Lumberyard and Amazon GameLift and Twitch for Games on AWS | Hacker News Reader

Lumberyard and Amazon GameLift and Twitch for Games on AWS

aws.amazon.com

114 pointsjeffbarr42 comments

Note: This isn't Open Source, though the source code is available.

> Q. Is Lumberyard “open source”?

> No. We make the source code available to enable you to fully customize your game, but your rights are limited by the Lumberyard Service Terms. For example, you may not publicly release the Lumberyard engine source code, or use it to release your own game engine.

What's really interesting is that you can't use Lumberyard to communicate with game servers if they are hosted on a competing cloud platform; only AWS or server hardware you own:

> Q. Can my game use an alternate web service instead of AWS?

> No. If your game servers use a non-AWS alternate web service, we obviously don’t make any money, and it’s more difficult for us to support future development of Lumberyard. By “alternate web service” we mean any non-AWS web service that is similar to or can act as a replacement for Amazon EC2, Amazon Lambda, Amazon DynamoDB, Amazon RDS, Amazon S3, Amazon EBS, Amazon EC2 Container Service, or Amazon GameLift. You can use hardware you own and operate for your game servers.

Did you think they were giving the game engine away out of altruism? They obviously want to have someone create the next World of Warcraft tied to AWS.

That'll never happen and it's very unlikely to be their intention. Any game on a similar scale/scope as World of Warcraft doesn't need such atrocious license terms.

UE and Source both have much more reasonable license terms and smaller-scoped games will go with Unity unless they build their own engine.

My guess on what they are probably looking for is probably desperate indie or small-size studios that want something bigger than unity, they'll give a bunch of AWS credit to them in order to try lock them in. Even that is a long shot.

Edit: Had another read through this. Honestly I still don't see what the attraction is. There's bigger names out there with better engines and more favorable agreements.

Edit 2: Mixed up engines.

UE wants 5% of revenue full stop. (After the first $3k/quarter.) But even for a small shop like ours, with 7 employees, that's very expensive, PLUS we pay for servers on top of it.

Forced/licensed integration may not be for some people, but this isn't that expensive, comparatively, in the end.

If you are making a AAA game in UE4, Mark Rein fully expects you to visit him in his office and negotiate a custom license. That's the way UE1-3 operated. 5% is the default for the masses and for most projects it's much cheaper than a custom license would have cost.

Right, but we're not making AAA games. When comparing Unity Pro licenses to UE4 and others (now Amazon) for smaller game studios, UE4 is still compelling technologically. That said, it's licensing for a small studio like ours if far more expensive than Unity.

My point was basically that Amazon's pricing isn't outrageous, comparatively.

Their pricing is fine (it's very competitive as I mentioned below). Their license terms are not fine. Lock in is awful.

Lock-in!

Ok! But why is that bad? Unity3d is lock-in. Flash was lock-in. Hell, writing your own engine is lock-in, because rewriting that chunk of technology is so time consuming it could easily bankrupt a smaller game company.

Of course Amazon wants lock-in. It's what they're selling. But we all make decisions about who we want to be married to and to what degree we want an open marriage. All of us in the game (et al) industry are "locked in" in many ways.

I asked Jeff Barr if there were any plans to build an open source community around the Lumberyard code. Answer: "Yes, I believe so!"

Whoa.

https://twitter.com/jdub/status/696975443606396928

I'm guessing this is to take on what Microsoft has been doing Azure? https://mva.microsoft.com/en-us/training-courses/microsoft-a... that sort of thing? It's a great idea, because I think there's room for improvement in this space!

It's interesting to see what's happening in the cloud-backed gaming space. We're working on similar ideas at http://improbable.io but with more emphasis on persistent and unsharded worlds. I have the feeling that Lumberyard/Gamelift does "scaling" in the sense of "adding more shards", not "adding more compute power to the single unsharded world".

This seems to be a fork of an old version of CryEngine

Makes sense. There were rumors last year that Amazon spent $50-70 million to license CryEngine from the near-bankrupt Crytek.

http://kotaku.com/sources-amazon-spent-big-bucks-on-cryteks-...

Is it just me, or does $50m seem like a deal in terms of buying a top-end video game framework that will compete with the likes of unity and unreal?

Uh. It's late, so I might not be processing this right, but this might be the most insanely amazing thing I've seen Amazon do, or maybe any technology company. They just, from what I can tell:

a) Commoditized game engines

b) Released an open source game engine that can build AAA games

c) Are going to bootstrap an open source community around the game engine while supporting it

d) Are not forcing the use of any of their services to develop games with this engine

As someone who has been working in VR the last year or two I've been saying for a while that it's going to be important that we see a rise of open source in game engine technology the way we've seen it with the web before we really have arrived at having VR be the next platform. This has felt more like a dream than a real possibility: I didn't in my wildest dreams think we'd see an open source modern game engine anytime in the next 5 years since I assumed it'd have to be developed from scratch. And here we are having Amazon put this out, which might very well overturn the entire industry overnight.

As pointed out elsewhere it's not open source.

IT IS NOT OPEN SOURCE - though you do get read access to the source code. You're right that you aren't forced to use their services, but you are also NOT permitted to use other cloud providers. So, it's either AWS or your own hardware. Still, this does seem rather amazing. One thing you left out is that there are no revenue sharing requirements either. You're not even required to use this for game development - you can use it to build movies, training simulations etc.

Not sure what you mean by a?

It's not open source.

They are forcing you to either use their services or your own hardware if you want to use any sort of backend.

a) Arguably Epic Games or even Unity did this earlier

b) it is not open source

c) it is not open source

d) they are forcing you to use AWS for multiplayer servers and that kind of stuff, which makes sense because otherwise they would not make money

Might want to check the ToS also for Lumberyard as there's some interesting terms. https://twitter.com/XenosNS/status/697069050837475328

It's not open source, but games tend not to be anyway. It's certainly a very impressive development as Amazon are the most trusted "PaaS" vendor even if they're not the cheapest.

(have Amazon ever shut down any of the AWS services?)

SimpleDB is the only service I can think of that they've shut down since I've used the service

SimpleDB is still operating. In fact, I imagine it'll get way more popular alongside Lambda.

It is deprecated, and they've moved people away from it when possible, but last I heard it was still "on" for the folks who wouldn't move. They just won't let anyone else in.

Maybe Amazon is serious but I personally doubt it.

If you really want to be the next big engine you need large support teams. Dev rel people, tech support people, you need to organize conferences, constantly put up tutorials, evangelize, etc..

Maybe Amazon will do that but most tech companies that barf out an open source engine have generally not really committed to support them past the 2 or 3 engineers that started them and who will only continue to support them for 6 months until either they get bored or until their boss tells them to get back to their real job

Of course if Amazon is serious and they actually have a budget for 100s of people on the teams supporting this then great!

They might not put 100s on it (MVP, MVP, MVP), but they have traditionally done very good work in developer relations, technical support, documentation, example code, etc. This doesn't seem like a half-arsed project at all. I imagine it's a strategic move.

I was under the impression the engine itself is not theirs, i.e. not being developed by them internally. I.e. Crytek will work on the engine as usual, Amazon will take care of cloud stuff.

That's sort of the vibe I'm getting too - they're dipping their toes in, but not investing too much until they've gauged developer interest. If they'd been really committed to this, they'd have outright bought Crytek instead of just keeping them alive by licensing their engine.

Crytek is likely one of few companies that is nearly impossible to buy as it's privately owned family business. Almost like Valve that isn't most expensive company exist, but unlikely going to be sold ever.

Rumor has it they'd have gone bankrupt without this licensing deal - if that's true then I doubt they'd have said no to an acquisition offer if the price was reasonable.

Can I use GameLift client SDK for browser games? It doesn't look like it but can't find a clear answer.

(Also GameLift server SDK is Windows only, requires visual studio http://docs.aws.amazon.com/gamelift/latest/developerguide/se...)

"$1.50 per month for every 1000 daily active users."

I'm not sure what that means precisely. But any interpretation seems prohibitively expensive. (That's for GameLift)

That'd be $1500/mo for 1,000,000 daily active users, right? (Which doesn't include infrastructure costs.) That doesn't sound prohibitively expensive...

It sounds completely reasonable to me ...

You're off by 30x. It's $45,000 per month for 1 million DAU.

Here's an example from the horse's mouth.

"10,000 DAU (10,000 / 1000) * $1.50 * 30 = $450"

http://aws.amazon.com/gamelift/pricing/

Ouch, yes: It's $1.50 PER DAY per 1000 daily active users. (I think everyone is so used to AWS prices being per month...)

If you can't extract more that $1.50 a month from a 1000 active users then you should probably pack it in and try something else.

You also pay for the EC2 instances on top of that.

Which will be far more than $1.50/month to support that many users. I guess.

Can anyone with experience in game development give us a ballpark figure on how much this would cost if you were't tied to anything and just wanted to set up multiplayer infrastructure on your own?

It sounds cheap to me too, ends up costing you 9 cents per copy assuming all users who bought the game play multiplayer for 5 years.

The pricing is really good. The licensing terms are awful, but $1.50 for 1k daily active users is nothing.

Three use cases to consider.

Subscription-based games (eg. MMOs), averaging $11 / month. Worst case scenario (all your users are active daily), that's less than 0.02% of your income going to the engine.

Freemium/microtransaction games. Considering $1.50 is somewhere between 1/10th of a transaction to 10 microtransactions depending on the game, you need roughly an extra 1-100 microtransactions per 10k daily-active users to cover the cost. That's nothing.

Ad-based games... 1000 users should be worth more than $1.50. That's not gonna happen, why would you use this.

Reply on news.ycombinator.com