Jump to content

Now that Second Life is in the cloud...


Recommended Posts

...wouldn't it be lovely if Linden Labs took a look at what other up and coming virtual worlds are doing? Increase prims, increase region sizes, keep the prices what they are now. 512 x 512, and 75,000 prim allowance to start, with the option of increasing it to 100,000. Start with private regions, then mainland. It would force rebuilds across the grid - but would that be a bad thing? So much of the mainland looks awful. Good sim builders are always rebuilding anyway.

  • Haha 3
Link to post
Share on other sites
49 minutes ago, RaeganBriana said:

...wouldn't it be lovely if Linden Labs took a look at what other up and coming virtual worlds are doing? Increase prims, increase region sizes, keep the prices what they are now. 512 x 512, and 75,000 prim allowance to start, with the option of increasing it to 100,000. Start with private regions, then mainland. It would force rebuilds across the grid - but would that be a bad thing? So much of the mainland looks awful. Good sim builders are always rebuilding anyway.

Wait, you think that now that all services are running in the cloud, the work is done?
Not by a long shot! First fo all, there are services that did not take kindly to being uplifted. The map is one such example.
LL also has to tweak a lot of stuff because running software on Amazon Cloud Services (ACS) is something completely different than running in their own datacenter where you control literally everything.
Once that process is done we are one, maybe two years further along into the future.

So, fun little ideas (especially the 512x512 regions which would take up 4 times the database space) but completely infeasible.

Link to post
Share on other sites
26 minutes ago, Profaitchikenz Haiku said:

And please could we have the scripts run percentage uplifted? If anything is sagging around the navel and in need of support, it's that.

that is actually already in the pipeline because it has been determined that scripts run with more wiggle room now.

Link to post
Share on other sites

Also remember OZ linden himself just said in a recent live stream, AWS costs most than their data center where they had physical hardware,  so keep that in mind, nothing is going to get cheaper for a very long time at this point. 

Link to post
Share on other sites
6 hours ago, RaeganBriana said:

It would force rebuilds across the grid - but would that be a bad thing? So much of the mainland looks awful.

Can't really force rebuilds or hope for that it might meet arbitrary standards of "quality".

Give people more prims, some will just rez more literal fullbright prims, some wont change anything because they left SL a decade ago and are still paying for the land, the rest of us will rez more hyper detailed gacha tat.

  • Like 1
Link to post
Share on other sites
7 hours ago, bigmoe Whitfield said:

Also remember OZ linden himself just said in a recent live stream, AWS costs most than their data center where they had physical hardware,  so keep that in mind, nothing is going to get cheaper for a very long time at this point. 

I told Oz that was going to happen a year ago. He insisted it wasn't.

AWS is not usually cost-effective when you have compute-bound dedicated servers running 24/7. AWS is cost-effective for loads that vary with time of day and season, because you can acquire and release resources, paying only for what you're using. But SL sim servers run busily even when no one is in a region.

Link to post
Share on other sites

From experience, hardware/platform decisions are always made on a commercial, not a technical basis, so it is interesting that they have gone down what may not be the cheaper (initial) approach, but I suppose looking at it over time, scalability and offloading a lot of in-house responsibility might have been their main objective, spend a little more now to save a lot in the future ?

Link to post
Share on other sites

I seem to remember that a few years ago SL did have a thing where empty regions stopped running until someone dropped in. I don't remember any announcement from LL that they'd stopped doing this, but I assume they have. Perhaps too many Meeroos died.

If AWS were leasing server time by the second, reintroducing something like this might be an attractive proposition. And the resulting evolutionary pressure on the many species of breedables inhabiting SL would be a positive thing.

Link to post
Share on other sites
3 hours ago, Ardy Lay said:

Second Life was mostly running on 10 to 12 year old hardware that was over-due for replacement, by some business standards.  Don't forget that.

/me slaps his maxxed out dell 2950 Gen 2 machine,   this puppy is worse than a mining rig at times.

Link to post
Share on other sites

I'm not sure if it has anything to do with Second Life moving to the cloud, but ever since they did I'm experiencing a huge drop in UDP bandwidth. Textures are loading with a whopping 20 Kb/s if I'm lucky. It literally takes forever to load textures. I would appreciate it very much if someone would open the tap a bit further because this is breaking Second Life. Sometimes it's okay but there are times when it's totally unplayable. And no, it's not a local issue. I have more than enough bandwidth available locally. I wonder if anyone here has the same experience?

 

bandwidth.jpg

Edited by Sherman McCallen
Link to post
Share on other sites
9 hours ago, animats said:

I told Oz that was going to happen a year ago. He insisted it wasn't.

AWS is not usually cost-effective when you have compute-bound dedicated servers running 24/7. AWS is cost-effective for loads that vary with time of day and season, because you can acquire and release resources, paying only for what you're using. But SL sim servers run busily even when no one is in a region.

I think part of it depends on what a company would have to do to keep running their own servers.  My understanding is that the LL data center was out of space.  So they would have to build or rent a new larger location.  Building something new is definitely expensive and renting may or may not be better than AWS depending on the location - assuming there was space available somewhere to build or rent.  Then they would still have had to buy tons of new servers to replace the outdated ones.  Sometimes, it is a matter of weighing current required expenditures versus costs over time.  A company's options can end up severely limited based upon what sort of planning has or has not been done in previous years.

Over time, I do agree that AWS is seldom ever cost effective if your servers need to be up running at full throttle all day every day.  

Edited by LittleMe Jewell
Link to post
Share on other sites
4 hours ago, Ardy Lay said:

Second Life was mostly running on 10 to 12 year old hardware that was over-due for replacement, by some business standards.  Don't forget that.

LL stopped telling us what they were doing in the datacenter about 10-12 years ago. Not quite the same thing. There is no reason to assume LL's datacenter operations were not proactively maintained right up to the very last day of operation, this includes cycling hardware and consolidation

Link to post
Share on other sites

I suppose 10 years ago the virtualization was not quite as good as it is today, neither were the servers. Space seems to be a poor excuse, the number of residents hasn't exactly stayed the same either. I suspect this using cloud services thingy is merely about cutting the costs. The sad thing is that those of us who stayed are now softly pushed towards the exit by this poor performance.

  • Like 1
Link to post
Share on other sites
6 hours ago, KT Kingsley said:

I seem to remember that a few years ago SL did have a thing where empty regions stopped running until someone dropped in. I don't remember any announcement from LL that they'd stopped doing this, but I assume they have. Perhaps too many Meeroos died.

If AWS were leasing server time by the second, reintroducing something like this might be an attractive proposition. And the resulting evolutionary pressure on the many species of breedables inhabiting SL would be a positive thing.

Region Idling is still working as it used to be.

  • Thanks 1
Link to post
Share on other sites

There are alternatives to AWS. You can rent space in a data center and install your own hardware, which is what LL used to do. You can lease a bare server with hardware support and install your own software. (I have one of those for a non-SL project.) That's more common today than owning the hardware, because the data center keeps a stock of spare machines and will replace them as necessary. Plus there are lots of other options for web-oriented servers that aren't relevant here.

In general, leasing bare servers is the most cost-effective alternative. But to get good rates, you have to lease for at least a year, so capacity planning becomes important. With AWS, you can scale up or down as needed.

1 hour ago, Coffee Pancake said:

LL stopped telling us what they were doing in the datacenter about 10-12 years ago. Not quite the same thing. There is no reason to assume LL's datacenter operations were not proactively maintained right up to the very last day of operation, this includes cycling hardware and consolidation

We know that towards the end, servers were not being added, because LL announced you could not buy new regions until cloud uplift.

Now you can't buy SL new regions online because the "land store" is apparently not yet integrated with the AWS system for adding more capacity. AWS lets you add more servers in seconds. The land store has to talk to the AWS ordering system properly for that to work. LL support is doing that manually for now.

Link to post
Share on other sites
5 hours ago, animats said:

Now you can't buy SL new regions online because the "land store" is apparently not yet integrated with the AWS system for adding more capacity.

It was announced in the last Concierge meeting that the land store is now back online.

https://secondlife-status.statuspage.io/incidents/509gysn88pqv

Edited by Lucia Nightfire
  • Like 1
Link to post
Share on other sites
7 hours ago, Coffee Pancake said:

LL stopped telling us what they were doing in the datacenter about 10-12 years ago. Not quite the same thing. There is no reason to assume LL's datacenter operations were not proactively maintained right up to the very last day of operation, this includes cycling hardware and consolidation

 

  • Like 1
  • Thanks 1
Link to post
Share on other sites

Given that SL is kinda a 'zero net gain' the higher end hardware you have - seems like the upgrades that make the most sense would be taking advantage of multi-threading and modern high end graphics cards, but given that is almost impossible to do - we just have to enjoy what little improvements they can dish out

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...