Around the Verse - Serialized Variables - Star Citizen Videos - MMORPG.com

System

Around the Verse - Serialized Variables - Star Citizen Videos - MMORPG.com

Chris and Erin Roberts host this week’s episode, which focuses on how serialized variables make networking a game of this size and scope possible. Plus, Los Angeles provides a studio update.

Read the full story here

Octagon7711

Wonder if we'll see SQ42 this year or a live demo?

MaxBacon

this cringy intro lol Roberts brothers: immersion and fidelity 

The network bit was pretty interesting, I like the technical side of things, it was always what I saw as the biggest challenge these devs have to deal with to make one MMO of a game with this much simulation, physics, AI, side of its graphics.

The network engineer bit:

"How do you network a game like Star Citizen? One of the first times I asked myself that question was shortly after I joined the company. I think my first reaction was one of sheer terror. Tears, panic..." - Clive Johnson, lead network programmer

That poor network engineer needs some bacon, good to know they are finally achieving some results in that front, network bind/unbind will also be a big one; this is that thing nobody sees but this is the literal skeleton of the PU.

IceAge

Citizen who ?

hfztt

I might have missed it, but that net code he talks about, is that the new net code that is going into 3.0, or is it for a later release?

MaxBacon

hfztt said:

I might have missed it, but that net code he talks about, is that the new net code that is going into 3.0, or is it for a later release?

Serialization will with partial bits, there was already some of it pushed into 2.6.1 earlier this year, it will be ongoing past 3.0.

rensta

ugh,,,,ugh?....ugh..... I'm drunk. but,im closer to a released game then this space whateverrrrrrrrrrr

donger56

Cool story so how about that release date? No? Ok, nothing to see here then....

Maurgrim

Yawn.

Tiamat64

I always get wary when a developer needs to explain why something is possible. At that point, you're in theory-land instead of something either proven to work in actual practice or something you actually can start working on now (else you wouldn't have to explain why it's possible because instead you'd just start doing it), and that's not a good thing.

I remember asking the developers of Warhammer Eternal Crusade if they really thought their vision for the game was possible too because of how massive the networking scale would be and they pointed me to a demo and bunch of other things that showed it was "possible" too. In the end, the final game didn't come close to the scale of what they originally intended, to the point where it's likely they never were even able to really try because it was so ludicrously out of reach. Hundreds of players vs hundreds of players in an open persistent world? They couldn't even get past small teams in instanced matches.

Saying something is "possible" is nice and all, but if things are at the point where you have to reach for things to prove that (as opposed to actually just starting to do it), you're definitely on tenuous ground as it is.

Octagon7711

Release date? I'd settle for a beta date.

gervaise1

Tiamat64 said:

I always get wary when a developer needs to explain why something is possible. At that point, you're in theory-land .

I think a part of the issue here is "filling airtime".

Backers want to have details of how the game is progressing - and even if a company has just posted extensive updates about what they have been doing there will be someone who didn't get the memo who asks how are things going. And when the update is pointed out you nearly always get someone who says: "yes but that was last week what has happened since then?"

And needless to say game development is like watching paint dry - well it would be but with modern fast drying paints I'm not sure the analogy holds anymore!

Fast forward and in the case of SC there is now a commitment to produce weekly updates - including weekly videos. And anyone who has done weekly team meetings will know that describing the weeks "hard results" quite often doesn't take that long! So what to do with the other 40 minutes. One of this week's topics was networking - and clearly this is not so much what they are going to do but what they have and what they continue to do. There have been other topics and for anyone interested in what the "glamourous world of what game development is like" they can discover that its like a lot of other jobs! Hard grind.

In the next few weeks people will be able to try out what has been happening during the last few months - when 3.0.0 launches - and networking will certainly be a feature. Backers first and then others through a "free fly" weekend.

MaxBacon

If we poke around we'll see the Backers have asked countless times for what is going on in the network front; so one can't say they are doing segments about things people have not asked for and this is a topic that wasn't brought up constantly in past times.

There have been very spread pieces of information here and there, and this is one more in-depth technical look at what was done, is being done and is planned to be done, ofc for something that isn't finished it will be much about what it is and will be but that's the nature of giving info on something from its concept to its implementation.

Wizardry

So they are saying the SIZE of a game has much to do with networking?Way out to lunch if that is the claim they are making.

YOUR visible and allotted bandwidth matters,quality of visuals,how many players,animations and overall static bandwidth needed per visual area.
SIZE of a game is almost completely pointless when talking about the NETWORKING.

PS...Did you see that FAKE smile the first dude after the Roberts squared bros?
Also sort fo turns me off when they just read from a piece of paper or off another screen.It sounds fake and really sounds more like rambling than someone who knows what is going on.

Bottom line is that first dude knows likely NOTHING of what is going on,he was merely told what to say and began to ramble on endlessly reading off a screen.

Kefo

Tiamat64 said:

I always get wary when a developer needs to explain why something is possible. At that point, you're in theory-land instead of something either proven to work in actual practice or something you actually can start working on now (else you wouldn't have to explain why it's possible because instead you'd just start doing it), and that's not a good thing.



I remember asking the developers of Warhammer Eternal Crusade if they really thought their vision for the game was possible too because of how massive the networking scale would be and they pointed me to a demo and bunch of other things that showed it was "possible" too.
In the end, the final game didn't come close to the scale of what they originally intended, to the point where it's likely they never were even able to really try because it was so ludicrously out of reach. Hundreds of players vs hundreds of players in an open persistent world? They couldn't even get past small teams in instanced matches.



Saying something is "possible" is nice and all, but if things are at the point where you have to reach for things to prove that (as opposed to actually just starting to do it), you're definitely on tenuous ground as it is.

You just described a lot of SC's promised features and gameplay lol

MaxBacon

Wizardry said:

So they are saying the SIZE of a game has much to do with networking? Way out to lunch if that is the claim they are making.

YOUR visible and allotted bandwidth matters,quality of visuals,how many players,animations and overall static bandwidth needed per visual area.
SIZE of a game is almost completely pointless when talking about the NETWORKING.

Isn't it blatantly obvious what that they mean by saying that, why the need for such rant? "when you are attempting to build a game as large and complex as SC" <<< duh

The size of things also clearly matters for the network, if you travel to a city in a planet, the scale of it's a factor in the network, if you have one solar system populated with tons of non-static NPCs you another have yet another factor in the network, the scale impacts the approach you have to take network-wise.

Even the big ships are a factor in the network, who is unaware only sees one asset, who is aware knows the big ships can go up to hundreds of networked elements. Size indeed matters in the network front.

hfztt

Well the most interesting thing the network guy says is that with this new system in place, their networking for handling a single area is about as optimal as they can make it. In short once this tech is fully done, we will know how many players per area/instance/scene the PU will support. Sure they might push a few more % out of it, but by and large, that will be it.

MaxBacon

hfztt said:

Well the most interesting thing the network guy says is that with this new system in place, their networking for handling a single area is about as optimal as they can make it. In short once this tech is fully done, we will know how many players per area/instance/scene the PU will support. Sure they might push a few more % out of it, but by and large, that will be it.

It's just one part of what is ongoing for the network of the game.

The network bind/unbind is another big one to stop sending updates to all clients and do a colossal cut in updates, right now all players are updated if one Enemy NPC AI spawned in the or one airlock was opened in the instance. We can easily figure how much the servers and also the clients get degraded performance by the crazy amount of updating it does currently.

So we see here, the serialization optimizing bandwidth and amount of updating as much as possible, then the network bind/unbind to be a big performance gain by sending those updates only to who needs them.

Once network bind/unbind is here that will be after 3.0, they have another biggie that is object container streaming, the object containers is a massive thing spread into many tasks that does compartmentalize the game-world and seems it's a key tech for having multiple servers supporting the game world.

I'd say we might see player increase after network bind/unbind is finished, and big improvements in that area only when there are multiple servers handling the players in the Stanton system.

adamlotus75

Serialised variables. WTF is this, coding class? Where is the game

rpmcmurphy

adamlotus75 said:

Where is the game?

This was bothering me so much that I decided to put the question to Spoon Boy & Neo...

• Spoon boy: Do not try and question the existence of the game, that's impossible. Instead... only try to realize the truth.
• Neo: What truth?
• Spoon boy: There is no game.
• Neo: There is no game?
• Spoon boy: Then you'll see, that it is not the existence of the game that matters, it is only the cash shop.

Herase

donger56 said:

Cool story so how about that release date? No? Ok, nothing to see here then....

Haha, of course there's nothing to see because the video content isn't for you.

The majority of you who posted have little to no interest in the game whatsoever, but are complaining about the video content that the actually backers and people interested asked for.

.....feel you guys need to find something better to do tbh  

hfztt

MaxBacon said:

hfztt said:

Well the most interesting thing the network guy says is that with this new system in place, their networking for handling a single area is about as optimal as they can make it. In short once this tech is fully done, we will know how many players per area/instance/scene the PU will support. Sure they might push a few more % out of it, but by and large, that will be it.

It's just one part of what is ongoing for the network of the game.

The network bind/unbind is another big one to stop sending updates to all clients and do a colossal cut in updates, right now all players are updated if one Enemy NPC AI spawned in the or one airlock was opened in the instance. We can easily figure how much the servers and also the clients get degraded performance by the crazy amount of updating it does currently.

So we see here, the serialization optimizing bandwidth and amount of updating as much as possible, then the network bind/unbind to be a big performance gain by sending those updates only to who needs them.

Once network bind/unbind is here that will be after 3.0, they have another biggie that is object container streaming, the object containers is a massive thing spread into many tasks that does compartmentalize the game-world and seems it's a key tech for having multiple servers supporting the game world.

I'd say we might see player increase after network bind/unbind is finished, and big improvements in that area only when there are multiple servers handling the players in the Stanton system.

My only objection here is: It is NOT what the guy in this video is saying, so unless you are telling me you know better than this guy, then whatever...

MaxBacon

hfztt said:

My only objection here is: It is NOT what the guy in this video is saying, so unless you are telling me you know better than this guy, then whatever...

You are wrong in your understanding of it.

As quoting relevant parts: "Serialized variables... it’s actually a critical system that's a core element of what we call Star Network."

"... Personally one of the best thing about serialized variables is that they’re one of the cornerstones that we’re using to build this big seamless persistent universe. Achieving it now is going to require spreading the load over multiple servers because it’s seamless we’re going to have to have these servers communicating with each other."

"So hopefully you’ll agree that serialized variables are not only a key part of the current technology but are a building block for future ones."

I'm not sure where you understand that this piece of tech is the entire network and not part of what's going on in the network front. You can also read this dev post that explains the importance of the work in object containers for game's network, side of network bind/unbind.

And you may check the global schedule for the object container tasks, the streaming the dev mentions is scheduled for 3.2 (that by the global schedule seems to be 2018). You also see Serialized Variables there listed, 5/15/17 to 9/13/17.

hfztt

MaxBacon said:

hfztt said:

My only objection here is: It is NOT what the guy in this video is saying, so unless you are telling me you know better than this guy, then whatever...

You are wrong in your understanding of it.

As quoting relevant parts: "Serialized variables... it’s actually a critical system that's a core element of what we call Star Network."

"... Personally one of the best thing about serialized variables is that they’re one of the cornerstones that we’re using to build this big seamless persistent universe. Achieving it now is going to require spreading the load over multiple servers because it’s seamless we’re going to have to have these servers communicating with each other."

"So hopefully you’ll agree that serialized variables are not only a key part of the current technology but are a building block for future ones."

I'm not sure where you understand that this piece of tech is the entire network and not part of what's going on in the network front. You can also read this dev post that explains the importance of the work in object containers for game's network, side of network bind/unbind.

And you may check the global schedule for the object container tasks, the streaming the dev mentions is scheduled for 3.2 (that by the global schedule seems to be 2018). You also see Serialized Variables there listed, 5/15/17 to 9/13/17.

I do understand exactly what he says. Funny thing is you do not. What he says is that when the containers are done, they cannot do more to improve the network within one area of the game. What you quote is how they can also use this tech to communicate across area borders, which, yeah, makes sense.

I understand what this guys says. I really do. You, quite obviously, cannot help but mix in some fanfic about what this means.

MaxBacon

hfztt said:

I do understand exactly what he says. Funny thing is you do not. What he says is that when the containers are done, they cannot do more to improve the network within one area of the game. What you quote is how they can also use this tech to communicate across area borders, which, yeah, makes sense.

I understand what this guys says. I really do. You, quite obviously, cannot help but mix in some fanfic about what this means.

I'm basing on what you said, that once they finish this tech we will know how many players per instance/area the PU will support, what is not the case. You have from network bind/unbind, object containers, and other work ongoing, that are pillars that directly impact the performance of the network per area/server, hence they are defining factors in player caps.

I got the impression you were thinking that once serialized variables tech is done the network is "pretty much it", but he says specifically the update messages are going to be as optimized as they’re going to get.