Kama Center down, Vallone overloaded worse than ever

[Qie Niangao]

3 hours ago, Qie Niangao said:

Again, all these are going to be relevant to business success in moving to the cloud.

The more I think about this, the more I worry.

We're never going to know what kind of service agreement the Lab buys from its cloud provider, are we? Or enough about the cloud-hosted SL server architecture to be able to guess what's happening when performance tanks (as if we could now).

And we'll never know when cost-saving measures are applied to the cloud contract(s).

Have to hope they know what they're doing. It's not as if we have any service contract with the Lab.

Interesting business, Second Life. Nice work if you can get it.

[Lucia Nightfire]

Ok, after sleeping on it, literally 10 minutes after I made the last comment, I realized an exponential cost is too harsh.

There still needs to be some additional cost per script or per script memory.

Right now, there is only a <0.5 LI cost per link IF that link contains ANY amount of scripts.

So, if a link contains 1 script or 500 scripts, it still costs <0.5 LI. This isn't fair in any regard.

Also, in regards to the % of land owned allowance, it was suggested a few Server UG's ago that there could be a sharing protocol that pulls available script timing from parcels that are not using their available script timing, but to make this fair, whenever a parcel that has had script timing pulled needs some script timing back, that amount has to be given back. There cannot be a first-come-first server basis or we still have the problem where one land owner's objects or their guests can consume any amount of available script timing in the region.

 

[Profaitchikenz Haiku]

6 hours ago, Qie Niangao said:

We're never going to know what kind of service agreement the Lab buys from its cloud provider, are we? Or enough about the cloud-hosted SL server architecture to be able to guess what's happening when performance tanks (as if we could now).

I get some deja-vu here from past experiences when software I was working on was shifted from 16-bit to 32-bit systems, and sometimes actually performed slower on the new and supposedly superior platforms. In the end we learned what to do to get things running again. (And the Ariane rocket program also learned about  integer rollaround taking far far longer when twice as many bits were involved

Edited September 1, 2019 by Profaitchikenz Haiku

[MBeatrix]

21 hours ago, Qie Niangao said:

[...] I think we should recognize that it's not the only "scripts issue" that affects SL. Some issues have been around a long time and are pretty serious. [...]

Indeed. And maybe some new as of late... Or it's just another drop in SL's performance, I don't know.

Anyway, I've been wondering if moving everything to the cloud is as a good idea as LL wants us to believe it is. So far, I haven't seen anything working better than before stuff started to be moved to it.