As some of you may have experienced today, there was a massive slowdown of the internet in areas that linked in with google at pretty much any junction point.
Some websites were speedy and fine, but those linked to google were so slow as to be unusable. This takes me back to many conversations I have had with friends in my tenure in the technology industry about Networked Computing (or Centralized Computing).
What we have here is a nice gentle example of why a centralized computing nexus for the world is a really boneheadedly stupid idea. I don’t mean the internet of course, but what I do mean is a resource on that internet (a decentralized updateably structured system) that is linked into by pretty much everyone for core functionality.
As should be supremely obvious today, we have a problem. What would have happened had google gone down? What would happen in ten or twenty years if google goes down? or… What would happen if google became the eponymous Skynet of our worst mechanized nightmare?
To put simply kids, here today is your wakeup call to what happens when one service or collection thereof is not just used, but relied on, by nearly everyone.
To all of you who wonder why you should have a personal email address as well as a google one, your question has just been answered. To any company out there considering the use of google as a major thread of your IT backbone… stop now and consider what happens not if, but when, google is interrupted for a length of time (choose your length here).
The bottom line here is simple. Is google pretty damn awesome? Yes. But at the end of the day people, open your eyes and understand that maybe there is something to ‘DeCentralized Computing’ after all.
The Cloud is nice, the cloud is your friend, But the Cloud is not your answer to applications, the cloud is not your answer to permanent storage or email or the host of your blogs or websites. The cloud is your friend… till it slows down, turns off, or gets taken over by something smarter than we..
Then the rest of us with our decentralized outposts will keep on ticking, taking advantage of productivity whilest the rest of you hope pray and wait on some other folks, some other company to get you up and running.
Chipmakers, Video Card makers, and everyone who produces hardware that is even remotely thinking about cloud based computation for graphics or an e-topia where we’ll only have access points to a computational cluster ala william gibson’s Cyberspace, please please remember that when the cloud turns off, And it will, you’ll still need folks with something more than a dumb terminal to get it back going again.
Keep your PC’s people, Keep your applications. Demand a seperate downloadable, home useable application for everything that exists in the cloud. Someday you, your company, your next dinner, your next shipment of much needed medication may, in fact, depend on it.
And for those of you who think I’m screaming at heaven and am a resident nutter, just know that next time you can’t access your gmail or your favorite site is down… I’ll be chatting with others, giggling at the mass hysteria.
So, welcome to the world of the here and now and say hello to your new machine overlord, Google, be thy name.
4 Comments
Which raises the issue of Second Life vs. OpenSim. While so far nobody’s life depends (I hope) on Second Life, we are at the mercy of Linden Lab and their centralized databases. Does that give OpenSim any potential competitive advantage? Only if assets can be made portable, and it looks like the law will have to change if that’s ever to be made possible.
Entirely right. Second Life is a walled Garden, a very large walled garden, but a walled garden nonetheless.
I think that for the way SL operates an asset server cluster makes sense, but there shouldn’t simply be one asset cluster, Real time backup versions should be placed in data centers around the country that can be switched to immediately on a hiccup and they should never be used simultaneously.
Your right though, assets need to be portable. And really the simplest and most effective solution would be for Linden Lab to release their Grid technology for open use but tie them into the asset cluster that they control with the overhead on the cluster side for access rights/permissions (So that no one can full perm everything just by setting up a sim region at home).
You still run into the one company being in defacto control, however its far more stable than today’s method. You’d also have exponential growth of Second Life that it becomes a standard across the internet and in future iterations it might be possible that some folks would set up their own asset servers or have a system of asset server nodes that might not necessarily be owned by linden lab, but updated by them. And if something hellish were to happen, that could then be opened up without permissions being opened.
Alot of theoreticals to be sure, but its not really much different than the way it operates with the internet backbones to date. In fact Asset Clusters might even be incorporated into the routing system at the high level, like Cisco Routers are.
The ideas are there, we’d just have to see allocation to those ideas.
Hadn’t considered opening the asset server. However, copyright law still kicks in; we’re copying the asset to the simulator, then to the viewer. So not only does the end-user need a license, so does the simulator owner. The problems here aren’t technical, they’re legal. Of course, nothing precludes Linden Lab unilaterally changing the terms of service again (and probably in a ham-fisted and disruptive manner, so as to retain their patent on that business method : ), but creator outrage in that event would be justified — Linden Lab would be unilaterally taking assets intended for use in SL and distributing them beyond that. I’ve personally uploaded assets to SL whose license would prohibit using them under an architecture such as you describe.
legal hurdles are what they are, legal hurdles and certainly we’d have to agree to some method whereupon the assets remained owned and avialble by us to those we wish I agree. But some method as described would need to be in place to obviate the walled garden centrism.