Ashes of Creation Alpha Two Server Meshing Technology Preview



Read more about Ashes of Creation ➜ https://ashesofcreation.mgn.tv

We are excited to give you an in-depth preview of our server meshing technology in Ashes of Creation!

Reminder that Ashes of Creation is an open and transparent development project. As such, you will see work-in-progress art, systems, and mechanics. Leave us your thoughts on our progress in the video comments!

💬 If you have any feedback on what was shown, please share it with us over on our Forums!
– For Server Meshing Technology Feedback https://forums.ashesofcreation.com/discussion/59653/feedback-request-alpha-two-server-meshing-technology-preview-shown-in-june-livestream/p1

Follow us on
👉 http://www.facebook.com/AshesofCreation
👉 http://www.Twitter.com/AshesofCreation
👉 http://www.Twitch.TV/AshesofCreation
👉 https://www.instagram.com/AshesofCreation

🗣 Join the discussion by connecting to our official discord http://www.Discord.gg/AshesofCreation

📺 Watch live the last Friday of every month at https://www.twitch.tv/ashesofcreation

source

28 thoughts on “Ashes of Creation Alpha Two Server Meshing Technology Preview”

  1. My friends, THANK YOU for tuning in to today’s development update. We are cloooooose to A2 coming!!🎉 Today’s showcase was the most technically oriented we’ve ever done, and it’s all in an effort to share with our glorious audience what exactly goes into making Ashes and what is being tested in A2. I know these kinds of updates aren’t as sexy as pure unadulterated gameplay, but they are important and I am excited to get your feedback on what you saw and heard today. We will be announcing the A2 start date at the end of this month, and we have begun sending out test invites to A1 users. Stay tuned for more updates ❤🎉

    Reply
  2. slow clap
    Got damn, this is good.
    Just when I thought that only Star Citizen was doing server meshing, these guys was like yup, were doing it too. Boys, we are on the verge of another great MMO boom and golden age with this new tech. GOT DAYUM this is going to be good.

    Reply
  3. In some games if I get kicked from server I cannot get back in because my server token is still active. I get a popup telling me I am still in the game obviously I am not. Then I have to wait up to 15minutes before I can log back in so the server registers I am not really there anymore. I get d/c'd from Soulmask sometimes but I've always been able to jump right back in so my d/c downtime has been the least I've known in any multiplayer I've played recently. Other games there is not quick way to get back in bunch of login screens so in essence they've slowed me way down in rejoining. I end up quitting these poorly run server games I miss out on content not only because I d/c'd but because I also have such a long time to reconnect that I died or missed all the action.

    Reply
  4. I don't quite understand the implementation. If actor A is on server 1 and actor B is on server 2, that means actor A/server 1 has a proxy repr. of actor B and vice versa for actor B, whereby the proxy means that a network connection is not required. However, if actor B (a proxy in actor A's eyes), moves to server 1 (the server which actor A has a network connection with), does that mean actor B disconnects from server 2 and connects to server 1 (destroying the proxy nature of actor B for actor A)?

    If this is the case then what problem do you actually solve? In your worst case scenario, one game server will still have to handle all player-controlled actors (if they all stand in one networked unit, for instance). The upside to this approach is that you spread the non-player controlled actors across multiple physical servers (assuming most of them don't translate between network boundaries), which is great. But why not just perform the split for non-player controlled actors and then still let all player-controlled actors remain on a single physical server and do away with the proxy related stuff for player-controlled actors altogether? It also seems to me that if you have actor A and actor B constantly jumping between network boundaries, then you will have a lot of execution time spent in creating/destroying networking connections and proxy reprs.

    If you really require the splitting of player-controlled actors across multiple game servers, then you ought to do away with the notion of a 'boundary'; it's entirely asymmetrical to the problem you're trying to solve–splitting client load between servers. Simply leverage the functionality you use for creating proxy reprs. of clients and load balance game clients automatically across game servers regardless of their position in the world. You can then re-balance the game clients across the game servers as you see fit and ensure that the game client does not have any power of what game server it is connected to (so we can't all stand in one networked unit and push a server's capacity).

    The other benefit to this is that your scale is then bottlenecked by your proxy repr. method not game servers as such. I imagine you simply have your game servers communicate client state in an internal network such that 50 game clients on server A becomes simply one network connection from server A to server B. This means that even though you have a vertical hierarchy of game servers across the game world (which is, essentially, sharding), they are all flattened by the proxy implementation.

    Reply
  5. They definitley bought this tech from Roberts Space Industries. Why change the combat system Sheeeeeeeeeeeeeeeesh. Just let tab target die like what is this mario. Nobody plays mario anymore cause the combat sucks. You AoC Apocalypse combat was EPIC AF. EPIC!!! THAT IN AN MMO!!!! EPIC. But this… Yawn been there played that.

    Reply
  6. I found this very fascinating to hear about the server meshing. It is frustrating to see so many games that can only show a small part of a land. I like the idea that there is finally a possibility of having a game world that can rival the surface area of an entire planet. Also, the solution of how to make players see each other across servers, more fascinating stuff.

    Reply
  7. Isn't the solution they are using at a high level.. basically what everyone is trying to do? The actual golden nuggets they've created is the optimization for each of these functions, hence why it's all custom built. I get they have to dumb stuff down for the general public, but this kind of info is staple in this problem domain; their solution and magic is not something they can explain in this video.

    Congrats on bringing the tech stack together!

    Reply
  8. I'm sorry, that presenter is way too boring, I can listen to star citizen talk about server meshing for hours on end.
    but I couldn't take 5 minutes of this one, why? you all sound nervous as hell, for one, and that presenter isn't doing anything to help alleviate that.
    The tech is amazing, I know because of what star citizen is doing with it.
    Glad I backed both SC and AoC.

    Reply
  9. microservices comes to videogames finally. meanwhile in web engineering we are wondering if they are even good at all lol. This seems like a really good approach to distributed data and parallel processing. It just makes sense, its a shame that the king of multiplayer engines doesnt support server meshing out of the box. I mean it would only make sense to have this on top of the world partition feature. Hopefully Epic is working on this exact problem and we dont have to lean on ripping the engine into pieces which Im sure now means thay are now locked to UE 5.x and wont be able to easily update the engine in the future.

    Reply
  10. As a star citizen backer-tester i am very interested in how Intrepid will handle server meshing. As Star Citizen is having a tremendous amount of issues with it.

    Reply
  11. As a programmer who's a newbie to game networking this was exactly the type of content i was looking for. I am not completed with the video yet, but the requirements of the game seem very very difficult.

    Reply
  12. But having everyone in the same gameworld can be a gameplay problem. An area might be totally overcrowded (central traders and banks) or completely empty and too large. Kind of hard to design an area if you dont know if there will be 3 players or 30,000 at that location..

    Reply
  13. Overjoyed you're taking on server meshing as an important endeavour. should be a world of fun and hoping for ultimately a smooth and fluid experience.

    Reply

Leave a Comment