The thing we are unclear on is “shards”. Shards is a term from the world of MMOG’s. These games are replicated in numerous servers and players are divided among them. BUT… players in shard-A on server #1 cannot see or interact with players in shard-B on server #2. This greatly reduces the update data that has to be sent. In some cases chat has a global feature that reaches across all shards/servers enabling you to find friends whatever shard they are in. But, getting together often requires a relog.
Fall Breeze
Will the Lab go with shards as we know them now or go with a scaling tech that will allow a thousand plus people in a single virtual area? That is unclear from what we know now. But, we do know that as more avatars enter an area the data needed to keep them all updated on what every avatar is doing increases exponentially.
Think about it. If avatar # and #2 have a stream from one to the other updating each other on what they are doing that is 2 channels. Add a third and that measn #1 and #2 have to update #3 and #2 update #1 & #2, so that is now 4 channels. When #4 enters each has 8 channels so this channel count can be described as the number avatars squared = N2.
So, there are known physical, technical limits. We just have no idea on how the Lab is going to handle them. Shards, to limit how many per server have to be kept updated or… server scaling to handle the data load?
Links to pages below…
I wanted to continue the comment discussion on the CTRL-ALT viewer and SL rift support all week-end but didn’t find the time. But this article is also well suited for the message that I have in mind:
I think that the consumer version of the oculus rift is approaching fast. So far, Oculus only says Q1 2016 which could be 3 to 6 months. This is my pure speculation but I am convinced that Oculus aims at a release very early in Q1.
The Software Development Kit 1.0 for the consumer version will be released in December 2015: http://www.oculusvrnews.com/oculus-sdk-1-0-for-rift-and-mobile-coming-later-this-year/
Experience from earler SDK updates show, that software publishers need 2-4 weeks to update to a new SDK version. So the software would be ready in January 2016.
Also the first the first headset has already rolled off of the production line in September 2015. They are currently optimizing the process but the production line is more or less ready: http://www.roadtovr.com/the-first-oculus-rift-has-rolled-off-the-production-line/
If I am right and the Oculus Rift starts shipping in January 2016, then LL has +/- 12 weeks to get Project Sansar to Open Beta status. This is impossible. I really hope that they are not trying to rush things. They will need to make a lot of design decissions and they haven’t even opened up their Alpha to an application stage.
That’s also the reason why I was so surprised to read at your and David Rowe’s blog that the Lab is thinking of adding a rift support to SL. We are only weeks away from the consumer version of the Oculus Rift shipping and LL hasn’t even any concrete plans when this feature will be put into their development circle? No wonder that I was (kind of still am) thinking that LL doesn’t plan rift support for SL…
….ummmm…. Do you know that the Lab has an Oculus Rift Viewer out? And has for several months… You’ll find it on the Alternate Viewers page.
There is also the CtrlAltStudio Viewer built on Firestorm that one can get.
VR Headset support for SL is currently a viewer thing. That viewer support has already been added. The server side support for VR is mostly about performance. That is an ongoing battle for SL. Because supporting the VR level of performance is incompatible with the basic philosophy of keeping SL backward compatible Project Sansar was started. So, since how much VR support can be added to SL is limited, I don’t see them adding that much more, that is VR specific. I do expect they will do all they can to improve performance. Both Sansar and SL developers will learn from each other in that regard.
Sansar is in Alpha with a select group of people playing with and testing it. A couple of weekends ago the Lab hosted a VR meeting and while attendees had to sign a NDA the speculation is Sansar was demo’d at that meeting as a focus of what was coming for VR. Because there was and is a login page for Sansar the speculation gained credibility. With less support is the idea the Lab was showing off what can be done with VR using the Lab’s platform for VR: Sansar.
I doubt the Lab is panic rushing, but I do believe they are feeling pressure to meet a time line. For some time they have talked of their release schedule. I think one has to put a little salt with what they say to distinguish how much is wishful and realistic.
While we will have the ability to use VR headsets, I am not sure how much stuff their will be to support them. Because of that I expect early adopters to visit SL. The lag for all but those with high end machines is going to be a problem. Sansar will be way better in that regard, but it will be an obvious beta and lack the variety of SL.
There will VERY likely be unforeseen problems with Oculus and other headsets. It is going to be a chaotic start for VR. Lots of people will be polishing. In that respect the Lab has a huge head start because of their experience with SL. And the smaller SL team has the Sansar’s team to draw on.
If you click the tag Sansar at the bottom of this article you can pull up all the articles with Sansar information.