Uru redefined! Proposal list - add your own thoughts.

[Lontahv]

I don't know how difficult this would be in the context of Plasma architecture, but considering that Cyan didn't do it, it must be a bit of a challenge.

I think there's one reason why Cyan didn't do what you suggest:
Plasma's concept: Client is god

But say they did want to add some filters, there's another problem (this would be faced by us if we try to add filters to what can be done):
How do you tell the difference between a cone-kick and a cone being warped around by a player using Plasma's physics (without adding huge load on the network and server)?

[Chacal]

By writing an application firewall that runs on a different server and filters traffic in real-time.
If it can keep up with the traffic then it won't be a bottleneck.

You train it in two ways:
- analyze normal traffic and set it as baseline;
- analyze abnormal traffic and store a signature of such events.

You then add rules for processing traffic accordingly.

This isn't different from current NIDS we use in IT security.

The Plasma principle of "client is god" is wrong and outdated. The only reason I can think of for Cyan keeping this model is lack of resources.

So far we've been lucky. Uru never attracted mainstream attention. Let it become an interesting target and the hacking will begin.

[D'Lanor]

Lontahv is right. There is no way to differentiate between "good" and "bad" game messages. They are identical. If you go blocking specific types of message you most likely end up with a crippled game.

The Plasma principle of "client is god" is wrong and outdated. The only reason I can think of for Cyan keeping this model is lack of resources.

I think the reason is not wanting to rewrite all of the current content from the ground up. Everything is based on this communication model. It is deeply embedded in all of the client files and changing it would mean making a clean start without any content. Not to mention all of the tools that would become worthless.

So far we've been lucky. Uru never attracted mainstream attention. Let it become an interesting target and the hacking will begin.

There is not much to gain by hacking. Uru does not have a grinding factor. And simply showing off hacks will just get you banned.

btw, home run UU shards were hacked regularly but nobody was ever interested in the game. The hackers just wanted somewhere to send spam or DOS attacks from.

Another question is, how does security fit into open source? How do other open source projects cope with that when any security implementation can be grabbed from the source by those seeking to circumvent it?

[Nalates]

Judge me by my size, you do, hmm-m?
Pressing that button with the help of the force, well do I can, from down here. *reaches for a space somehwere below the button*

lol , that'd be a little excessive, but there are intelligent creatures other than humans in the D'niverse. Like the monkeys people from Channelwood, and there are probably others as well somewhere in the Ages..

I like that idea.

What works in MOUL, is available in MOUL (the model with animation skeleton) and what is in the D’niverse and restricted by IP rights may disallow the use of those creatures/people as AV’s.

I only remember two AV’s in Uru, people and bahro. I wonder who gets to be a bahro.

Would Shroomie be an AV? The pod age animals? Would any model with an animation skeleton be able to be used as an AV? – In other games it is a matter of there being animation for a given model and the ability for it, the model, to handle the basic game action that decides.

Would a Shroomie AV be restricted to swimming in Teledahn? Or would she show up in Ae’gura? We might need to mount the gun on the Great Stairs to keep Shroomies out.

While I like the idea, it will create some IC problems. I suspect Cyan will limit use of a Yeesha AV and may be others, like Shroomie and the bahro.

Nalates: You (and plenty of other people) seem to consistently and grossly underestimate what plasma can do. (as well as cherry-picking features that offer considerable high ground to SL) I'm not sure if you somehow forgot that KI images are taken locally and then uploaded to the server for storage in the vault, or if you noticed that those images were then downloaded from the vault by all other clients and displayed on in-game objects like imagers.

True, I may underestimate what Plazma can do. As to giving high ground to SL, no. But, I will continue to point out what they are doing and the technology they and OTHERS are advancing. Some of which these changes will likely mean we need to adopt.

Taking a picture client side and uploading it to the vault server does not constitute a major server load in my thinking. I did not forget that Uru/Plazma does that. However, the number of times a picture downloads compared to the number of times a shirt text downloads is a different thing. Large numbers of people downloading a picture at the time would be rare. Groups of people downloading a shirt texture at the same time would be much more common. Most VW’s deal with this issue and are advancing the technology.

The picture is all under game control. Shirt textures made in Photoshop/Gimp are not going to be. Some means of controlling those images (size, file type, etc.) will be needed.

So, while I hear you say adding shirts is easy and not much change is needed. I have my doubts.

Well... what you do is... you come up with a story about how if you hack our game, our backer will pull funding and it will be gone for everyone. This serves to turn your entire community into a vigilante organization that ostracizes anyone who might be considered a "hacker" because of course they wouldn't do that if they loved the game like a true fan.

This seems like simplistic view of the people that may come into MOOS. If only Uru fans are allowed, an IC story might work. If we make the game available and try to increase its popularity, we are going to have people that couldn’t care less about the game and just want to hack it fun and cause problems to show they can. SL calls them grief’ers. Most competitive games have punk-busters… We have a whole class of blackhat crackers that want nothing more than to take over a server just to say they did. Often when they can’t, they just like to mess things up for others. Telling them they could blow the game off the net by hacking would be like honey to a bear.

I hear lots of discussion that Uru servers only have to pass info through and are not the source of lag (server network meaning). I am just now getting people talking about servers catching and storing and distributing image files. Is the current system up to that as we add shirts, then may be pants, then may be… ? I walk in a room with 50 AV’s and those fifty have get my shirt and I have to get their shirts and we are already lagged (as in overall performance)… seems like that would add more.

Lontahv is right. There is no way to differentiate between "good" and "bad" game messages. They are identical. If you go blocking specific types of message you most likely end up with a crippled game.
[…snip…]
btw, home run UU shards were hacked regularly but nobody was ever interested in the game. The hackers just wanted somewhere to send spam or DOS attacks from.

Another question is, how does security fit into open source? How do other open source projects cope with that when any security implementation can be grabbed from the source by those seeking to circumvent it?

(emphasis is mine)

OMG! We got there…

Chacal's point about a firewall is on point. When one allows the client to make additions to the databases, add a image for a shirt, things get complex. If we allow fans to add models (tables, chairs, books), it gets even more complex. When crackers come in and try to chameleon their way in it gets even more complex. It usually requires human supervision over an automated firewall system.

D'Lanor’s question and my next logical questions will totally divert this thread from topic of proposal to implementation considerations. May be we need a new thread.

[Chacal]

I think the current message-based, client-server model can be kept, but some sanity-checking must be performed on the server side. It does not metter if it is an open-source project. The idea is that the server would only accept proper messages (hence the application firewall idea). This would narrow the surface attack greatly.
This would have little impact on the client side, so there would be no need of modifying content.
We could also rework the way the client does some things. For instance, instead, if you run from point A to point B, instead of sending a lot of "I'm running in this direction!!!" until you reach point B, the client could send "I ran to point B". It's a little less real-time, but a lot more efficient, and it solves the problem of clients using fly-mode.

[greendragoon]

Considering that a t-shirt image would be rather low-res (far lower that a KI image,) I would hope that it won't put too much strain on the servers. I mean, it's about the size of a windows desktop icon, 64x64 at the most.

As for more complex clothing models, I think you'll find that the crowd capable of creating new models is much smaller. What may work better is a system where new clothing models are submitted for inclusion directly on the server. Really, I suspect that most of the clothing added will be in connection to rewards for new ages.

Just my thought.

[Nadnerb]

We could also rework the way the client does some things. For instance, instead, if you run from point A to point B, instead of sending a lot of "I'm running in this direction!!!" until you reach point B, the client could send "I ran to point B". It's a little less real-time, but a lot more efficient, and it solves the problem of clients using fly-mode.

For one thing, both messages are already sent, and they're one message. This doesn't have any mitigating effect on flymode.

for proof, take a look at "physical.sdl". There, you will see what gets transferred whenever a client sends or receives an update on a physical object. (this includes avatars)

point3 position
quaternion orientation
vector3 linear
vector3 angular
plKey subworld

There isn't really a problem here... the transfer is already quite efficient.. the problem that occurred with flymode was that the hack that was first used to do it didn't consider such seemingly trivial things as sending physical updates over the network. When you move your avatar normally, your client sends updates every few seconds, depending on how often you change velocity.. (turn, etc)... but the first flymode, being the hack that it was, was a python script that used the warp method on the avatar physical several times a second while netforcing it. (this is why it looks jerky when you use flymode) The result was many of these messages a second, compared to 1 every few seconds under normal use. This is actually a good example of something that could be caught by an automated filter. Of course, after it became evident that this kind of thing was a problem, a "shard safe" version of flymode was created for use in the adminki, which didn't netforce all of the messages. This would be indistinguishable from normal use.

I'd like to point out to Nalates that when Chogon said:

take what you know about Prologue and Until Uru and throw it out the window. Because that is what we did when we totally re-wrote the servers for MOUL.

He said SERVERS. Yes, SERVERS. We have the MOUL client, we can read it's files. (remember drizzle, and the libplasma tools?) So please don't claim that we are contradicting Chogon when we talk about what Plasma can do. It's not true. Keep in mind that if you have a client, the functions supported by the server are implied by the functions that the client expects. So sure, the internal architecture may differ quite a bit from UU, but we know it still does the same job.

Well... what you do is... you come up with a story

This seems like simplistic view of the people that may come into MOOS. If only Uru fans are allowed, an IC story might work.

You seem to have missed the point of what I was saying. There was sarcasm in my previous message. If you can’t get the point of plain text and respond to it as though you knew exactly what I was thinking, then how can I expect to have a reasonable conversation with you? Oh yes, of course, that's right, the internet is not the best conduit for subtle dialogue. I realize that sort of thing doesn't work too well through text, so I'll explain myself. This "story about funding" was my impression of what Cyan already did. OOC. They said that they would lose their funding if their game was hacked in so many words during MOUL. They also instituted a censorship policy over on the MOUL forums that left many with a bad taste in their mouths. I was not seriously suggesting that we do what Cyan did, as I think it was a very bad idea, hence my sarcasm.

and also why I'm not sure it's a good one.

The picture is all under game control. Shirt textures made in Photoshop/Gimp are not going to be.

This is exactly why I thought that, while implementing custom clothing textures is quite possible, such an option was a bad idea. Imagine my surprise when you, the one who wanted to bring your "avatar and your stuff" from other virtual worldz into Uru, decided to toss this argument up as well.

Plazma

Okay, maybe I'm being irrational here.. but why do you do this? Because it looks like you're either really lazy and just don't care, or you are actively mocking Plasma. How clever and mature. Hey look, I can do it too! It's good that we can have these little talks without stooping to childish goading.

[Jojon]

I don't know just how hardcoded the two Uru user avatars (male and female), are into relevant Plasma builds, but as Aloys hinted; anybody who considers adding a third, with different physical properties (such as anybody with a different height than the standard ones, say a child), is going to have to be prepared to keep creating proportional versions of any custom animations introduced with new ages. (unless they mind some other animation, less adapted to the environment, playing...)

[Nadnerb]

Well, theoretically, it would be a new clothing item, rather than a completely new avatar. The avatar used by the engine is determined by your gender. Your gender is either Male, Female, Kodama, Engberg, DrWatson, Sutherland, Victor, Yeesha, Zandi, RandMiller, Cate, or Bahro. >.> But all of the above use only the Male or Female (or possibly Bahro) animations depending on which size of armature they use.

[Nalates]

It doesn't change any points being made here. Just sayin'.

And in the same light that I'm not trying to change any points already made. But take what you know about Prologue and Until Uru and throw it out the window. Because that is what we did when we totally re-wrote the servers for MOUL. Thanks, Chogon

(Reference)


OK… Nad… I’ll stand corrected on Plazma client side rewrite. For this point we will ignore the change from Havok to PhysX and all the comments about all that had to change for that. For the meaning I think that you mean to imply that would not be pertinent anyway.

If the servers were doing just pass-through and were not a source of lag (meaning server side only and network latency), what was it that they rewrote and why? Are the servers now trying to handle some of the client side load to improve performance? What changed from Prologue/UU to MOUL that improved things? What is it that still made MOUL laggy (overall performance)?

If poor performance was not on the server side, not in Plazma, not in PhysX, and not in DirectX/OpenGL then where was it coming from?

Greendragoon is right, I think, the images will likely be small (64x64x3=12k). Some VM’s use larger images for more detail, 128x128 (49k) and 512x512 (786k). With compression those numbers drastically shrink, but that depends on the individual image. Allowing the addition of user created images would require some changes to control them and assure only small images were uploaded. But, adding shirts as a reward and forcing creation of them to move into the age creation area is a way to mitigate those issues. However, it reduces the number of players that can add them to the game. I see the later choice as the best workable in the short term.

@ Nad… I did not respond to what I thought you were thinking… I responded to what you wrote, which was my point in the PM you are plagiarizing. Since there are no emotes or clues and I don’t really know you… and I try to avoid projecting my interpretations, I am not ashamed to have missed it.