Thoughts about Metaverse Economics

Here you might discuss basically everything.

Moderator: SC Moderators

Post Reply
Jon Seattle
I need a hobby
I need a hobby
Posts: 648
Joined: Mon May 29, 2006 6:18 am

Thoughts about Metaverse Economics

Post by Jon Seattle »

The following are just a few thoughts about economic organization of second life. I am just beginning to think about this, but it seems relevant to NB’s longer term plans.

The Lindon’s primitive allocation on land area are second life’s rough way of allocating computer and communication resources to members and insuring limits on the complexity of any any visible scene. I say rough, because they are only an approximate measure of the resources consumed. In many cases texture downloading and rendering may be more expensive that rendering primitive shapes.

There are a number of projects that involve the development of open-source systems similar to second life. My guess is that one of these will become a feasible replacement in the next three years or so. The Croquet project (http://www.opencroquet.org/), for example, is rapidly approaching this point.
Once everyone can download and run a server, the economics of primitive and land allocation change radically. Primitives may still be a commodity in shared servers that function like directories, but when anyone can run their own server on the grid, primitives cease to be a scarce resource.

I really do not know how second life will respond to these changes. Especially if it becomes relatively inexpensive to set up competition to second life.

What is left, once the primitive problem is solved? Setting up a system for the exchange of intellectual property. If you teleport to someone else’s sim, do they get rights to everything you bring with you? Probably not. What are the rules and social conventions that govern IP rights? Should they set by US law?! What are the technological means for enforcing property rights across servers?

User avatar
Rubaiyat
Casual contributor
Casual contributor
Posts: 42
Joined: Tue Jun 06, 2006 2:37 am

thoughts

Post by Rubaiyat »

Jon,

I am familiar with croquet and see other potential issues with each person running their own server as much as I see benefits. I would however like to see some sort of private sandbox mode that could run without the network, kind of like Hiro's house in Snow Crash.

I think you also have an interesting point about dealing with IP issues as one crosses servers, but if I may point out the issue exists already with caches and routing. Email for example passes through multiple machines on its way to the destination. Granted it is cached on servers awaiting the pop client but it is still cached.

I am interested in issues of economics, but my interests run in a different direction. I am fascinated by the fiber of the metaverse and what can be built upon the 'givens.' I was recently on a plane thinking about how distance is depricated in SL and what that actually means and how to build upon it.

rs

User avatar
Gwyneth Llewelyn
Forum Wizard
Forum Wizard
Posts: 1183
Joined: Thu May 25, 2006 8:00 am
Contact:

Post by Gwyneth Llewelyn »

Linden Lab is in the 3D hosting business. While things like megabytes (either for disk space, or for bandwidth calculations) make sense for 2D hosting (anybody does have the idea that an image will take more space than words, so one understands that one should be charged more for many images than for a bit of text), for 3D hosting, other measures have to be employed that are simple, easily visualised (as 3D is such a visual medium!), and abstract enough to make some sense.

"Land" is a cool first approach, because iRL, we can envision easily that a bigger parcel should cost more than a smaller parcel. However, unlike what happens in RL, an empty parcel in a 3D virtual space does not take any resources (or only minimal ones). A small parcel crammed full of buildings does consume way more than a large one which is completely empty.

So "land" is not a good analogy. Actually, the very old users of SL remember that at the beginning, land was not parcelled. You paid for prims instead (one prim costed L$1 to rez), since the most important resource in SL are textures — they're the ones filling up the databases (object data is relatively small).

On average, every prim has 7 faces (yes, even "tortured prims"), which can mean up to 7 textures to download. It follows that tying prims to textures is a rather good metric: more prims mean more textures, and, although a tortured prim means more polygons (and more client-side rendering), the number of textures to be downloaded is the same. So, from a [i:286n9nl2]server[/i:286n9nl2] point of view, a prim takes always the same amount of data to transmit (which is negligible), and since it has an average of 7 textures tied to it, one can pretty well make an equivalence between prims and textures.

This is from the server's point of view; naturally, from the point of view of the client, one heavily-tortured prim "costs" much more to render (more polygons!) than a simple cube with blank faces. Still, for LL's core business (hosting 3D content), that is irrelevant — it has only relevance in the time to render a scene on the client, not on the effective cost of storing and transmitting the relevant data (prims and textures). Also, no server CPU is required (or server memory) to store complex prim data; a simple cube takes as much a share of memory and CPU than a very twisted prim with hundreds of polygons. This is the beauty of the prim-based system (unlike meshes, which are used almost exclusively on most other 3D platforms) — it actually follows a rather simple model, and one that ties in pretty nicely with 2D hosting (ie. calculating hosted space in MBytes and traffic generated in bandwidth).

What Linden Lab did was tie a "physical" measure of effective disk space (stored textures and prim data) and potential bandwidth (assuming an average viewing of a certain area — an empty sim does not consume outgoing bandwidth, from LL's PoV, while a busy sim will be always consuming bandwidth) to a RL measurement which makes more sense for us humans: land. So, tying land to available prims is a rather good compromise. One can naturally envision that bigger land means bigger buildings and thus more textures, so you should pay more for it.

A mechanism where prims are not tied to land is harder to enforce (as said, LL used that approach in the very beginnings of SL, where you did pay for rezzed prims, but not for land...). This is because that people can upload textures (and you could pay LL a monthly fee based on how many textures of yours are in their databaases) but they might never get used; on the other hand, someone who uploaded a very popular texture might get it being used by thousands of users. The alternative would be to charge for visible prims (as before), but don't tie it to land — but if LL abandoned that approach 3 years ago, it was certainly for a very good reason!.

So pay-per-texture is hard to enforce (and probably unfair — why should a popular texture be penalised?? This would discourage good artists/texturisers to produce them...), and a different mechanism would have to be employed. I think that the hardest bit to implement in this economy is how to differentiate use [i:286n9nl2]in potentia[/i:286n9nl2] and real use. You can have huge amounts of prims in your inventory; but if they're not placed in-world, they just consume tiny amounts of info (ie. just the prim descriptions, which are compact and small); once you place them in-world, they only consume bandwidth if someone sees it. In SL, truly, if a tree falls without anyone looking at it, it doesn't make a sound :)

Also, a highly primmy building with huge textures, that nobody sees, consumes nothing; a blank cube which is seen by 40+ avatars, 24/7, consumes an incredible amount of bandwidth! So clearly the method that we currently have is far from perfect — it works well for averages, but not for the extremes.

Eventually I could imagine than in the future the notion of dwell could be "reversed", ie. popular places, since they consume more bandwidth, would cost more. However, this is totally contrary to the spirit of LL, which rewards (even if indirectly) the places that attract more people, and thus encourages builders/texturers/event hosters to be creative in the ways they get people together. "Taxing" them would cripple the whole concept.

In conclusion: I don't think that a future SL (as long as it's run by LL!) will get us an alternative to the land-prim-texture model. Even if prims get replaced by meshes, there will be a similar land-mesh-texture model as the basis for "scarcity of resources". We might need to look towards other platforms to see different models; for instance, how will you pay for access to a OpenCroquet server? This is somehow a problem that the OpenCroquet promoters rarely discuss, since it's supposed to be a "free platform", with "access for free", and "someone else" paying for the costs of hosting, bandwidth, etc. I imagine that if someone would host a paying model of OpenCroquet, they'll do it the traditional way: pay per MByte and per traffic like on a "normal" 2D web server, and get rid of abstract concepts.

But I happen to [i:286n9nl2]like[/i:286n9nl2] the abstract concepts, since they make a whole lot of sense for this "metaverse thingy". Sure, they need some polishing — but what doesn't in SL? ;)

"I'm not building a game. I'm building a new country."
  -- Philip "Linden" Rosedale, interview to Wired, 2004-05-08

PGP Fingerprint: CE8A 6006 B611 850F 1275 72BA D93E AA3D C4B3 E1CB

Jon Seattle
I need a hobby
I need a hobby
Posts: 648
Joined: Mon May 29, 2006 6:18 am

More on the future..

Post by Jon Seattle »

Gwyn’s explanation of the current Lindon economy is really excellent; and yes it is a very clever arrangement. Nevertheless, an open source technology will offer competition and force the price of 3D hosting (and the cost of land!) to fall. A good analogy is what happened to the bit ISPs when the market opened up for smaller providers.

For educational projects, my guess is it will greatly lower the cost of hosting. The OpenCroquet project's talk may confuse some people as many Croquet developers are mainly in academia. For example, my tiny group at a US university runs a number servers that could host educational 3D projects at no additional cost. The servers run in any case, and this would make it easer to make good use of the resources.

Also as many people found the resources to run their own MOO and Mud servers (a kind of text-based VR: think SL with creative writing instead of 3D objects), I suspect that enterprising people will also find the resources to run their own 3D worlds.

Will the various systems with different ways of connecting different hosting platforms to form a single virtual world? Will there be technical standards that allow an avatar to walk from one to another? My guess is that there is little economic incentive for commercial hosting companies like Second Life to do this. They might do it anyway!

Democratic communities could decide to use open platforms as they become available. Two consequences:

1. A 3D community is composed of it’s people and content, not its hosting service.

2. A community (unlike an individual users) is more likely to want to run its own host, or at least lease a host system. That way they can take control of their own rules for resource allocation.

If we want to grow more democratic 3D communities in the future, we might find the technical landscape changed.

User avatar
Pelanor Eldrich
Veteran debater
Veteran debater
Posts: 246
Joined: Fri May 26, 2006 10:07 am

I would be interested in NeuCroquet

Post by Pelanor Eldrich »

It'd be a *bongload* of work, even with Brie, but I'd like to see the possibilty of our community running on Croquet servers, just to try it out.

Pelanor Eldrich
Principal - Eldrich Financial
User avatar
Aliasi Stonebender
I need a hobby
I need a hobby
Posts: 586
Joined: Fri May 26, 2006 12:58 pm

Post by Aliasi Stonebender »

If I have a problem with Croquet, it's that advocates have been trumpeting how an open-source virtual world will totally own the competition... yet, despite possessing a few interesting features, overall it's fairly primitive yet.

Truthfully, I'd see SL going open-source as the greater possibility.

User avatar
Pelanor Eldrich
Veteran debater
Veteran debater
Posts: 246
Joined: Fri May 26, 2006 10:07 am

Yeah croquet is very primitive at the moment...

Post by Pelanor Eldrich »

It's like SL 1.0 alpha or something. First big prob is a lack of easy to use dev tools.

Pelanor Eldrich
Principal - Eldrich Financial
Post Reply

Return to “General Discussion”