Google and Their Server Farm

JR writes "CNet has a very interesting story about Google, operating systems, and where Google may be going. The upshot is that they may make OS issues totally irrelevant by supplying everything anyone needs over the web from their mega-server-farm."

Comments

Her Pie-in-the-Sky Dream is What?

[filmmaker]

"Ajax, which is short for Asynchronous JavaScript + XML, combines JavaScript, dynamic HTML, and XMLHTTP to, in essence, let you build Web-based applications that run as quickly and seamlessly as local software."

Great. If the author of the article gets her pie-in-the-sky dream, the future of virtually all client-side computing will lie in the hands of javascript code. For certain applications, like ones with small, text data sets, a system like Ajax could "feel" like a desktop application. The bandwidth just isn't there for video or even industrial photo work. I wouldn't want to run a batch script to modify 5,000 images in the Ajax analog of Photoshop. Better not be a fiber network without any limits on network transfer.

Besides that, who wants anything but light-weight or at least, non-critical, data and applications to be out on the network. Gmail is a perfect network application, but my financial software or any number of other things? No thanks.

Re:Her Pie-in-the-Sky Dream is What?

[Alan Shutko]

Besides that, who wants anything but light-weight or at least, non-critical, data and applications to be out on the network. Gmail is a perfect network application, but my financial software or any number of other things? No thanks.

Online tax software has proven to be very popular over the last couple years, so not everyone shares your qualms.

Re:Her Pie-in-the-Sky Dream is What?

[Kaa]

Today, think of the benefits from PC virtualization:

Ooookay, let's see...

compiling would be done over a huge grid of computers,

People who compile will have their own computers for sure. Isn't the general consensus that the everthing-is-a-Google-web-app world is for the unwashed masses? :-)

video games would be faster because the client/server communications barrier would no longer exist (well, it still would exist, but it'd mostly be sending images to the user's computer, and then the user sending short commands back),

ROTFL. Welcome to the world of X Window, VNC, and remote displays.

But let's check if the games would be faster :-) Let's say the game runs at 1600x1200 resolution. That means a single screen is 1.92 megapixels. Each pixel needs three bytes of RGB data, so that's 5.76Mb for a single screen. We want to have 60 fps for twitchy games, so we need the bandwidth of 5.76 * 60 = 345.6MB / second which is around 3.5 Gigabits/second. A dedicated OC-48 line (2.5 Gbits/sec) won't cut it, we'll need at least OC-192 going into each house ('cause more than one person might want to play games simultaneously).

Yeah, definitely, this will solve all the network lag problems...

all your data would be automatically backed up and secured,

Until the rats in a warehouse in Calcutta chew through the backup tapes...

and the world would have less environmental damage due to outdated computers with lead parts.

Umm.. what would be that thing that talks to Google servers -- the one with the screen, the keyboard, the network interface, the video chip, the sound chip, etc. etc.? Maybe a computer?

Re:Her Pie-in-the-Sky Dream is What?

[aixou]

We'll see an Audrey-like Linux Box with a Firefox and nothing else and it'll be called a GoogleBox. You can do your e-mail, web browsing, photo organizing, document writing, and music work on this box and you never need to run scandisk, install AV software, deal with adware, etc. etc. etc.

Plug into your cable modem and go.

It's not what I need or you need but it's what most people need.

Absolutely not. You highly underestimate the average user if you think their computing needs will be satiated so simply.
Such a box would face that same problems that the "other OSes" (i.e. non-Windows) are today. People can't just walk into the store and pick up a game or other Application and use it. People can't walk in to the store, buy a scanner, and expect it to work.

There is too much talk about the mythical user that only uses checks their email and browses the web. As far as I can tell, this user is the exception rather than the norm. Real users use all of that and more.
It's like the saying about Microsoft Office, that even though few people use more than 10% of the features of Office, everyone uses a different 10% and thusly Microsoft can't really cut out the bloat without pissing a fair chunk of users.

It's the same way here. Everyone may use a minimal amount of software, but they all use different software, and to try to fill their needs with such a simple box is ludicrous.

Regardless though, what makes google so special that anyone should trust their entire computing experience to them? I thought computing monocultures were a bad thing in general. Why is it OK for google to have more control over a user than Microsoft ever had?

If I trust my computing experience to a web-based system, I am trusting it to too many fault points for comfort. What happens if the web goes down? Google gets hacked? DNS server goes down?

There are just too many dependencies in such a system for it to ever work (dependence on your net connection, that google will continue the service, that hardware makers will support the box. etc etc)

Brilliant

[tabkey12]

Thin-Client computing by another name, again. Wasn't convinced 20 years ago. Still not convinced now. I don't want to have a useless PC just because I stopped paying the $20 a month subscription to the applications.

Re:Brilliant

[xtracto]

Now, worst than that. Imagine what will happen if for some, any reason your data is lost in their servers.

Of course, as the EULA will state, the service goes with no warrantay and AS IS. So after that you will just be screwed.

And there you have another point, I sincerely preffer to buy a house than to rent it, if I rent software, they will have me grabbed-by-the-b4115 until I die, and surely DRMd in some way. It is similar to iTunes, once they grab you, you pay, or scream...

Sincerely I think that approach is just useful as sun approach, for "processing" tasks, no information storing or "application rental"

Not going to happen anytime soon

[dtolton]

Seriously if I had a dime for everytime someone predicted the demise
of the desktop, I'd have a couple of bucks.

Here is the problem I have with her theory. Her points were all
logical and well laid out, essentially that most people aren't system
administrators and that they don't back their data up, don't secure it
etc. While that is true, it doesn't necessarily lead to people giving
up the desktop in favor of a thin client. Giving up your desktop is
an emotional decision, and there are a lot of factors that weigh
against that.

In the long run, maybe ten, fifteen or even twenty years in the
future, this type of service may be much more prevalent. But I don't
think something like this will change over night. Think about how
much computer systems have really changed in the last ten years. Not
that much if you really stop to think about it. What she is
predicting is a *massive* paradigm shift to say the least. Microsoft
didn't have the clout to pull it off, probably because no one trusts
them enough. Do you trust Google enough to give them *all* of your
data? I'm not sure I trust *anyone* that much.

Re:Google will likely try to do this.

[micromoog]

In the future, the network will be just as dependable as any other public utility. When "the network is down", people will treat it just like when the power's out today.

Using someone else's computers....

[thewiz]

I don't know about other folks in general, but I do know that I like my privacy. I'd rather have a computer on my desk, behind a firewall, where I can keep my private information private. It's all well and good to say that storing your data on Google or Yahoo or MSN allows you to access it from any computer on earth, but you run the risk of the computer you are at copying the information you access.

Wether it's a malicious keylogger, trojan, or simply the paging space / file, your information get copied to the PC at the internet cafe you are using. Suddenly your private information is no longer private. Any savvy computer-literate person could access that copy of your data. Give me a laptop or desktop where I can encrypt the data and only I have the decryption passphrase any day.

Slow Down

[rokzy]

>they may make OS issues totally irrelevant by supplying everything anyone needs over the web from their mega-server-farm

yeah, try that line again when 90% of their stuff isn't (USA + Windows only) and/or beta.

Re:Google will likely try to do this.

[jarich]

In the future, the network will be just as dependable as any other public utility. When "the network is down", people will treat it just like when the power's out today.

Uh huh...

http://today.java.net/jag/Fallacies.html

Essentially everyone, when they first build a distributed application, makes the following eight assumptions. All prove to be false in the long run and all cause big trouble and painful learning experiences.

1. The network is reliable

2. Latency is zero

3. Bandwidth is infinite

4. The network is secure

5. Topology doesn't change

6. There is one administrator

7. Transport cost is zero

8. The network is homogeneous

ASP?

[14erCleaner]

Hey, maybe they can become a big Application Service Provider!

Oh, wait, that was two buzzword generations ago. How many words are there for "mainframe" anyway?

It won't work and here is why...

[dfj225]

I don't think a plan like this will ever gain acceptance more than a small percent of computer users and here is why:

- The first thought that came to mind is business. The company I currently work at would have a heart attack if anyone suggested using a thin-client like solution with Google storing all the data. So I guess Google might sell their technology (like they currently do with their search servers) but this really wouldn't be any different than buying a file server and desktops.

- I don't see bandwidth getting fast enough in even 5 or 10 years to support a video or photo editing app. I can't even imagine having to upload a whole DVD's worth of video to Google before I could start to work with it.

- Another similar point would be application load time. Google Maps and other Axis based technologies load and run fast because there is a relatively small amount of JavaScript being sent to the browser. Could you imagine something the size and complexity of Microsof Word being sent to your browser everytime you wanted to edit a document? I think something like that would bring any browser to a crawl.

- What about customization? I like to be able to install new software on my computer. The few times I have had to deal with shared hosting for websites, it has been annoying that I couldn't install new software that I wanted to try out. Especially when my host had outdated versions of something like PHP or MySQL.

So, those are my thoughts. The only crowd I can really see this appealing to are the WebTV, just surf, email, and edit docs crowd. They might be really happy not maintaining a computer and having their data available anywhere. However, I think a small portion of computer users would fit into this category.

Personally, I would much rather just use VPN to access my home shares while on the road than have to use some sort of thin client.

What Google or someone else should really do is create VPN software that is easy enough to use that anyone can set it up. I think that would appeal to many more people than a thin-client. Plus as hard drive space gets cheaper and cheaper, it shouldn't be an issue to have the same software installed on your laptop as your desktop.

Re:Not surprised

[danheskett]

Except that people can easily buy enough power to satisfy their needs for a small premimum on top of what a terminal costs. Look at around at the so called thin-clients available. Even the thinnest of them has enough power to be a "fat-client" with substantial processing power.

Add on top of that people have routinely rejected thin-clients. Bandwidth and latency are big problems. I expect acess to my files and data with low latency. That means viewing my 8MB digital photos without waiting for part of all of it to come over a wire. I expect it to be available to me all the time.

Google is great, but Google is not above the law of physics. People - just average users - have 20 or 30 or 40 or 80 gb of data on their PCs. No matter how great Google gets, providing this amount of data quickly, securely, with low latency and high-availablity will prove out of reach. Even with Google's highly skilled team of programmers, making a decently response web-mail client, or map tool is a pain in the ass. And it's still below par. Despite how great Gmail is, it's not nearly a rich as Thunderbird or Outlook 2003.

Monopoly?

[MikeCapone]

Since when are we happy about monopolies in the making? Google is cool now, but can we trust them to stay that way indefinitely?

Well, it's not done yet and they still have competition, but I'd feel a lot better if these next generation things that are supposed to be used by the whole internet community were open and democratic like Wikipedia and not close and proprietary - however cool they are - like Google.

Re:Not surprised

[BobTheLawyer]

I don't agree.

By publishing your website, you are granting an (implied) licence to the world to create cached copies of the website. Were this not the case, your web browser's cache and your ISP's proxy server's cache would in constant copyright violation.

The argument Google would use is that they're just going a step further in having a publicly available cache. Whether the implied licence extends to this is arguable: I have no special knowledge of US law but under English copyright law they have a pretty good case.