Survey Suggests Substantial Market for Cloud Streaming Second Life -- Which an Expert Says is Feasible at $99/Year!

35% of NWN readers who took my survey on paying to stream Second Life said they'd be interested in doing so, with the most popular rates $10 a month for up to 30 hours, immediately followed by $40 a month for unlimited use. Almost all regular NWN readers are SL users, and with 255 of them taking the survey, I'd say it suggests a pretty substantial potential market. Yes yes, most respondents said they were not interested in paying to cloud stream, but again, as they are likely longtime SLers who already have a system that runs the world well enough, that's not a surprise. I'd say 1 in 3 is too large a market to ignore.

And here's the interesting thing: An actual streaming subscription rate would probably be cheaper than the most popular options I suggested. I checked this with an expert, Michael Zyda, Director at the USC GamePipe Laboratory, and here's the ballpark estimate he gave me. "Same as iCloud," as he put it to me. "Basically $99/year." He even suggested a varied pricing model that's free on the low end: "Tiered like Dropbox. So free piece, like Dropbox, and then more money for superusers." (And I'd suggest marketing the free version to new users, to give them a high quality taste of SL.) And adding such a plan would address three significant problems for Second Life: Poor retention rates of new users, trouble transitioning existing users to the more resource-heavy mesh viewer, and perhaps most important of all, the ongoing loss of SL land revenue.

For just imagine, say, if 100,000 SLers opted for a $10/month streaming subscription: We are then talking about $12 million in new yearly revenue for Linden -- and a better world for all. How about it in 2012, Linden Lab?

Update, 6:30pm: Just heard from another expert, this one a virtual world/game developer extremely well-versed with SL deployment architecture. He told me this streaming could probably be done by deploying Second Life on Linux VM. "Then the cost will be lot cheaper," he estimates, "like 10-20 cents/hour." If not on Linux VM, then costs would increase to about $1-2 an hour.

Comments

[Ann Otoole InSL]

If it makes SL completely functional from an iphone or android I bet a lot of people involved in SL commerce would use it from time to time. But it has to have full functionality. The "basic viewer" thing would not work. Must be able to access and share inventory, make payments, send IMs, etc.

[jjccc coronet (@JJcccART)]

Second life is free right now and your saying 1 in 3 are prepared to pay for it. so 2 in 3 are not prepared to pay for it which means the end of second life if they try that one on

[Ezra]

Doesn't sound like that guy knows what you were even asking. iCloud and Dropbox are priced based on gigabytes of data stored. Cloud rendering has nothing to do with that.

We have clues about how cloud rendering would actually work with Second Life, since more than likely Linden Lab would use something like Otoy.

Based on http://www.otoy.com/media/press/launch.html we know that for every 3,000 screens streamed in HD, Otoy needs 500 GPUs and 250 CPUs.

With your proposed scenario of 100,000 users subscribing to use a cloud rendering client, that means Linden Lab would need north of 8,000+ new CPUs and 16000+ GPUs; the GPUs a server cost Linden Lab has never had before.

Focusing just on the CPU requirements though, Linden Lab would need Opteron 6100s according to that Otoy spec, which are 8-core at their lowest end. So 64,000 new CPU cores.

Based on the last peek Linden Lab gave us into their infrastructure http://www.infoq.com/news/2008/12/Second-Life-Ian-Wilkes we know Second Life runs off a bunch of 4-core Xeons. Each core hosting one full region per core.

We know at present Second Life is what, 30,000 regions? Not all of them even 'full' regions. So proposing Linden Lab having enough infrastructure ready for 100,000 cloud rendering subscribers at any given time is proposing Linden Lab buy and operate enough hardware to run the simulators as they exist now more than twice over.

And at just $99 dollars a year per user.

You don't find it a bit ridiculous to champion this while at the same time proclaim Linden Lab can't possibly lower the cost of tier?

[yoshiko Fazuku]

for those that say it would be neet on a ipad iphone or what not, wait till you get your 4g wireless bill.

[Hamlet Au]

"Doesn't sound like that guy knows what you were even asking."

Ezra, Mike is quite familiar with SL and how it's architected and deployed, I wouldn't have asked him otherwise.

[Ezra]

Citing he's an expert and saying that you asked him questions unbeknownst to us really doesn't clear up why iCloud and Dropbox pricing schemes apply at all to cloud rendering.

A more indepth analysis from either him or you would be greatly appreciated, that is if your goal was to convince this would be a good idea vs. all the other things Linden Lab could do with a whole lot of new servers and pricing schemes. Like lower tier for starters.

[okiedokiecow]

If it were extremely dependable, 99 per year is not a bad price!

[Ordinal Malaprop]

There is no streaming subscription rate.

Let me say this again. There is no streaming subscription rate. It is all imaginary. "Your world your imagination" was never meant to cover this sort of thing.

By all means wander on making posts on the basis that this technology is going to work - perhaps eventually it will, though it's been years since the first time it was breathlessly promoted as the future of online 3D - but please do tag it all as "imaginary with no connection to anything".

[Gaga]

I would be far more interested in paying $50 to a $100 a month to Linden Labs for a premium account that allowed me to access my SL avatar ID and inventory from a cloud so I could teleport to other CLOSED Opensim grids that have been vetted, approved and licensed by LL. I think this would be a far better use of cloud technology and I feel sure it would increase premium account holders dramatically, not to mention growing a bigger SECURE market for virtual content which Linden Labs would still largely dominate. It would also allow them to move away from land sales business model that is almost certainly doomed to collapse with falling prices in Opensim grids within the next few years as more abandon SL for the greater freedom and lower costs of the open metaverse.

With greater opportunities for people who can't afford LL's high setup/tier costs to build their dream sims and grow businesses in grids like OSgrid (currently with over 10,000 regions and growing rapidly) LL could grab a golden opportunity here to secure their future.

[elizabeth (16)]

$99 a year

i would get that if i got a 1024m as well

[Joey1058]

Personally thinking, is 35% worth sustaining a business that is clearly spinning it's wheels in terms of an innovative direction? I believe one of the reasons Wil Wright came to Linden Lab was to redirect SL in the direction of the Sims. And of course the Sims is a game, not a community. Parent company EA's unofficial position has been "we've dropped below profitability, shut it down". How many iterations of the Sims have there been?

My opinion is if you want a true "SL experience", that direction has gone to the OpenSim movement. "Second Life/The Sims" ultimate destination is Facebook.

[shockwave yareach]

The issue is not what OS the rendering farm runs on. The issue is not how many users a single computer and GPU can support. The issue is not how much it will cost us to use a streaming service.

The issue is that ATT and Verizon won't support the astronomical datarates this requires. And frankly, I don't need nor want SL on my ipad or iphone -- trying to use SL on a virtual keyboard would be enough to drive me insane.

Nobody is trying to find a way to make WOW work on such things. WOW is making tons of money still, even though it requires a stout computer. So the problems with SL's misfortunes are not the requirements. For if that was the case, then Blizzard's execs wouldn't be filling hottubs with hundreds and splashing about naked in them. Streamed SL will still be SL, and without marketing and an improved first day experience, you'll just have streaming churn as well as unstreaming churn.

If the lab cannot bother to fix core issues and deal with the real problems inside the funhouse, simply adding another door to get in won't do any good.

[Arcadia Codesmith]

It solves the problem it solves -- it makes SL look good and work smoothly on any device capable of handling a live stream.

It's not a magic solution to all our woes, and I don't think anybody claims it would be.

$8-10 a month for unlimited access strikes me as fair (with the status quo remaining for people who can't or won't subscribe).

[Adeon Writer]

I world prefer to see SL running natively on a tablet. It's entirely possible.

[Alazarin]

Ugh... in spite of all the fancy talk about cloud streaming/rendering all I can see is being priced out of SL because the subscription rate would go past what I could afford. MySpace is dead, I've been locked out of FaceBork, Google+ is just another spammy/sucky FB clone... The internet seems to be getting smaller by the day.

[Pussycat Catnap]

It still has to get past mobile company caps on 3g/4g monthly bandwidth. A limit of 1gb / month by most providers in the USA means you'd not get far in SL.

[Ezra]

@Adeon

Me too.

iPad 2 is said to be half as powerful as Xbox 360, and the iPhone 4S almost as powerful as the iPad 2. This year we're getting a whole bunch more new smartphones and tablets including the iPad 3 and first run of Windows 8 tablets. And of course everything Android.

As is, there's games like Rage Mobile and engines like Unreal and Unity on smartphones. The smartphones and tablets running them most definitely have the graphical power to run Second Life natively.

There's the bandwidth concerns though. Most of us have Second Life caches larger than our mobile data plans. Streaming assets over the wire mobile isn't plausible with data caps.

Cloud rendering doesn't fix anything when it comes to mobile data caps, since streaming the client in HD video makes a cloud rendered Second Life client as plausible as watching Netflix on the go with a data cap. Worse actually since cloud rendered applications are dependent upon a hefty upstream as well as downstream.

I'm sure Linden Lab has plans for smartphones and tablets, but cloud rendering is a solution that doesn't solve problems Second Life actually has, except maybe running in the browser.

[Ciaran Laval]

At the very least Linden Lab should be investigating cloud services for people who own sims but would rather pay by the minute rather than a flat monthly fee, for private users this could prove to be very cost effective, for commercial ventures, the flat fee will win.

However I've gone off on a tangent from the concept of mobile computing and Second Life there.

[Nightbird Glineux]

I think streaming via a home ISP (cable or DSL) is a separate problem from streaming to a public WiFi hotspot or via a 3G/4G mobile data plan.

I have a Kindle Fire. Somebody wrote in a blog comment somewhere about using Splashtop Remote Desktop to stream Second Life to his Android tablet from his desktop computer over his home WiFi network. I'm going to give that a try just for giggles. :)

[Nightbird Glineux]

BTW, my computer is a 2010 13" MacBook Pro with 8GB RAM and an SSD. It CHOKES on the current mesh viewer :( so I'm taking talk about running a client on an iPad 3 or other current or near-future tablet with a shaker of salt.

[shockwave yareach]

@Nightbird: You may want to doublecheck your graphics settings. There is a glitch in the installer for v3 and it puts every single feature to absolute max. Dial them back down without 16xAA (good lord) and reasonable draw distances and you'll probably choke less. The problem is the settings the installers sets, not the viewer. Fortunately it's easy to fix.

[Pussycat Catnap]

It doesn't crank them up for everyone. I had to dial the settings up from where it put me. It tries to pick something based on what it thinks your system can handle - much as it always has.

But its guessing apparently could use some looking into in current versions.

[CJ]

Second Life should cut a deal to put the game on STEAM - and benefit from Valve's market of millions of gamers!
STEAM, if you haven't heard of it before, is the iTunes of PC computer games.
Thousands of games made availabl. To millions of players - a couple clicks and STEAM downloads, installs, updates, and maintains all your games. STEAM even helps maintain your Windows graphics and subsystems software - making sure your games will work just right.
STEAM could introduce a whole new population of players to Second Life.
As a promotion, Linden Labs can get Valve to list Second Life right on the front login page so instantly - over 4,000,000 players worldwide will see Second Life - many for the first time. Linden Labs probably loses sight of the fact that millions of people have NEVER Heard of Second Life - but would love to try it.

Oh, and gamers tired of spending all their disposable income on MMO in game gold for hack & slash quests are a new source of revenue and profits from 'pre-screened' self selecting customers. Anyone using STEAM probably already has a library of dozens of games they've spent hundreds if not a few thousands of dollars on buying & playing.
STEAM users also tend to have 'High End' gaming systems with the best graphics & CPUs, broadband connects for fast real time gaming, and matching real world income levels to afford all the latest hardware, software, & related internet technology.

Second Life viewer needs put on STEAM - the sooner the better!

[Shug Maitland]

I think some ppl are missing the point. A streaming SL *option* would make a dent in user retention. We probably all know ppl who struggle to run on equipment that I would call inadequate but is what they can afford. This would open the choice between a downloaded client on a $2000+ machine and streaming SL @$10/month to a $500 machine.

[Dale Innis]

I'm sure those people are experts, but I'm not sure they actually understood what it was you were asking them.

As other people noted, neither iCloud nor Linux VM have anything at all to do with the actual expensive resource here, which is the rendering hardware.

To do this kind of streaming, you need to render the world in realtime, separately, for every user using the streaming service. So that's the "rental" cost of a decent graphics card, times the number of people-hours using it, reduced by whatever economy of scale (if any!) you get from having lots of rendering going on in the same room.

Does that come out to $99/year? I dunno! :) But answers involving iCloud or Linux VM aren't likely to be relevant...