Mūsų naujas serveris

UPDATE Rgs 19, 2013: Iš www.kamremake.com perkėlimas į naują priimančiąją baigtas! Jei turite kokių nors su svetainės klausimus praneškite mums. Bendruomenės TeamSpeak serveris kamts.eu yra taip pat prieinama.

Kadangi projektas prasidėjo KAM perdaryti išaugo daugiau nei mes tikėjomės. Dalyvių Online multiplayer skaičius nuolat piko metu per 120. Tiek liepą ir rugpjūtį mūsų svetainėje www.kamremake.com perkelta apie 2 terabaitų duomenų srautą (žr. diagramą į dešinę), ir tai tik vienas iš mūsų veidrodinių serverių, siuntimų dauguma atsirasti kitais veidrodžiais teikiamų gerbėjų. Kita vertus, mūsų masinės pagrindinio serverio kuris išlaiko žaidimų serverių sąrašas paslaugos beveik 2 mln individualius prašymus per mėnesį (vidutiniškai 40 per minutę) nuo maždaug 20.000 unikalių IP adresų (žr. statistiką žemiau).

Kai mes įsteigti šiuos serverius mes ne tikėtis, kad senas žaidimas, kaip riterių ir pirklių perdirbinys taps toks populiarus. Dabartiniai serveriai mes naudojame, yra bendrai pavaduojančius skirtos talpinimas mažų svetainių, ir dažnai nėra pakankama srauto kiekio gausime. Per pastaruosius kelerius metus mes turėjome tam tikrus klausimus su mūsų serveriais, dėl prastovų skaičių dėl perkrovimo arba problemos su paslaugų teikėju. Kita problema yra tai, kad mūsų pagrindinis serveris yra patalpinta Australijoje (Lewin savo gimtojoje šalyje), kuris yra priešingoje pasaulio pusėje į Europą, kur mūsų žaidėjų dauguma gyvena, o tai reiškia papildote serverio sąrašą žaidimas yra labai lėtas šių žaidėjų.

Dabar mes turime išspręsti šias problemas. Bendradarbiaujant su knightsandmerchants.net (Tarptautinė kam forumas ir ventiliatorius bendruomenė), mes nuomojamas virtualus privatus serveris (VPS) Nyderlanduose kuri dabar šeimininkai šias paslaugas:

  • KAM perdaryti multiplayer pagrindinio serverio
  • knightsandmerchants.net
  • www.kamremake.com
  • KAM bendruomenės TeamSpeak serveris naudojamas ne kamts.eu
  • Kitos paslaugos mes norime ateityje, pavyzdžiui, multiplayer sąskaitų ar pasaulio žaidimo pokalbiuose

Norėtume padėkoti Tomas (T * Ruzs anti-V!) Apie administratorius knightsandmerchants.net už jo sunkų darbą ir pagalbą organizuojant šį.

Kaip tai veikia tave?
Virtualus privatus serveris suteikia mums daugiau kontrolės, patikimumo ir išteklių, nei mūsų ankstesniuose serverių. Čia yra keletas pakeitimų, galite pastebėti:

  • Ši svetainė bus įkeliami greičiau ir parsisiųsti bus greičiau
  • Multiplayer serverio sąrašas žaidime bus įkeliami greičiau
  • Svetainėje ir masinės meistras serveris bus labiau patikimas (mažiau prastovų / blakės)
  • Bendrijos TeamSpeak serveris kamts.eu prieinamas
  • Daugiau paslaugų gali būti prieinama ateityje, pavyzdžiui, multiplayer sąskaitas, jei mes nuspręsime įgyvendinti, kad

Kaip galite padėti?
Serverio mums kainuos pinigus nuomai. Būtume labai dėkingi, jei galite padėti, siekiant padėti išlaikyti serverio veikimą ir remti būsimą plėtrą Kama Remake. Aplankykite mūsų dovanas puslapį:

Aukoti KAM Remake

Let us know what you think in the comments 🙂
Lewin ir Krom.

23 Responses to "Mūsų naujas serveris"

  1. Hey,
    Ich hätte da eine Idee zu den Multiplayerkonten, wie wäre es denn den “Spendern” ein kleines Visuelles Gadet, zur Verfügung zu stellen.
    z.B. ein goldener Rahmen im Forum oder irgendwie in der Spielelobby ein Hinweis darauf das man euch unterstützt hat.

    • We don’t speak German sorry, I read your comment with Google Translate. That’s an interesting idea, we’ll consider it if we do add multiplayer accounts.

  2. when will u fix AI code cause right now they cant build any forces they build all nesecaarry building collect weapon making 5 recruits but still 0 force pls mention about it

    • The standalone AI is not finished yet. Making the AI automatically decided where to place troops, how many to make, when to attack the player, etc. is very challenging from a programming perspective (it has to be able to work on any map under many different conditions). Hopefully in the next release the standalone AI will be more functional, but it’s not a simple feature to implement and is going to take a while before it’s completely finished.

      In the campaigns the AIs are given a script so they know where to place troops, and their village is already created at the start, so for now you can play the campaigns or co-operative multiplayer missions which do not require the standalone AI.

  3. Hi devs,

    I have some suggestions to further improve the website (a high traffic is a good sign for a download website – but a bad sign for an information website). I know you’ve already greatly improved it from initially 3MB for the index page – which was out of the question – to now 800kB. But there is still some room for improvement.

    1) you send out almost 200kB for the RGraph.js’s that are not needed unless for the statistics page. If you skip that, you save 20% of your website traffic (not downloads) per month! If you want to keep it, please minify it. That still gives you 10%.

    2) do you really need the jquery-ui? I didn’t see any spot where it’s used. Another 10%.

    3) the image which you inserted into your last article – make it jpg, not png. Another 10%.

    4) you’re already using google analytics. So there is no point in not using the jquery hosted by google. It has caching enabled, so the user saves some extra CPU power. Gives 20% and makes the page smoother.

    If you take into account a fair share of people with dynamic IPs, you’ll end up with a figure like 800 real users that join the multiplayer lobby per day. That is impressive!

    • eric theodore c.

      Hi Devs,

      you could also make your homepage open source (on github or so). people could make improvements and you have to care a little bit less about optimization and stuff.

      cheers

      • We’re using WordPress, so it’s not really practical to make it open source. What kind of improvements were you suggesting? People can always email us if they want to update the content on a part of the website.

    • This is site IS a download site, it is our USA mirror so the high bandwidth is not a bad thing. I would bet that 99% of the 2TB of traffic each month is from downloads not page views (we don’t get nearly enough page views to use that much traffic). Also, using lots of bandwidth isn’t a problem because we do not have a bandwidth limit at the moment and the server is on a very fast connection. So as long as it’s not too much for the user to download (which I don’t think it is), why bother optimising it? What benefits will it bring? I think we’d be better to spend that time developing the Remake.

      I’ll look into optimising/minifying the javascript files you suggested.

      JPGs look blurry for images with not many different colours like graphs. Also, those two images are 6.2KB and 6.4KB. No point optimising that 😉

      JQuery UI is used for lightbox images I believe.

      Pages on this site do load slowly because this server needs to query the master server to display the statistics on the right column. That effectively doubles the load time because our server has to download another page from the master server (which is in a different country) to get these statistics. Once we’re moved to the VPS it will be querying the same server so it will be almost instantaneous. This will greatly improve page load times.

      • Hi Lewin,

        it brings an improvement for low bandwith users. There’s still a great lot of people out there who only have like 1MBit/s connections. For example, I have 3MBit/s. Or think of all those mobile users with their limited connections. So each 100kB is another second of waiting. So it’s very appreciated if you do some optimizing.

        The lightbox you’re using is the colorbox – it requires jquery, not the ui. I haven’t checked all subpages, but the main ones don’t need the ui.

        Also, the loading time does not come from the masterserver. I’ve checked that too. You see the wordpress base there, because the provided code starts almost all requests at once. And a 200kB image takes longer to load than the 3kB masterserver query. I’ve tested it on a 100MBit line in europe. The 200kB image takes 900ms to load, the masterserver response on average 150ms. But that doesn’t even matter – the site doesn’t wait until it gets the response from the master server. As I remember the code, it places an arbitrary number at the beginning which is then asynchronously updated. The master server therefore does no delay at all. Did you change that?

        You’re right with the VPS, it will improve the loading time. But not because of the masterserver, but because your current server is way too slow. It takes already more between 1 and 2 second to run the wordpress program. You find it in the difference between the successful contact and the beginning of the content delivery. That’s way too much to just blame a initial masterserver query.

        But I understand that you only see it as a download site. I didn’t know that because I use it as a source for new information and also an overview over the current online play.

        Minifying is very easy, you can use an online minifyer like that: http://jscompress.com/
        For example, it shrinks the RGraph.common.core.js from 98kB down to 30kB. And the RGraph.line from 88kB to 40kB.

        • No, it downloads the actual values from the master server and sends them as part of the HTML of the page. The values are then updated periodically for the user with Javascript. From Australia it takes apache benchmark about 690ms to download that file from the master server. Our webserver for this site is currently hosted in America so it will be slightly less than that (USA is closer to Europe by cable than Australia), but I’d guess it’s still at least 400ms. So during that 400ms our site is stalled waiting for a response from the master server (it will not send the page to you until it has received a response). So by moving the master server to the same server as our website will reduce page load times by maybe 400ms, which is quite significant.

          I’m not sure our current server has speed issues, we’ve never really had any problems with it. You probably just find it slow because you’re used to local European websites, connecting to America would add a bit of extra time (living in Australia I know these things because a lot of websites are hosted overseas :P). And that extra 400ms or so from waiting for the master server probably makes it seem a lot slower than it really is.

          • I’ve tested now over several days on different times. Just the waiting delay is 1100ms from central europe.
            The overall time to query the master server is 160ms on average from here, so I guess your 400ms might be correct for the US. Which makes 700 ms of delay that you can’t blame the master server query for.

            Your host is using nginx, which is known for better performance than apache, so I conclude that it’s the hardware that speeds it down.

            But you’re right, we europeans are very likely to be spoilt by fast responding websites. We want to keep that 🙂
            Now let’s stop discussing, this already took longer than it would have taken to apply the fixes. I just wanted to give some ideas for the new server.

  4. hmm ye thats nice idea guys and to AI u can make more maps for single player ( transfer maps fromm multiplayer to single mission skirmish) and for AI u can make time push i mean for example at 3h they atack or when they reach army limit for example they need ( on different maps different amount of army to protect city and rest of it can atack nearby enemies ) cause now if they build army they dont do anything i tested it today i put game on 27 hour games when i left home and after it nothing change since 2hour mark they still keep massive amount of army without any action

  5. I would like to, again, congratulate you on what you are doing, and the unbelievable amount of time, money and love that you invest into this project. I will support you financially as soon as I get the bank transfer complete. Thanks a lot for what you are doing. Spasibo, rebiat 🙂

  6. I am an SEO expert. I would like you to know that changing servers is very negative for your google rank of http://www.kamremake.com. Keep that in mind. If you can I advise you to keep kamremake.com on the same server forever.

    • Keeping it on the same server forever is not practical. Also, I very much doubt that moving servers will affect our google rank. All we are doing is changing IP address, the domain and URLs will not change (the only difference google bot will notice next time they crawl the site is that our IP has changed). Everything that I’ve read on the subject says it should have no impact. Can you please explain why it would affect our rating and give some citations?

    • Expert, are you serious? Get out script kiddie, you got no idea about SEO!

  7. Lewin’s right. This is bull.

  8. Knights & Merchants forever!

  9. will there ever be a Australian dedicated server? to much lag on the other dedicated servers.

    • Good to see a fellow Australian! 🙂
      Unfortunately there aren’t many Australian players (I’ve had contact with maybe 5 over the years) so if we did have an Australian server it would be hard to find anyone else to play a game with. We could probably organise to join at a specific time and play but that’s hard to organise.

      I also don’t have the ability to host a permanent Australian server at the moment, I can host one occasionally from home but I’m often not here and I’d rather not leave my computer running 24/7.

      I’ve found that playing on American or European servers works ok, your actions are a bit delayed but that’s not such a big deal in an RTS game.

  10. Congrats, that’s all I can say, I always loved this game and even today I play it constantly. Do you plan on remakes of any other games
    Ex: heroes III?

    • We don’t plan to remake other games. We remade KaM because we both loved it and were disappointed by the bugs and limitations of the original game (broken multiplayer, no map editor, no support for custom missions, etc.) We don’t make money from this project, we just do it for the experience and because we enjoy it. Unless we felt the same way about another game, we’d have no reason to spend 4-5 years of our hobby time remaking it.

Palikite komentarą

Jūsų elektroninio pašto adresas nebus skelbiamas. Privalomi laukai yra pažymėti *

*