[Announce] FPS Networking Kit

I'm not sure this is abandoned. The developer has said that he is very busy with other stuff at the moment. However, I've essentially put our project on hold while we wait for this so it would be great if we could just get a definite 'yes, still working on this' from fholm. I know there's no ETA at the moment, but will we be seeing it released this side of Christmas? Thanks!

 

When it is ready?

 

i dont need to fps kit , i need network for fps games
u can creat a new kit only for network for fps game ?

 

Well I wouldn't say its abandoned as you can see he has been slowly doing minor updates to the webplayer so I like to think he is still working on it.

 

It's not abandoned, but due to the fact that I have a full time job, plus a five month old plus moving houses atm, I'm a bit pre-occupied

 

Glad to see you're still around

 

fholm, put the project on kick starter.. we can then fund you to do it!!

 

No, that's not going to happen sorry, it will be done when it's done (yes i am still working on it, but busy with real life + moving + wife + kid, etc.)

 

I need to say that I LOVE THE HUD

 

ok, it looks really good. Cant wait to see what you end up with..

 

will buy this in a heartbeat, even for $200

rest of the FPS kits in the store are laughable at best

 

As this thread has gotten some attention in the past week and I know there are a few of you waiting for this to be released, I figured it would be a good time to post a status updated. So, what has actually happened so far, how much is done and how far off is the release? Sit tight and I will give you all the info.

So, I first have to rewind a bit to the early planning stages of this networking kit. I was originally throwing around several different networking libraries as options for this: The built in, uLink, Photon and Lidgren. The absolutely first version of this kit actually used the built in networking in Unity, but it proved far to limiting to be of any use. I also evaluated Photon but had to dismiss it as it doesn't give you access to any of Unitys game world on the server. uLink was also cut out of the picture as it's far too expensive for most people to have access to. The choice was simple and I went with Lidgren, while it does take more work to use a library as basic as Lidgren it comes with several benefits: 1) It's free, 2) It can run inside of Unity and 3) It gives you (almost) full control over how and what data is sent.

Lidgren as definitely the best choice to begin working on this kit with, but I am sad to say that it has been proving difficult to mold it completely to my will and my idea of what this kit is supposed to do. Lidgren has a TON of features, message queues, channels, unconnected messages. Many of these features are of no use to 99% of all games, there are also a bunch of internal problems inside of Lidgren like bugs, quirky thread safety, not being able to run in single-thread mode, it plays a bit loose and wild with the garbage collector and several pieces of the core components have linear scaling causing them to slow down heavily under load. Lidgren still has a solid choice, but it just has too many problems and a source code which is incredibly complex.

About half way through implementing all features I realized to be able to get exactly what I wanted, I had to build my own UDP library. The library I developed, which is done and working has the current working title "Lynx". It has several features not available in any other networking library for Unity today - the most specific one being a concept called "most recent state data"-delivery which is a priority based delta-compressed state delivery mechanism borrowed from Tribes II and and Halo: Reach games. "Lynx" is completely game-agnostic, targeted at up to 256 players (no "MMO"-scale, but large enough I think) and performs really well. It's mainly targeted at fast moving and high precision games (like FPS:es) and has several features built in to support these types of games.

Anyway, "Lynx" will be released to the community hopefully within a month, maybe a bit longer. I have been working on the FPS Kit at the same time, building it on top of "Lynx" and pushing more features into the "Lynx" side of things. The FPS Kit is looking to be released about two months after the Lynx release. So a realistic time table is this:

Middle to end of Februari: Lynx
Middle to end of April: FPS Kit

This might be a bit longer then most people want to wait, but I'm pretty much refusing to compromise on quality here and want the code and functionality to hold the highest possible standard.

 

Good to hear! I really agree with "refusing to compromise on quality", I've waited quite a while and will keep waiting, for a solid FPS networking solution. The custom Lynx library sounds great as well. Thanks for keeping at it!

 

when is it coming out? can you make a quick video on youtube showing it please?

 

Ughh.. Learn to Read?


Aside from that, I am glad to hear on this update, there is truly nothing like this around. To purchase this and then create a first person game around it. I guess I will work on my models and environments in the mean time. Don't go to heavily into any 2d or 3d graphics. Keep it simple/clean , most people who purchase this will replace everything.

Looking forward to the release,

 

Just chiming in as another person who is very interested in this. Looking forward to seeing the end result of Lynx.

 

Absolutely cannot wait for this.

 

Lynx? Ready?

 

So is this ever coming?

 

Hope this is still happening, I'd rather like to use it.

 

Please fholm, this would bring so much happiness to so many!

 

Cant wait fir this to come out

 

any updates for this ?

 

Keeping the faith here with a bump.

 

Stumbled upon this thread and now i'm really looking forward to this coming, i'd happily pay up to 80 bucks for this even though i'm broke at the moment XD

 

Still would love to see this. Has anybody heard from fholm regarding the progress? I do hope he's still working on it.

 

January the 5th. This is the last day that Fholm has writing a message. You have faith. Good.

 

Another bump. Holding on to hope!

 

I really want this i hope he hasn't quit it ;-;
bump

 

It's still in the works, I've just had very little free time to work on my hobby projects in the... well, past year it is now. having a kid, buying a house, renovating said house, etc.

 

Any chance you would be able to release the compass?
The one you used in your demo

 

Hey, it will all be released as a pack or not at all. I am working on it (I've had to take a break for like 6-7 months at least), and i recently released the low level UDP library I am using in the kit, called UdpKit: http://forum.unity3d.com/threads/215161-Open-Source-UdpKit-a-networking-library-for-NET-Mono-Unity which is 100% open source/free

 

How are peoples feeling about being windows-only? Basically I am having huge issues with the CharacterController in the default Unity physics system, it is super slow and gives me troubles when trying to run 64 player maps. Just simulating the player input on the server eats almost 3ms per frame with 64 players.

Maybe some of you know this, but I have been working on integration for the new PhysX 3.3, and its currently working pretty well for Windows at the moment. Those of you interested in this kit, what are your feeling on that?

The other solution would be to forgo the client prediction movement, and instead use completely client based movement where the clients just report their position to the server.

 

I'd suggest you to include both solutions to the kit(client prediction movement and client based movement) and allow user to select the right one based on requirements, platform, etc. Another suggestion - please add voice chat to this project or at least documentation/sample how to add it.
P.S. Very interested in your kit. I'd pay up to $100 for it.

 

Doing both versions is not an option, because it would require me to integrate two different physics engine into the networking component, i am going to stick with the default Unity physics for now.

Yes this will be released, and the reason it is like delayed for a year is due to me having an amazing son and moving to a house, etc.

 

I would've happy with a windows based system to have client prediction included.

 

I should probably have been a bit clearer, I will do the unity physics version and keep the client prediction. If it ends up being so that it's to expensive server-wise, then that is up to each project using it. I have gotten the update time for 64 players down to ~1.2ms with the unity physics, which is good enough I think.

 

I like what I've heard so far. I'll be quite pleased to see how you do this when you release your system. I'm working on a first person game and I'm somewhat disappointed that there aren't enough resources for networking already.

You're doing good work, keep it up!

 

Wow, I just spotted this thread after replying to your physx 3.x integration thread. This networking solution and / or physx 3.x would completely transform what I'd be able to do with my game. I'd pay big $$$ for this stuff. Keep up the great work!

 

Just saw this thread for the first time, have been watching the physx 3.x thread. Want both. Windows only is fine by me, if that helps you get it out faster. You can always work on expanding to other platforms at a later date - and windows is the primary platform for FPS games. UFPS says they're working on multiplayer/networking for their next big release, so will have to see what they come up with (no news on what they intend to implement).

There are a lot of networking solution assets out already however ...
No physX yet though, if you know what I mean...

 

Please, please, please finish this. I am currently working on a multiplayer game and have tried to do most if not all this myself. While I have a working system, it is FAR from anywhere near what I would like it to be. I just found your rewinder asset on a whim and can't believe I missed it before.

I will gladly pay 100$ for something like this.

 

I am working on it, as fast as I can, but the reality is that right now I have something like 2-4 hours per week to spend on this, due to day job + running my own company on the side + family + renovating house. It will get done, as some point, but I can't give you a date or timeline, sorry.

 

Please, please, please finish this!!!!

 

I have always wondered, why unity networking has a bad reputation, because i was hoping to use it for an fps. In this kit for instance, you're using your own networking library instead of unity networking. What exactly would be different, if you had used the same fps authoritative techniques with unity networking (in a headless server). Is unity networking unrealiable? Does the server crash too often? does it use too much bandwidth? Is it too slow (why)? Are too many packets lost? Is unity aware of this issue, if yes, why don't they fix it?

This is more a general question regarding unity networking, but you probably have good reasons for using/creating UdpKit for fps game networking

 

The reason it has a bad reputation is because it exposes very few features towards the end user of the API, and since it's a "black box" inside unity there's nothing you can do to change/enhance the things which doesn't fit the game you are making. So when you need a specific feature, say "unreliable" events/RPC, you need to hack this into unitys networking solution, and it becomes a nightmare to use very fast. There's also more specific issues with things like state sync not having unreliable delta compression (which is almost required for fast paced games).

Yes, the low level socket library is UdpKit, and the higher level "game networking" solution built on-top of udpkit is called Bolt. But the client and server both still run inside of Unity. It just doesn't use their networking solution.

No.

No.

Sometimes, yes - as it doesn't allow you to implement unreliable delta compression which is paramount for fast paced games with a lot of players.

I have never tested the serializing performance/overhead of it, but I doubt it's slower then anything else.

This has nothing to do with the networking library itself, so no more then any other library

 

Ok, many thanks for your insights.

 

Hi fholm,
Just was to chime in to note my interest in this package too.
Also, as a random question, would this package support (or at least as a basis support the later development of) vehicles that can be driven/stood on by multiple players at once, without players jumping around, etc?

Thanks and all the best with this package!

 

Hi fholm
Is it possible to be added to an email list, for when the kit is released?

I can not remember if this is specified, but I think that mechanim support would be great

 

The first "part" of this Kit is scheduled for public beta release within a few days, it's a networking engine called Bolt which has support for many advanced features not found in any other networking solution for Unity at the moment, this includes things like automatic position + state sync, automatic animation sync (using mecanim), hitbox recording for lag-compensation, etc.

 

awesome