though perhaps with reduced functionality.

    How big is the fleet database? They’re just text files, right? Text files are highly compressible, so you can get a lot of information in a small patch.

    You should do a snapshot of the fleet database on the server and include that (or a sampling of it) in the patch, and also point the game to look at the offline snapshot for fleets rather than online, and that final patch would allow the game to be playable even decades in the future.

  22. When a game uses an online server (Multiplayer) or online activation, its fair
    to promise the customers to make the serversoftware, or a patch available in the
    future to “unlock” the game,

    Lets say by stating: “When our game will become non marketable in the
    future or has few active gamers remaining, we promise to offer our serversoftware
    (or unlock patch) to the public.”

    If there are fans remaining, they could setup a gameserver privately.

  23. In my opinion, the time it’s okay to shut down a server which a game is dependent upon is when:

    There’s some kind of fan-supported alternative. THIS ALSO HOLDS TRUE FOR MMOS AS FAR AS I’m CONCERNED.

    If a game completely bombs to the point where you can’t afford to keep the servers going at all for any length of time (like APB), that’s different. But frankly I don’t think games should be designed so that the company is effectively holding the game hostage in perpetuity. Any game with a significant popularity over time should have enough IT nerds and Linux gurus and whoever else you need to run the magic box to keep the game going.

    If the company already made their money, then piracy is literally a non-issue (unless the DRM-obsessed are afraid TIME TRAVELLING pirates might retro-steal their games from the future), and the potential to run the game should eventually be turned over to the fans.

    I haven’t played enough MMOs to have my heart broken a lot, but I have been through Tabula Rasa’s and Hellgate’s closings. At least I can still play Hellgate single-player.

  24. My suggestion would be that when the numbers start to turn so that you really wonder if sales will pay for the server over the next year, then announce that in a years time the server will likely be shut down. I’d then setup an account for donations, and tell server users that if in a year donations to the account equal the cost of running the server, then the server will be online for another year.

    This I believe solves the 2 key problems, 1) judging just how much people would miss the server if it was shut down, 2) provides the necessary money to run the server.

    You keep doing this on a rolling yearly basis, until the yearly donations don’t meet the server costs.

  25. Something relating to GSB galactic conflict in particular:
    Would it be possible for you to patch all the collected enemy fleets into the game if you were ever to shut down the galactic conflict server? That way GSB galactic conflict would be preserved for posterity :-)

  26. Yes. A local data base, which is simple enough so user can edit/expand it, could be a way to go post server (if the option of a releasing it is not possible)

  27. Well ive been thinking of hosting a private server for some games, such as World of Warcraft, Flyff, Ryl, or anything really. Well what does it actually take to learn how to Create the server, hosting it, fixing it when needed?

Comments are currently closed.