The State of the Internet Operating System

macslocum writes "Tim O'Reilly: 'I've been talking for years about "the internet operating system," but I realized I've never written an extended post to define what I think it is, where it is going, and the choices we face. This is that missing post. Here you will see the underlying beliefs about the future that are guiding my publishing program as well as the rationale behind conferences I organize.'"

Comments

Dumb terminals and smart people don't mix

[elrous0]

This whole "Internet OS" thing reminds me of the periodic resurgences of the dumb terminal/thin client idea that goes back to the mainframe days. It seems like every ten years or so, everyone is talking about thin clients in every office, with the OS and apps running on some offsite server somewhere (now with the added twist of multiple servers over the internet). Ostensibly this is seen as a good way to save IT money and overhead. But in every actual deployment I've seen, it only causes hassles, additional expense, and headaches.

Back in the 90's we tried this at my old university. We networked all our computers and put all our apps on a central server. Even though this was all done on a local network (much more reliable in those days than the internet), it was still a complete disaster. Every time there was a glitch in the network; every student, professor, and staff member at the university lost the ability to do anything on their computer--they couldn't so much as type a Word document. Now, with little network downtime, you would think this wouldn't be so much of a problem--but when you're talking about thousands of people who live and die by the written word, and who are often working on class deadlines, you can imagine that even 30 minutes of downtime was a nightmare. I was skeptical of this system from the get-go, but got overruled by some "visionaries" who had bought into the whole thin client argument with a religious fervor. Of course, long story short, we ended up scrapping the system after a year and going back to the old system (with a significant cost to the state and university for our folly).

Re:Dumb terminals and smart people don't mix

[Em Emalb]

Actually, it's all just one big cycle. When I first broke into the IT world, PCs were a bit of a novelty in most businesses. Then, the PC explosion caused things to move towards a "client-side" setup, with faster desktops, laptops and not as much horse power required on the server side. Then, in an effort to save money, tied in with servers/CPUs/memory becoming cheaper, and security concerns, companies started (or have started) to slowly pull things back from the client side and put more emphasis on the server side of things.

That said, I'm sure it won't be long before we go full-circle again.

One final thought, I do not want any "OS" that's supposed to run on my computer to be running on the internet. Corporate networks, in my experience, are typically much more prone to solid uptimes, unlike the internet. Plus, if something goes down on my network, I don't have to depend on someone else to fix it.

Re:Dumb terminals and smart people don't mix

[drinkypoo]

It's also dumb. Even if you bought a low-end Intel Atom machine, why would you want to waste that CPU letting it be a dumb terminal? Put that CPU to work by enabling it to do tasks independently even if the network connection fails.

I weep for OpenMOSIX. I was hoping that the project would continue and ere long we'd be motivated to buy all one architecture in our house simply because all the machines would form a cluster almost without our involvement and just accelerate each others' tasks. A terminal cluster where the terminals also make the entire system faster is kind of an ideal dream.

Re:Dumb terminals and smart people don't mix

[david.given]

I weep for OpenMOSIX. I was hoping that the project would continue and ere long we'd be motivated to buy all one architecture in our house simply because all the machines would form a cluster almost without our involvement and just accelerate each others' tasks. A terminal cluster where the terminals also make the entire system faster is kind of an ideal dream.

What happened to OpenMOSIX, anyway? I used it very successfully to turn groups of workstations into build servers; they all ran OpenMOSIX, and then make -j8 on any of the workstations would farm out the build to all the workstations. And it all Just Worked, and there was bugger all maintenance involved, etc. I was really looking forward to it getting mainlined into the kernel and then it just all kind of vanished.

There's no indication of what happened on the mailing list --- it just stops. There's a new project called LinuxPMI that claims to be a continuation but there's no mailing list traffic...

Re:Dumb terminals and smart people don't mix

[snarfies]

According to Wikipedia, "On July 15, 2007, Bar announced that the openMOSIX project would reach its end of life on March 1, 2008, due to the decreasing need for SSI clustering as low-cost multi-core processors increase in availability."

P or NP

[daveime]

It seems the hardest and most time-consuming problem with Internet operating systems is figuring out how to work offline.

And the easiest solution, which seems to escape almost everybody, is "don't work online in the first place".

Re:P or NP

[starfishsystems]

Not really. Your situation of working offline is a particular case of working online. It just happens to have high latency. So the easiest solution, for the user, is one which generalizes to encompass high latency.

The converse is not true. Of course you can retain the capabilities of an offline environment even after you add a wire to it, but those capabilities do not generalize to managing the resources on the other end of the wire.

The easiest solution to implement is a pencil and a piece of paper. Oh, you want capabilities too? Well, that's different.