SL Moving to the Cloud

[LittleMe Jewell]

1 hour ago, ChinRey said:

Oh, that explains it.

So, you buy 100 Lindens on LindEx (never mind you can't buy such a small amount - this is an illustration), that's 100 Lindens added to SL's "GDP". You give those 100 Lindens as a tip to a DJ, that's another 100 Lindens added to SL's "GDP". The DJ buys a new shirt for the money - another 100 Lindens added. The shirt merchants uses the money to buy a new template and the template merchant finally sells it on LindEx for hard cash. Those 100 Lindens you bought added 500 Lindens to the economy. But it's not generated money, it's recycled money. There's a slight difference there.

As the old saying goes, there are three kinds of lies: lies, damned lies and statistics.

Yeah.  I'm not sure how they'd easily calculate a partial on that though.  Say I buy 100 L$ and use it to buy something from a creator.  If the creator does not cash it out, but instead uses it to upload 10 textures for new creations, then no money has come out of SL, but the economy did grow from the money that I put into it.

 

Besides, I have it on good authority that statistics are correct 62% of the time and only made up the other 76% of the time.  

 

Edited December 1, 2017 by LittleMe Jewell

[Parhelion Palou]

The comments in the video about number of users and the SL economy was just to show the audience that SL isn't a dead dog. LL's position is that they're making a lot of money from SL. They want to keep making money off of SL for as long as they can. One way to do that is to move the hardware to the cloud. Then they don't have to deal with hardware maintenance or upgrades any more. Assuming their cost drops (especially if they're shutting down empty sims) then that's potentially more profit now, but it also means SL can decline further and still make a profit. They might be down to one employee pushing the weekly roll button (tradition, ya know), but they'd still be making money.

[Oz Linden]

One point we'd like to correct: Moving to the cloud does not necessarily mean that we would stop running regions without avatars in them. For the time being at least, our goal is to run SL and its regions just as we do now - full time, with a substantially similar schedule for when and how they are upgraded. Doing otherwise could, as some have correctly pointed out here, introduce problems with backwards compatibility and mobility.

Whether or not we will also be able to create new land products that are not always running, how to decide when they run and when they don't, and what the pricing might be for such offerings are among the things we've had discussions around internally, but not made any decisions about yet. The cloud will broaden what we might do in this and a number of other exciting ways, and we look forward to exploring them, but that will happen only after we've gotten the current service uplifted.

[Phil Deakins]

11 hours ago, Oz Linden said:

Whether or not we will also be able to create new land products that are not always running, how to decide when they run and when they don't, and what the pricing might be for such offerings are among the things we've had discussions around internally, but not made any decisions about yet.

That's an interesting sentence. It says new land products, and not revised land pricing or sizes. I can imagine a possible new land product that fits - cheaper sims that are only turned on for specific hours, when the owners are awake, for instance.

So I'm still not holding my breath for any revisions to mainland sizes or costs. I hope I'm wrong, but I doubt it.

Has anyone mentioned roughly when SL is likely to be fully on the cloud?

Edited December 2, 2017 by Phil Deakins

[Parhelion Palou]

2 hours ago, Phil Deakins said:

That's an interesting sentence. It says new land products, and not revised land pricing or sizes. I can imagine a possible new land product that fits - cheaper sims that are only turned on for specific hours, when the owners are awake, for instance.

So I'm still not holding my breath for any revisions to mainland sizes or costs. I hope I'm wrong, but I doubt it.

Has anyone mentioned roughly when SL is likely to be fully on the cloud?

The senior systems engineer said it would hopefully be next year. She's probably high enough in the hierarchy to have the 'overly optimistic' syndrome -- though not as bad as Ebbe predicting Sansar in alpha by the end of 2014. Assuming she's half right, it should be 2019.

[Phil Deakins]

14 hours ago, Oz Linden said:

The cloud will broaden what we might do in this and a number of other exciting ways,

I agree with the word 'exciting' It does sound a lot different to the recent 'exciting' increase to 90 days of transactions history lol.

I can imagine a new substantially cheaper 12 hour sim, or 8 hour sim, or even 4 hour sim, for instance. That would be an exciting development - I think. Or perhaps the ability for the sim owner to turn it on and off as needed, paying a basic rate plus time turned on.

The facility of the system being able to activate and deactive sims rather quickly does open up a bunch of interesting possibilites.

Edited December 2, 2017 by Phil Deakins

[ChinRey]

14 hours ago, Oz Linden said:

For the time being at least, our goal is to run SL and its regions just as we do now - full time, with a substantially similar schedule for when and how they are upgraded. Doing otherwise could, as some have correctly pointed out here, introduce problems with backwards compatibility and mobility.

The big question then is sim crossings. Do you have a plan for those?

I an only see three ways to handle that on a cloud based grid. One of them will make them considerably worse than they are today, the other two should make them considerably better.

Another important question is of course how will this land prices?

Beyond that, if you can find a cheaper hosting solution that offers the same quality of service as you have today, that's great for Linden Lab, go for it! It's not really something that concerns the average user though.

[LittleMe Jewell]

7 hours ago, Phil Deakins said:

 I can imagine a possible new land product that fits - cheaper sims that are only turned on for specific hours, when the owners are awake, for instance.

 

What use is a sim that is only alive during my waking hours or when I log in, other than me putting my home there?   I suppose for someone that does a lot of building and needs something just for that purpose, it could be useful, but it couldn't be anything that you typically have the public accessing.

I could see a cloud setup, with quick spinup and shutdown, an easier and cheaper way for event organizers to do events.  Spin something up quickly and then release it all in a week or two - LL might do that for a cheap fee.

[Phil Deakins]

2 hours ago, LittleMe Jewell said:

What use is a sim that is only alive during my waking hours or when I log in, other than me putting my home there?   I suppose for someone that does a lot of building and needs something just for that purpose, it could be useful, but it couldn't be anything that you typically have the public accessing.

Maybe not useful to you, but you followed the question with 2 perfectly good reasons for such a thing And I'm sure there are more.

[Callum Meriman]

On 12/1/2017 at 2:15 AM, Rhonda Huntress said:

 A virtual server just sits in RAM (possibly non-volatile memory at that) and when it is idle, it simply does not get any CPU cycles. 

I'm not sure this could happen. Any time dependent scripts in the frozen sim, say a marketplace delivery box amongst others would stop working if the sim was empty. Then of course, if it came live there would be a flurry of events and some spectacular lag.

[Theresa Tennyson]

2 hours ago, Callum Meriman said:

I'm not sure this could happen. Any time dependent scripts in the frozen sim, say a marketplace delivery box amongst others would stop working if the sim was empty. Then of course, if it came live there would be a flurry of events and some spectacular lag.

Think of it as if you're running an application on your personal computer. If you want to open a file and the application you're using isn't running it will take quite a while for the application to start up and open the file. If the application is already running, though, opening the same file will be much quicker.

You're also assuming that the simulation of the region has to be an all-or-nothing thing. There is no reason that the always-on scripts for non-physical activities like marketplace deliveries would be dependent on the complicated and resource-hungry simulation of the physics and avatar interactions of the region. It should be very possible to leave the non-physical functions on all the time and turning the physical functions on only when they're needed.

[Theresa Tennyson]

20 hours ago, ChinRey said:

The big question then is sim crossings. Do you have a plan for those?

 

A vast number of regions in Second Life are either adjacent to no other regions or no more than, say, three other regions. The single regions would have no reason to deal with sim crossings and the small groups could come up as a single entity. They'd need to because a region has to be running to be visible and you often see into more than one region at once.

[Callum Meriman]

1 hour ago, Theresa Tennyson said:

You're also assuming that the simulation of the region has to be an all-or-nothing thing. There is no reason that the always-on scripts for non-physical activities like marketplace deliveries would be dependent on the complicated and resource-hungry simulation of the physics and avatar interactions of the region. It should be very possible to leave the non-physical functions on all the time and turning the physical functions on only when they're needed.

I can see one out sailing (diaganol path, 64M crossing) - in your scenario - and the boat crosses into a "sleeping" sim. That 5 second hand over just became a crash as all the sleeping daemons spool up to a running state after the boat leaves the sim.

Edited December 3, 2017 by Callum Meriman

[Phil Deakins]

I shouldn't be concerned about details of what might happen with sleeping sims. Oz said that they are not going to do anything like that at first. They are doing it sensibly, imo - get SL up and running as it is right now. When that's done and working fine, then they'll no doubt look at such changes as turning sims on and off. One step at a time.

[Theresa Tennyson]

39 minutes ago, Callum Meriman said:

I can see one out sailing (diaganol path, 64M crossing) - in your scenario - and the boat crosses into a "sleeping" sim. That 5 second hand over just became a crash as all the sleeping daemons spool up to a running state after the boat leaves the sim.

If you can see into it it's not sleeping. That's how region idling works now.

[ChinRey]

2 hours ago, Theresa Tennyson said:

A vast number of regions in Second Life are either adjacent to no other regions or no more than, say, three other regions. The single regions would have no reason to deal with sim crossings and the small groups could come up as a single entity. They'd need to because a region has to be running to be visible and you often see into more than one region at once.

That's how Kitely is doing it and that's exactly why I ask. Generally Linden Lab shows little or no interest in mainland so it's quite possible they have overlooked this factor. For the people who like mainland and the large themed pricate continents- and especially for the sailing community - sim crossings are already a serious issue and if it gets even worse, it will be a significant reduction in the quaity of the experience. It's good to see Oz mentions specifically problems with mobility as soemthing they want to avoid but even so, some reassurance that they are taking sim crossings seriously would be nice.

[Theresa Tennyson]

11 minutes ago, ChinRey said:

That's how Kitely is doing it and that's exactly why I ask. Generally Linden Lab shows little or no interest in mainland so it's quite possible they have overlooked this factor. For the people who like mainland and the large themed pricate continents- and especially for the sailing community - sim crossings are already a serious issue and if it gets even worse, it will be a significant reduction in the quaity of the experience. It's good to see Oz mentions specifically problems with mobility as soemthing they want to avoid but even so, some reassurance that they are taking sim crossings seriously would be nice.

So, your position is that Linden Lab has little or no interest in what Premium members pay for, and is in fact the foundation of the original concept of Second Life, and therefore have forgotten about it? I'll give you another chance -- you're sure this is the position you want to be on record as holding?

There seems to be an attitude in this thread that the only way of doing things is how things are done now, and therefore the current setup will need to be moved to the "cloud" exactly the way it is rather than Linden Lab saying, "This is what we have to do. How can we do some or all of it by using the cloud?"

This attitude is common in many threads in the forums. Some people just can't comprehend that there is more than one way to do what's being done now. To quote the movie Chicken Run, "Babs, there is no farmer..."

Edited December 3, 2017 by Theresa Tennyson

[Phil Deakins]

13 minutes ago, Theresa Tennyson said:

There seems to be an attitude in this thread that the only way of doing things is how things are done now, and therefore the current setup will need to be moved to the "cloud" exactly the way it is

You make it sound like it's incorrect, Theresa, but this is what LL, in the guise of Oz Linden, said in this very thread...

On 01/12/2017 at 10:06 PM, Oz Linden said:

For the time being at least, our goal is to run SL and its regions just as we do now

It seems pretty clear to me.

ETA: And, FWIW, I'm on record many times saying that LL shows little to no interest in its paying customer. That includes premium customers. ChinRey is not alone in that sort of opinion. I stand shoulder to shoulder with her. I'd prefer to stand face to face with her - very very close - but I'll settle for shoulder to shoulder

Edited December 3, 2017 by Phil Deakins

[Qie Niangao]

I'm not sure where folks are getting their assessment of current cloud technology -- particularly communications latency between instances -- but I have no reason to doubt that a cloud-hosted Mainland could at least compete with the little SONET ring the Lab had (has?) between datacenter sites. Granted, sim crossings sucked then, have always sucked, and will continue to suck forever and ever, amen. Simply migrating the existing architecture to the cloud, as-is, needn't greatly affect performance -- although it probably wouldn't save much expense, either.

The prospect of offering some new, more affordable land "product" that more aggressively "spins-down" the sim than the current idling, that's a whole separate consideration. I'd guess such a product is just not gonna be available for anything but void-surrounded single-region islands. It's less clear how much more aggressive than current sim-idling they could get and how much that would affect what folks expect of their "land" ownership. If it supports neither scripts nor extra-avatar physics at all, well, that's still a product some folks would buy. (I mean, it would be adequate for 99% of LEA installations, for example.)

I am, however, concerned that Mainland is left further and further behind in these offerings. The big step-function was when Estates were offered the grandfathered rate buy-down, putting them at par with Mainland region tier cost despite the many Estate advantages and controls denied Mainlanders. Meanwhile, there's been no break at all to Mainland tier for... a decade? can that be correct? .. during which time Estates have acquired (and repeatedly adjusted fees for) Homesteads and OpenSpace products.

At this point, if the Lab is so ashamed of the Mainland that they can't bestir themselves to update the product offering, just give us a date and we'll get out.

[ChinRey]

On 3.12.2017 at 1:25 PM, Theresa Tennyson said:

So, your position is that Linden Lab has little or no interest in what Premium members pay for, and is in fact the foundation of the original concept of Second Life, and therefore have forgotten about it? I'll give you another chance -- you're sure this is the position you want to be on record as holding?

Yes.

Edit: Although to be fair, it's not nearly as bad as it used to be. Back in one of the earliest discussions about Sansar on this forum, somebody asked about mainland and how something like that could be transferred to Sansar. The Linden who replied - I think it was Ebbe but it's a while ago so I may be wrong - was taken completely by surprise. The answer was essentially that it hadn't occured to them that anybody wold care about that and they needed to think about it. They obviously have done some thinking and mainland is no longer completely ignored as it used to be. But it's still not anywhere near where it should be.

Edited December 4, 2017 by ChinRey

[Love Zhaoying]

On ‎12‎/‎2‎/‎2017 at 4:25 AM, Phil Deakins said:

It says new land products, and not revised land pricing or sizes. I can imagine a possible new land product that fits - cheaper sims that are only turned on for specific hours, when the owners are awake, for instance.

Think about all the options:

- Bigger land are that supports fewer avatars

- Smaller land that supports more avatars

- Land that is only online when YOU (the owner) is online

- Land that is designed to be "completely disconnected" from the rest of the Map

- Land designed to run "better" with more/fewer objects or avatars (more L$/m2)

etc.

[Kadah Coba]

If land was cheaper, they'd sell more land.

At a minimum, making the non-adult mainland tiers either cheaper or each tier having more sqm would go a long way in doing something with all that abundant abandoned land everywhere.

If full private regions were cheaper, I'd consider getting another or two as many more options start to become more viable.

[Phil Deakins]

1 hour ago, Kadah Coba said:

If land was cheaper, they'd sell more land.

At a minimum, making the non-adult mainland tiers either cheaper or each tier having more sqm would go a long way in doing something with all that abundant abandoned land everywhere.

They can't make tier any cheaper, or it would be free.

I don't see a need to do something with the "abundant abandoned land". Because of it, loads of people have much nicer plots of land.

Edited February 6, 2018 by Phil Deakins

[Callum Meriman]

1 hour ago, Kadah Coba said:

If land was cheaper, they'd sell more land.

Changes to land need to be handled very carefully, it's a pretty fine balance. The risk could be with cheaper land that people don't buy more, or drop from a rental company, or the Lab find they can no longer pay for the developers and support (staff is where MOST of the tier goes, I imagine) so, service drops.

Last time the Lab tried a major change (Openspace debacle, Jack Linden, 2008) they badly stuffed it up, they burnt a lot of people, they lost a lot of customers, and seriously they lost a lot of good will that I doubt they can ever fully regain.

Jack Linden did a lot of damage to the world with his Openspace free for all, then snatch-back. I lost a sim then, had to abandon it. I don't forgive that stuff up of the Lab's at all.

From grid survey - The sharp 2008 peak and sharp drop are both Jack Linden's doing because he didn't really understand that balance.:

 

[Theresa Tennyson]

2 hours ago, Kadah Coba said:

If land was cheaper, they'd sell more land.

At a minimum, making the non-adult mainland tiers either cheaper or each tier having more sqm would go a long way in doing something with all that abundant abandoned land everywhere.

If full private regions were cheaper, I'd consider getting another or two as many more options start to become more viable.

In these threads I see a lot of people saying, "If land were cheaper I'd buy more land."

What I don't see so much is people saying, "If land were cheaper I'd give Linden Lab more money."

More land in Second Life isn't really an advantage to the Lab after the initial purchase of the furnishings for the new areas, and with the way the system is set up more land equals more expense - there's very little economy of scale if you grow Second Life from what it is now; it may actually be reversed.