One of the most involved, and accurate news sources covering Virtual Worlds posted an interview with Philip Rosedale, former CEO of Linden Research, Inc. The Guardian interviewer David Smith talks to (an audible tired) the founder of Second Life about the current state of affairs, and the future of the Virtual World. To summarize the state of the network with key metrics:
- 18,000 servers in USA in three data centers
- 16 acres per server
- 462 square miles land mass of Second Life (1,100 sqr. km)
- 100 terabytes of user created content
- 55% traffic is from Europe
- 250 employees and profitable
- Development scale is equal to that of an operating system
- Theoretically the grid can hold 900,000 users
- Currently Second Life holds 60,000 consecutive users at most
- Central databases are constantly at 75%+ capacity
“Bigger than London”
PacificRim Exchange creates a sharp observation and visualization on the user density of Second Life, in response to Rosedale’s claim “it’s a lot bigger than London”. By using the provided metrics, PaxRimX shows the user density is barely 1% of Greater London, explaining the sometimes deserted feeling Second Life can create. Second Life contains 3.27 avatars per 16 square acres while London reaches a total 308.39 real people per 16 square acres.
No competition
Rosedale constantly points out why he believes there is no competition in the open virtual world field. It’s not because there is no market (he says Second Life is profitable), but because of the scale of the project (and compares it to building an operating system). Furthermore the required security of protecting their currency would be another obstacle to create serious competition. He compares their protective system to that in place at companies like PayPal after stating Second Life ’suffers’ attacks from “teens having fun” every day.
Building the 3D Web
Rosedale goes on to confirm once more he is building the 3D web. His goal is the standardization of content, but the prioritization right now is on making the software and client more accessible, and the initial Second Life experience will be their mission ’till it’s done’.
“Not only do I think that [we are building the 3D web], I think that it’ll become more pervasive, especially if we open it up and standardize it. I’m not necessarily saying that this one company will control all those servers. No. We’re working on systems where you can have the servers outside of our company. But, yeah, it will become the Web. I believe that what we’re working on right now will become a more common way of using the Internet to retrieve information essentially.”
He points out people are still getting used to the 3D web, just as it took the web years to be adopted by the current 30 and 40′ers, and skips across the ‘why’ by saying “we could laugh at youTube videos together”. To me that would be one of the examples why we would not see Second Life as a 3D web because you’d be doing things we can already do a lot better without using Second Life.
Source: Virtualworldsnews.com
Posted on: May 6, 2008 | Share on:
Abhishek Saxena on 
Sophie on 
IYan Writer
said on May 9th
Couple of obvious silly points:
* “18,000 servers in USA in three data centers”, yet “55% traffic is from Europe”. Wouldn’t some geo-spacing make more sense?
* “Theoretically the grid can hold 900,000 users” = 18,000 times 50 users per grid. However, as asset servers regularly drop at 60k concurrent users, this is patently false.
Hype is good in an early stage of a companies’ life - but LL is at the point where some measured rationality might serve them better.
IYan Writer
said on May 9th
subst(”per grid”, “per sim”)
Oops.
Rick van der Wal
said on May 9th
Good point on the asset servers - I had a hard time believing ll could hold 900.000-million users. Still, the fact SL can only hold 50 per sim is pethatic, given the fact 95% (guess) of SL is deserted at all times. Wouldn’t it make sense to create a structure that ports capacity to sims that need it? AW for instance can run 16,000 users on one ’sim’ - and their locations run on PIII computers.
From what I’ve gathered the big difference is the clientside information in active worlds, VS the serverside info on SL, which seems like a problem they won’t overcome any time in the near future.
The interesting part to me is, if they are having nearly 50k-60k consecutive users already, and that also seems to be the max numbers before the asset servers blows a fuse, LL apparently sees no reason to really scale up yet. This seems to be making Philips claim of a 3D web on an SL like grid (and LL’s ambitions with that) somewhat hollow. I’m not sure whether or not this relates to their plans of adding servers in Europe, but it seems unlike to be a solution to their scaling problem.
IYan Writer
said on May 9th
There are several problems with SL Grid being the 3D web:
a) problems with the client/server implementation: the more I hear of the way SL architecture works, the more dismayed I am. From single point of failure databases to downright idiotic way calling cards work - makes me wonder if they all slept through the CS classes at the uni. This is not rocket science - the mundane CS problems have been solved for years and decades. Instead of focusing on the are where they innovate, they stumble on the basic building blocks - web land store, back-end database.
b) the client/server architecture itself - given the nature of virtual worlds, I think P2P is a much better way of implementing them.
Oh well, we’ll see what Mr. M does.
IYan Writer
said on May 9th
*area - Argh, can’t seem to post a reply without a typo.
Rick van der Wal
said on May 9th
M’s focus will be on a sustainable a business platform like facebook is a business platform - not some new internet or a country :) I think it comes down to Philips influence on the course of SL will be.
If you have a good article that has some more details on the structure of SL I’d love to read it! Like to know some more about it without having to browse trough 700 pages of wiki material :p.