Tag Archives: FJ Linden

Linden Lab developing long-term plan.

The Linden Blog notes that the recent embarrassing level of crashing & outages is highlighting to Linden Labs the need for scalability.

Scalability shouldn’t be confused with land mass, which there is planty of. The problem is only so many people can get online at once right now…… about 80000 tops!

FJ Linden explains in technical language the tight situation. Basically Second Life is at a bottleneck. It needs to be scaled up big time!

Imagine you have a burger stand, that to your surprise becomes amazingly popular. At first, everything is going well. You can serve people in a timely manner, and the place is packed.

Then a problem arises. Your customer base keep multiplying exponentially! (In this world, the entire globe can travel to your door easily without commuting.) So people are lining up at the door & down the street.

Seems good, but you can only serve so many people so quickly! Your kitchen staff (the database system) is exhausted from being constantly overworked. Yet more people keep coming. You’ve expanded your dining area enormously, but you now have too many tables to serve. People start complaining and getting angry because they can only get served half the time. They came for steak & didn’t even get a menu brought to the table…..

The quality of the burgers (the grid) starts to goes downhill fast as chaos grows. As your cooks try to work faster & faster, they start making more mistakes (inventory loss, lag, downtime). Now, those wonderful famous burgers just don’t taste as good as they used to anymore. New customers wonder how this place ever got so big & successful?

So you hire a man (restricted logins) to stand at the door to turn away a good many of your paying customers (both new and old) upon arrival.

This obviously is a terrible solution….. in the long term, it is no solution at all, and would slowly ruin this otherwise successful business.

What is a better solution?

Expand the “kitchen” & remodel, reorganize that workspace towards maximum efficiency – or as FJ Linden says rearchitecture. It’s got to be a tough job redesigning something so that it can be expanded ad infinitum. Tougher still when they have to keep the grid running, while expanding it. This is what the Lindens are working on right now. I’m glad to hear it…. kudos to them for attempting what must be a very arduous task.

When Second Life is always filled to over-capacity these days, how can it grow more? Booting out a portion of the user-base continuously is not a solution – its can only be a stop-gap temporary fix. The pressure’s on at the lab now….. we need a stable & scalable grid — and pronto!

Can they do it? I hope so…. good luck LL. 🙂

Advertisements

New kind of bot in Second Life!

It’s the LOLO Kickabot!

The kickable robot.

This bot is NOT used to fake traffic, bother people, or steal stuff. Its quite harmless. Very Happy

Fully animated: its a walking, free-roaming robot!

You can make your robot talk & even give it a name too.

Best of all, if anyone or anything bumps into the robot, it falls down, and the bot can be kicked around on the ground!

Finally we can get back at the bots.

To take a look at the LOLO Kickabot on Xstreet SL shopping site, please click here.

To go inworld to The LOLO Pet Shop and see the LOLO Kickabot, please click here. (Here you can see the bot live & even kick it!)

In other news…..

The official Linden blog announces some fancy techno-stuff they are busily working away on. Seems lots of people get thrown right off when lots of long technical words are used….. so I’ll try to give a little summary, to the best of my understanding….. I promise I will use no long or unusual words. 🙂

First FJ Linden admits the grid has been a little choppy lately, but this is because big changes are happening. Lindens have been building a special new network called “LLnet”, just for us! This new network will soon be running the SL for us. FJ says LLnet is ready so, next week they will start to move some of our avatars and stuff onto this new network.

Glad to hear they are not trying to just quick dump us all over there…. seems like a slow & careful process, hopefully avoiding grid crashes & downtime.

Next FJ describes the good things about LLnet. Right now we’re all running on this “VPN” system. It’s kinda slow, aged, and clunky. LLnet will be faster, less laggy, and not crash the grid so much.

But that’s not all…. The Lindens have been working on “improving asset storage“. This means keeping your data in a smarter way – more organized. Their new way of storing stuff will save Linden Lab lots of storage space and speed everything up for us too.

It will keep all those great things you like to play with all the time in a place where the grid can fetch them quick for you, when needed.

The things you don’t use much, will get tucked neatly away, so they are not in your way, where you trip over them all the time… You don’t want to sit your home gym out where its half blocking off your front door right?

Its kind of like keeping your car keys on a hook on the wall, instead of wasting time hunting for them each time, because last time you randomly threw them down behind the couch & they’re now hidden under that ugly sweater you never wear, and an old pizza box.

So, soon when you search for your favorite jewelry, it should pop right up in inventory the quickest.

3 cheers for Lindens deciding to put everything in its place & clean their room!

Next FJ goes on to tell us about “HTTP Dataserver and Agent Inventory Services“. Sounds complicated – I’m sure everyone skim read that part. All he’s basically saying here is: when the server and database talk to each other, they use complex language. Probably slows them down & gets them confused. Lindens want to feed them new code so they talk simpler, which should make things faster, and I think less likely to bork.

So, lots of good improvements coming down the pipes. Hooray! More stability. Less lag. Snappier grid. Should be nice to see. FJ says they are also working on some other stuff with the Central Database, but he’ll explain that later this January, probably because he thought his post was getting too lengthy to throw on stuff about another big complicated thing.

I’ll likely post a “Cole’s notes” version right up here for everybody, whenever that post appears.