Question for Lindens.

[Love Zhaoying]

8 minutes ago, Fluffy Sharkfin said:

And before that the infamous Massive Grid Stress Test of 2005 during which they tried to squeeze a whopping 4000 users onto the grid at the same time! 😮 

Quote

On Sunday July 24, we are planning to have a massive stress test on the
grid, inviting everyone to stay logged in between 7-9pm PST to prove
that the grid can withstand ~ 4,000 concurrent users. We've also planned
exciting give-aways and prizes as an incentive to gather as many people
together as we can.

While we did experience some problems with the asset server this past
weekend (which were fixed on Monday), our database load is currently
running very close to zero and we feel we can handle the increasing
number of Residents on the grid. So, be prepared to see Second Life at
its busiest July 24!

Expand  

(I can't remember for sure if they managed it or not, but have a vague recollection of the total number of users getting pretty close before the grid finally collapsed and LL declaring it a success anyway)

What's our concurrency these days? Sorry if question is teh dum.

 

[Fluffy Sharkfin]

1 minute ago, Love Zhaoying said:

What's our concurrency these days? Sorry if question is teh dum.

 

I honestly have no clue, I think there's a "total users online" counter on the viewer log-in screen but it's not something I've ever really paid any attention to.

[Arielle Popstar]

14 minutes ago, Love Zhaoying said:

What's our concurrency these days? Sorry if question is teh dum.

 

Around 29K most mornings around 3 am SLT up to 50K in the evenings on the weekends.

[Love Zhaoying]

Just now, Arielle Popstar said:

Around 29K most mornings around 3 am SLT up to 50K in the evenings on the weekends.

At least that's way up from 4k (the 2005 test quoted earlier).

[Fluffy Sharkfin]

8 minutes ago, Love Zhaoying said:

At least that's way up from 4k (the 2005 test quoted earlier).

According to the SL stats page on Daniel Voyagers blog SLs highest concurrency was close to 90k

Quote

Second Life user concurrency levels from 2009 to 2020 – source

2009/2010 User Concurrency Levels: Max: 85, 000 to 88, 000

2022 User Concurrency Levels: Max: 46, 000 to 53, 000

[Istelathis]

20 minutes ago, Love Zhaoying said:

What's our concurrency these days? Sorry if question is teh dum.

 

At the moment, according to a script I found on these forums we are currently at 42,759.  That is pretty impressive when you think about it, the high of the year appears to have been January 29th,  where we were at 53,495.  The high might be wrong though, I think I reset the script sometime in January and at one time it had reached over 55k.  

I know a lot of people think SL is dead, but those numbers are pretty healthy.  Not quite up to par with Roblox, but probably consistent with the history of SL if I am not mistaken.  There is no way to tell for certain, and the number is probably wrong but according to one site WoW is currently at 107,514

For FFXIV users using the steam client, it is at 24,867 with a high of today of 28,405 in the past 24 hours - but that does not take into account non steam members which I haven't a clue.  Everquest, on steam is only 289 players 😢 RIFT shows as 179 players on steam.. what a waste I really enjoyed that MMO. 

 

[Indiana Nexen]

People can't Shop right now, WTF is going on ?

The Status web page doesn't seem worried at all about this.

~ Sad

[Love Zhaoying]

4 minutes ago, Istelathis said:

At the moment, according to a script I found on these forums we are currently at 42,759.  That is pretty impressive when you think about it, the high of the year appears to have been January 29th,  where we were at 53,495.  The high might be wrong though, I think I reset the script sometime in January and at one time it had reached over 55k.  

I know a lot of people think SL is dead, but those numbers are pretty healthy.  Not quite up to par with Roblox, but probably consistent with the history of SL if I am not mistaken.  There is no way to tell for certain, and the number is probably wrong but according to one site WoW is currently at 107,514

For FFXIV users using the steam client, it is at 24,867 with a high of today of 28,405 in the past 24 hours - but that does not take into account non steam members which I haven't a clue.  Everquest, on steam is only 289 players 😢 RIFT shows as 179 players on steam.. what a waste I really enjoyed that MMO. 

I wonder about Sansar..and Meta.

lol!

[Arielle Popstar]

1 minute ago, Istelathis said:

At the moment, according to a script I found on these forums we are currently at 42,759.  That is pretty impressive when you think about it, the high of the year appears to have been January 29th,  where we were at 53,495.  The high might be wrong though, I think I reset the script sometime in January and at one time it had reached over 55k.  

I know a lot of people think SL is dead, but those numbers are pretty healthy.  Not quite up to par with Roblox, but probably consistent with the history of SL if I am not mistaken.  There is no way to tell for certain, and the number is probably wrong but according to one site WoW is currently at 107,514

For FFXIV users using the steam client, it is at 24,867 with a high of today of 28,405 in the past 24 hours - but that does not take into account non steam members which I haven't a clue.  Everquest, on steam is only 289 players 😢 RIFT shows as 179 players on steam.. what a waste I really enjoyed that MMO. 

 

Think a lot of us are starting to realize how large of a bot population is included in those numbers. As an example I check the concurrencies every time I log in and am surprised at the consistency of the 29K in the mornings. It is rather suspicious when I see a much greater variance in other grids and Virtual worlds where Bots aren't counted.

[Love Zhaoying]

4 minutes ago, Indiana Nexen said:

People can't Shop right now, WTF is going on ?

The Status web page doesn't seem worried at all about this.

~ Sad

Since search has been broken, maybe they are working on it.

[Zalificent Corvinus]

15 minutes ago, Arielle Popstar said:

Think a lot of us are starting to realize how large of a bot population is included in those numbers. As an example I check the concurrencies every time I log in and am surprised at the consistency of the 29K in the mornings. It is rather suspicious when I see a much greater variance in other grids and Virtual worlds where Bots aren't counted.

"In the mornings" 

As in "Hey its late afternoon/early evening on the other side of the Atlantic and all those people at home are logging in, they can't possibly count?"

 

[Istelathis]

14 minutes ago, Love Zhaoying said:

I wonder about Sansar..and Meta.

lol!

If the numbers are correct, 4 steam users

[Orwar]

1 hour ago, Love Zhaoying said:

What's our concurrency these days? Sorry if question is teh dum.

   Firestorm launcher's front page right now:

34 minutes ago, Istelathis said:

RIFT shows as 179 players on steam.. what a waste I really enjoyed that MMO. 

   I had a lot of fun in that game, it's a shame it didn't survive. 

[Silent Mistwalker]

1 hour ago, Gavin Hird said:

Before that, the crash in Feb 2008 was epic. The grid just got slower, and slooower and slooooower, and was gone for 3 1/2 days.

Somewhere around 2005 or so there was a viewer release that caused the grid to be down for several days. 

Then there was the time when they'd rolled out an update and the grid was down for a week.

There were quite a few times in the earlier days when the grid was down for days and weeks at a time.

Fun times. 🎪

[Coffee Pancake]

2 hours ago, Qie Niangao said:

Did I miss the part where it's explained how Lumiya was just fine on the Google Play Store, but any other viewer will require SL pass a purity test?

The Google Play store has far laxer requirements than the Apple store. These things are not alike.

2 hours ago, Gavin Hird said:

You keep claiming it will never happen. Try again!

It's been 20 years, I've been here for 17 of them and we've had enumerable false starts and flubs.

The only legacy of which can still be seen in the desktop SL viewer to do this day; Swipe to pan touch screen camera controls, added for a mobile hail Mary that crashed and burnt so hard LL couldn't even manage to revert the setting so it made sense for desktop users.

Mobile SL might be very possible.

Mobile SL made by LL, designed by LL, fit for purpose or use of actual SL users. NOPE. Hell no.

 

We wasted multiple TPV meetings and hours trying to bring LL up to speed on why AOs were important.

 

[Arielle Popstar]

1 hour ago, Zalificent said:

"In the mornings" 

As in "Hey its late afternoon/early evening on the other side of the Atlantic and all those people at home are logging in, they can't possibly count?"

 

The variance of the minimum concurrency has less range than the maximum, by a fairly large margin.

[Love Zhaoying]

59 minutes ago, Arielle Popstar said:

The variance of the minimum concurrency has less range than the maximum, by a fairly large margin.

That's just..mean.

 

[Prokofy Neva]

14 hours ago, Zalificent said:

No, it showed paid actors doing sci-fi/fantasy cosplay while hanging from Hollywood flying rigs, with some digital effects.

No avatars were harmed or used in the making of that dreck...

True. But the point was, you could have avatars like that. Some are even in the library. 

I always thought that one bit with the flying rig looked like some bad highschool play, or like that "Bee Girl" video by Blind Melon.

I was surprised it won awards but good on them.

[Paul Hexem]

16 hours ago, Gavin Hird said:

It you place a client in the iOS store, it immediately becomes available to 1.5 billion active users. Even if 0.1% of those downloaded and tried the client, it represents 1,5 million users, which easily could happen within days if hyped up by media.

This still doesn't address the question of what makes you think they'd actually use SL beyond maybe a single login, of which SL already sees plenty of.

SL can't retain users on PC, the platform it works best on. How is it going to retain users on hardware that makes it even more of a miserable experience?

Edited February 17, 2023 by Paul Hexem

[Zalificent Corvinus]

13 hours ago, Arielle Popstar said:

The variance of the minimum concurrency has less range than the maximum, by a fairly large margin.

So, your complaint is that people on the right side of the Atlantic are more reliable users of SL than people on the left side.

Interesting.

[Aethelwine]

18 hours ago, Coffee Pancake said:

We were banned on Twitch long before they settled on adult content  .. Catznip actually did all there app integration work back in the dark days before OBS, the final step of which was to show your application and apply for an API key. They sat on it a week and refused us because "it's impossible to determine ownership of all content on screen" .. we would not have gone to the trouble had SL been banned.

 

I am sure this is obvious but feel impelled to point out that by that argument Skyrim, Fallout, the Sims... in fact any modifiable game would not be allowed on their platform.

[Qie Niangao]

2 hours ago, Zalificent said:

15 hours ago, Arielle Popstar said:

The variance of the minimum concurrency has less range than the maximum, by a fairly large margin.

So, your complaint is that people on the right side of the Atlantic are more reliable users of SL than people on the left side.

Interesting.

I think we can agree that there are non-zero bots on the grid, and that they are in-world as much of the time as their bot programs can keep them—basically constantly.

Whatever the distribution of concurrency by non-bots, that bot concurrency will contribute a near constant to the total concurrency distribution.

Because the bots are logged-in all the time, the minimum total concurrency will include a larger share of bots than the maximum concurrency. The variance will be lower, then, at the "bot-rich" lower end of concurrency than in the "human-rich" upper end.

Undoubtedly the Eastern Hemisphere's "prime time" produces a lower concurrency than the Western Hemisphere's, but for the purposes of estimating how many bots may be in-world, I don't see why it matters whence they're logging in.

On the other hand, bots aren't the only possible reason for some more or less constant component of minimal concurrency. It's not that lower variance necessarily corresponds to constantly logged-in sessions (like bots), but that those constant logins will necessarily contribute little variance to the concurrency distribution.

[Zalificent Corvinus]

11 minutes ago, Qie Niangao said:

I think we can agree that there are non-zero bots on the grid, and that they are in-world as much of the time as their bot programs can keep them—basically constantly.

Whatever the distribution of concurrency by non-bots, that bot concurrency will contribute a near constant to the total concurrency distribution.

Because the bots are logged-in all the time, the minimum total concurrency will include a larger share of bots than the maximum concurrency. The variance will be lower, then, at the "bot-rich" lower end of concurrency than in the "human-rich" upper end.

Undoubtedly the Eastern Hemisphere's "prime time" produces a lower concurrency than the Western Hemisphere's, but for the purposes of estimating how many bots may be in-world, I don't see why it matters whence they're logging in.

On the other hand, bots aren't the only possible reason for some more or less constant component of minimal concurrency. It's not that lower variance necessarily corresponds to constantly logged-in sessions (like bots), but that those constant logins will necessarily contribute little variance to the concurrency distribution.

Mid afternoon, GMT, Tuesday, middle of the rolling restarts, concurrency drops to 28 or 29k, no great surprise. One week there was "The Great Bot Viewer Failure", and there was  a 2.5k drop in concurrency compared to the usual Tuesday afternoon.

 

STUPID people screamed that 26.5k for a Tuesday afternoon compared to 55k for late evening at the weekend  "proved that over half the grid is filthy bots"

The stupid people were, as usual, wrong.

The posts I replied to, were the usual nonsense that acts as if there are no real people west of the golden gate bridge or east of vegas

Edited February 17, 2023 by Zalificent

[Qie Niangao]

3 minutes ago, Zalificent said:

One week there was "The Great Bot Viewer Failure", and there was  a 2.5k drop in concurrency compared to the usual Tuesday afternoon.

My memory of this is really vague; is there somewhere to learn more details about this event? It being on a Tuesday, I'd guess it was an update of the main server channel, but it would be interesting to know how much of the grid rolled to that update before they started a rollback. Also, I'm just generally interested in what a bot-specific failure mode would be, and why it would affect all (and only) bots.

(That said, I'd guess the total population of bots to be lower than "common knowledge" would have it, but that's more hunch than anything.)

[Orwar]

27 minutes ago, Qie Niangao said:

Whatever the distribution of concurrency by non-bots, that bot concurrency will contribute a near constant to the total concurrency distribution.

   Whilst this is true, I don't think SL has anywhere near the bot-to-human ratio as many MMOs do. Aside from Tiny Empires and other HUD games, there's no resource farming bots in SL.

   Meanwhile, World of Warcraft at one point was estimated to have a ratio of 60% bots the other year. So if they have 100,000 concurrent users and only 40,000 of those are humans, and SL has 45,000 concurrent users and 5,000 are bots, it just tells us that our user count is actually better than one would expect from looking at concurrency alone.