The Easter weekend saw large concurrency on the grid and things held up relatively well, albeit with some glitches. As relative newcomers to daily immersion in Second Life, we’re behind the eight-ball in realising how restrictive the central server model is on community events specifically and end user enjoyment more generally.
An effective 50 to 60 avatar limit on a event is obviously restrictive, both for the people who miss the event and the fifty who may be dealing with significant lag during the event. Islands are now considered a great value proposition because of the likelihood of reduced lag. Our own in-world launch was modest but managed to crash the server once we got above 45 avatars. The ABC Island / Four Corners launch had quite a number of people trying to access the island unsuccessfully due to it being ‘full’.
Critics of Second Life say the issue is the centralised server model and I’m yet to see any significant rebuttal of the claim. Potential competitors like Outback Online are touting the peer-to-peer (P2P) model as being the solution, claiming a 10 000 avatar population at an event as being feasible. If P2P is able to replicate the virtual world experience at the level Second Life has achieved whilst dramatically increasing concurrency of population, then the stampeded is likely to be significant. That said, I wouldn’t be alone in both hoping and assuming that with Linden Labs going the open source route, a P2P model may be in the platform’s future. Or at the very least a significant performance breakthrough that makes more than fifty people in a room a bearable experience.
What are your thoughts on the issue – if a competitor offered better performance would you pull up stumps and go elsewhere?
If there was the same mix of freedoms and significantly better performance I’d go like a shot. That said, a system that lets us build, script and animate freely, import textures etc. and make money if that’s our thing is quite a bit ask.
SL is moving towards a web-api rather than a centralised system as described at http://blog.secondlife.com/2007/03/04/malleable… – what that will actually mean is anyone’s guess, but it could well be 100’s of people in a sim with no problem.
If there was the same mix of freedoms and significantly better performance I’d go like a shot. That said, a system that lets us build, script and animate freely, import textures etc. and make money if that’s our thing is quite a bit ask.
SL is moving towards a web-api rather than a centralised system as described at http://blog.secondlife.com/2007/03/04/malleable-messages/ – what that will actually mean is anyone’s guess, but it could well be 100’s of people in a sim with no problem.
To me, the fact that the whole internet will be 3D in a matter of years is a complete certainty. Just as certain to me is that Second Life won’t be the application that will “make it” – there are so many more issues with it than “just” the unbearable lag when building up a screen: very poor usability for one, difficult and annoying navigability another.
I am however also happy about Second Life as people finally start thinking about the next stage of the internet, past the mere 2-dimensional display.
As soon as somebody comes up with a better 3D space than Second Life, I will be gone. I agree with Eloise though, we would need the same types of freedom in such a new space.
To me, the fact that the whole internet will be 3D in a matter of years is a complete certainty. Just as certain to me is that Second Life won’t be the application that will “make it” – there are so many more issues with it than “just” the unbearable lag when building up a screen: very poor usability for one, difficult and annoying navigability another.
I am however also happy about Second Life as people finally start thinking about the next stage of the internet, past the mere 2-dimensional display.
As soon as somebody comes up with a better 3D space than Second Life, I will be gone. I agree with Eloise though, we would need the same types of freedom in such a new space.
I agree with your points and i must say that in my opinion lag is really an issue, as well as usability.
I would also like to add that the search engine is also really rudimentary and it wouldn’t hurt to have a more efficient tool, with for example some social features like digg or del.icio.us.
I think that a P2P architecture plus web2.0 features could unleash the real power of online virtual worlds.
I agree with your points and i must say that in my opinion lag is really an issue, as well as usability.
I would also like to add that the search engine is also really rudimentary and it wouldn’t hurt to have a more efficient tool, with for example some social features like digg or del.icio.us.
I think that a P2P architecture plus web2.0 features could unleash the real power of online virtual worlds.
I would say that the limit of 50 people is pushing the class 4 server limits to their extremes.
Call 5 servers ,which are slightly more powerful, have more processing grunt and can handle about 100 people well. Try the sim a ‘virtual festival’ to see this in action for yourself.
I think it’s worth remembering how many processes and actions that take place on each sim – there are many differenct facets in controlling a virtual realm – every object feeds off one another, there is a lot of scripting / collission detection happening every moment you navigate the map.
Not only this, but rendering more than 100 highly primmed avatars on a screen is likely to do more damage to the framerate on your personal machine than on a class-5 sim…
The peer-to-peer model, I feel, may be a bit of a failure. Can you imagine trying to send an IM to a person say, 15 people away ? The lag will most likely be endless!
I feel that because SL has such an open model (and becoming more open, all the time), it’s allowed people to create content and to sell that content at a very cheap US $ rate, that it has helped developed an infrastructure that many competing realms would have a long way to go to play catchup before any alternatives present themselves to me…
I would say that the limit of 50 people is pushing the class 4 server limits to their extremes.
Call 5 servers ,which are slightly more powerful, have more processing grunt and can handle about 100 people well. Try the sim a ‘virtual festival’ to see this in action for yourself.
I think it’s worth remembering how many processes and actions that take place on each sim – there are many differenct facets in controlling a virtual realm – every object feeds off one another, there is a lot of scripting / collission detection happening every moment you navigate the map.
Not only this, but rendering more than 100 highly primmed avatars on a screen is likely to do more damage to the framerate on your personal machine than on a class-5 sim…
The peer-to-peer model, I feel, may be a bit of a failure. Can you imagine trying to send an IM to a person say, 15 people away ? The lag will most likely be endless!
I feel that because SL has such an open model (and becoming more open, all the time), it’s allowed people to create content and to sell that content at a very cheap US $ rate, that it has helped developed an infrastructure that many competing realms would have a long way to go to play catchup before any alternatives present themselves to me…