Why Microsoft Should Fear Bandwidth

Mike writes "Microsoft should fear increasing bandwidth to the consumer more than any other single factor as a threat to their monopoly. The average user has no desire to be the sysadmin of their machine(s), and telcos and cable companies would be glad to take this task from them -- for a nominal fee, of course, as application service providers. The PC as we know it probably only has a decade or so left."

Comments

I call shens

[CdBee]

We've heard this how many times so far? The ideas been spinning around since the early 90s at least.

Repeat after me. As long as there are laptop computers there will be a strong demand for locally-installed software.

Repeat after me #2. Laptop sales have been steadily rising and will probably continue to do so.

Re:I call shens

[Anonymous Coward]

More importantly, people want to have control of their data.

Think of the following scenarious:
1) Company A hosts all the programs and data for Person B. Person B writes a document in Word 95. Now Company A upgrades Word to Word 2007. Person's B's documents look all wonky and he has to learn how to deal with a whole new word processor with dozens of new features that he never wanted. If Person B had his documents on his computer, he (or this "computer guy" he knows) could have upgraded MS Word when he dammed well felt like it (if ever) and learnt MS Word 2007 (and how to turn off it's annoying "features") on his own schedule.

2) Because hosting companies are liable for the stuff they hold on their servers in the same way corporate computers are liable for stuff on their computers, hosting companies will have to monitor what you're doing and check for "illegal content". No-one likes being watched 24/7.

3) If quality of service isn't up to snuff, moving your entire world with the exact same features will be a bitch.

4) Chances are, hosting companies will take advantage of priviledge information and forward information about you to "respectable companies" "so that they may serve you better". You can expect targetted ads to start trickling into programs. This isn't paranoia. Ad sponsorships are already embedded in movies that you pay for, on buses that are paid for with public tax money, and pay tv channels that you've paid for.

Re:I call shens

[Anonymous Coward]

Actually 1965, Multics was supposed to do this, the funny thing is that the "only" thing left of that idea is Linux and the BSD's which is far more threat to MS then bandwith.

--
Martin P. Hellwig

Re:I call shens

[SilentChris]

You know, I was toying around with that idea. Central windowing server (Windows, Linux, take your pick) in the basement. Tablets to roam around the house, connecting through Terminal Services and the like. WiFi is fast enough that this would work, and since most of these setups allow multiple terminal service sessions, every tablet would have identical screens and access to the same information.

Re:I call shens

[fm6]

The ideas been spinning around since the early 90s at least.

A lot longer than that. The concept of a simple "information appliance" has been around at least as long as the PC. That's the concept that's actually right for most people, and if computer development were driven solely by consumer needs that's what we'd already have.

But new technology isn't created solely by the market -- that just gives a massive economic incentive. The actual creation of new products is done by all those geeks and hackers who can't stop themselves from fiddling with technology.

The first serious consumer computer was the Apple II. Which was not what Wozniak had in mind when he designed it -- he wanted a system that other geeks, like himself, would enjoy playing with. So he built in concepts of openness and expandability that dominate PC design to this very day. Which means that everybody who uses a computer needs a tame Wozniak nearby to keep their computer working.

This has always been true, but Microsoft's security woes, driven by their need to fiddle and hack and featurize themselves to death, is rubbing everybody's nose in the fact. So all of a sudden, everybody's talking about "the end of the PC". And it's not a bad idea -- it's just not clear who's going to make that fundamental change. Because the people who create all the tech just don't think that way.

BTW, the rise of the laptop is hardly evidence of "the PC forever". A laptop actually is more of a "post-PC" system than any desktop, since laptops are, by design, much less configurable. And in theory, you could have a laptop which never needs to have anything installed on it. If you need new software, you just run it off a server. Not practical without pervasive networking, of course, but that's fast becomming a reality.

But laptops aren't going to be the post-PC either -- they're still designed around the idea that a computer is an open system you keep adding components to. It's just that the components are software rather than hardware. And despite all this talk of "Post PC", I don't see that changing any time soon.

Re:I call shens

[danila]

I don't think there is any sane user, who would be willing to accept that every time the WiFi LED on his laptop dims the laptop stops responding as well. Not to mention the fact this would place further tax the laptop battery.

No, this is a silly idea even for desktops and it will never fly, because local computing can be made just as good as remote and it doesn't have so many limitations.

Re:I call shens

[HiThere]

When you assume that this is working, you don't need to assume that all data ends up on the user terminal. So large downloads, e.g., won't take wireless connection speed. They'll remain resident on the main system. The terminal connection only needs to support, say, an X-Window connection. Nothing else. This MIGHT be managable over a 9600 bps connection. (I've never tried.) If it wouldn't, then perhaps a more intelligent protocol is needed. At all events, only the changes need to be transmitted, so one would certainly only need a connection faster than 9600 baud in bursts.

P.S.: saying that Wireless is half duplex it wrong. Some particular implementation of it may be half-duples. Shared I agree, unless one is using a focused beam for connections. (Low frequency Masers anyone? Do lasers exist for the RF spectrum?)

Since they are talking about a decade from now, assuming that current protocols and techniques will be the ones that are used it faulty. They might be, or the current approaches might be superceeded by something completely different. And beamed connections would certainly allow much more bandwidth than broadcast connections do, even with the requirement to echo pings so that position info is maintained current. OTOH, that could be too complicated for a portable system. But perhaps not.

Re:I call shens

[myowntrueself]

"People don't recognize 'intellectual property' people recognize tangible property."

Absolutely.

I think that theres, perhaps, a conflation between 'a property of something' as in 'a property of the sun is that it is bright' and property of something as in 'these trousers are my property'

So an idea, or a computer program has the intellectual property that it came out of someones head.

But that doesn't mean that its their property.

The above may or may not fly, its just a thought, but heres the killer of IP; if it isn't made out of atoms how can it be owned?

How can someone be said to be deprived of something just by someone else having the same idea? Theres nothing removed, nothing taken away, nothing that is in one place now which was in another place before.

Intellectual property is one of the great lies of our time.

People instinctively find it a flawed concept.

Except lawyers and politicians whose ability to reason seems limited to the debating club stlye of 'reasoning'

Wow, what aload of *BEEP*

[FinchWorld]

"this average person often has no idea how to fix the computer when it breaks, and no idea even how to perform the most basic maintenance on it to prevent such breakage. It's also vulnerable to hackers, phishing schemes, and hosts of other plagues.

With a car, for instance, this exposure to complexity is a necessary state of affairs. With inevitably increasing bandwidth, this is definitely not a necessary state of affairs for computers, and the time of the personal computer as we know it will soon be at an end, I think.

Most users have no desire to be the system administrators of their machines, and would gladly turn that task over to someone else for a nominal fee."

No they wouldn't hand it over to someone to do it for a fee when they know the likes of me will do it expecting nothing in return, even though I know its often good for £10 or a small bottle of whiskey.

Nerds us are the reason this won't work. Hell I've even started installing tight VNC on every computer I build for people know so I don't een have to bother to go around to them to fix (And for 2 people clean out there comp twice a week over vnc).

Might seem i have alot of time on my hands but all that whiskey and money is great for college students.

With Linux ...

[at2000]

we can boot the whole OS from the net with ease.

web based apps becoming very very popular/Active X

[acomj]

The popularity of web based apps (I've sold a couple for small offices) is astounding. Install one place and go. LAMP (Linux Apache Mysql Php) or java (JBOSS) makes this very convienient. Only one machine to maintain vs many installs across multiple computers. Of course if the one server fails....
At my company more and more things are moving to web based colabrative apps (Notes/ Bug tracking/ timecards..).

Active X was MS attempt to control this market by making web apps work only with internet explorer. Fortunetly it didn't catch.

Web mail is another web app that is astoundingly usefull and has driven this trend.

The main thing holding it back is web browsers are cludgy to develop real slick apps with. Javascript helps but.. Gmail is pretty decent.

Most people don't care what OS they are running if the web works and they can get what they want. Computer purchasers are very unloyal to brand names. It remains to be seen if they remain loyal to MS windows.

I can't wait

[Reality Master 101]

I have a wireless laptop that follows me around my house. It's only purpose is RDP into my home-office Win/XP machine. If the bandwidth was there, I'd love to eliminate the home-office machine and be able to get my desktop anywhere in the world from a "PC service provider". It totally makes sense -- let them worry about backups, hardware upgrades, etc.

Related to this, when is Linux going to get something like RDP? No, X11 isn't it. When you disconnect from X11, it blows away your desktop. VNC is closer, but boy does VNC suck compared to RDP. It's unbelievably slow. I know why it's slow, but that doesn't excuse the fact that it sucks.

Agree, but differently

[moduc]

M$ will have problem not because of the ISP will replace them, but because people will easily download software, and they will nolonger have as much the distribution advantage as of today (interm of bundling). If downloading OpenOffice takes 3 seconds, then users would much more likely to download it. Ofcourse, computers must also be much faster to have it install in couple seconds, and start up quicker. If not, then M$ stil have the advantage of being more seemingly integrated.

This is a good idea I think

[minairia]

I agree with the idea of a centrally managed environment controlled by ISPs. This would be perfect for older users, average non-tech users and children. If I was a parent, for instance, it would be great if our home PC was as managed in terms of software installs, web-sites that could be visited, etc. as our PCs at work.It would be especially good to have the system alert me if my children were sending e-mails or IMs involving sexual and/or illegal content with strangers on the net. My father virus infects himself about once a month, no matter how many times I warn about going to dodgy sites. I set up Firefox, but, somehow, he (and my mother) always find their way back to Internet Explorer no matter how hard I try and hide it.

Of course, such a system would need an opt out provision. I would not want my own personal use PC to be managed by anyone other than myself. I can imagine that when my kids got to a certain age they'd be allowed to use the "adults computer". I'd also be sure to make sure that, if my son or daughter developed an interest early on in IT and PCs other than just IM or music downloads that I'd give them access to an opt-out machine. Even with the risk of their being exposed to the dark side of the net, I feel it would be more important that they have a fully functional tool available to build their knowledge, if computers were their thing.

Some will say that the best way to control your kids internet access is to watch your kids. I agree, but, realistically, with the schedules we follow today combined with the nefariousness of the average teen boy in terms of finding ways to see naked chicks, dead people, etc., having the IT department of my ISP keep an eye on things would be a real blessing. Having the system prevent them from installing god knows what virus ridden dreck from the internet would save endless time spent in restoring systems, reformatting hard-drives, etc.

With the MPAA/RIAA lawsuites flying everywhere, as a potential parent, the last thing I want to find in my mailbox is a demand for hundreds of thousands of dollars because my daughter downloaded a Britany Spears song or two. (I blackly hate the RIAA but, as one guy on a budget, if they come after me, they win.) I know the risks and no ways to protect myself when using p2p networks, an average 10 year old, or an average 70 year old (my father just loves downloading movies) won't have a clue.

Thin clients?

[El Gordo Motoneta]

I've got one question: If we still have not managed to make standard thin clients popular (what percentage of the market is using sunrays, for example?) -- and these work over a LAN at 100Mbps or more, then what is making these people think that just because there's broadband, everyone will just jump over this thing? Thin clients are NOT a replacement for a desktop computer. They are, if they get them to work as good as a thin client can work, office workstation replacements. Nothing more. You can't play games or video at decent rates using remote applications. Not even over a 100Mbps line. Not even using Citrix. I've worked with citrix myself and while it's performace over a low-bandwidth link is impressive, the look and feel of a remote application is not anywhere near perfect at 100Mbps. It's just not the same. So, *maybe* some big company will start offering this service. My prediction is that a few people will subscribe to it, use it for two days at the very most, and go back to their "primitive" local applications as fast as they can. But i guess we'll just see.

It's mine!!

[nodehopper]

I don't see, even basic computer users, wanting to give up local control of their computers. My IT department is looking to roll out a Terminal Server as a way of saving IT budget. This will run the OS and applications on Thin Clients from a centralized Terminal Server. Many of the users immediately balk at losing control of their local computer. Even those who aren't very computer literate. It is just a normal human reaction to someone taking away control. Into this add the current distrust of anything being done over the internet. How many people do you know who refuse to do "X" over the internet? With "X" being: home banking, shopping with a credit card, give out personal info, etc. I know quite a few and they are mostly the less informed users. I understand what an SSL encrypted connection is and basic internet security where as average users don't. This entire concept just goes against too many facets of basic human nature to take off. In my opinion people would rather have a spyware infested mess of a computer of their own that allow some one from outside to take control away from them

Pressure Microsoft to Make Windows Hard to Manage?

[Spinlock_1977]

If service providers offered configuration management services, would they then pressure Microsoft to, when adding features, make those features more amenable to provider-based management rather than end user based management?

Does Microsoft invest in any bandwidth providers? Should future investments in this direction make us nervous?

What about the mono-culture problem? When a provider applies the latest patch and clicks the wrong button, will a million PC's get trashed? Who's going to visit all those homes? Will grandpa have to wait until Microsoft and the provider duke it out in court before someone drops by to re-install Windows?

Right outcome, wrong reasons

[Anonymous Coward]

Although I agree with some of the premises of this article, I think he has completely missed the boat on the reasons for it.

First off, we can already see the movement he describes. How many ISP's (particularly the big ones like AOL and NetZero) now advertise e-mail filtering and virus blocking? Microsoft has done such a poor job on it's software that ISP's had to move agressively in this direction or risk having all their mail servers and bandwidth pipes melt down!

In fact, Microsoft has done such a poor job that one is tempted to think that it couldn't possibly be sheer incompetence, it had to be part of some strategy on Microsoft's part. I think it might have been. Microsoft doesn't want secure computers! I think that, in the beginning, the entire security issue was going to be used to force a move to Microsoft's Palladium (or whatever they call it today!) and move final control of every computer into Microsoft's hands.

But their strategy has backfired! Now that ISP's are taking responsibility from their users for running virus and spyware controls, how much of a leap is it for them to start providing application software, also? They had to upgrade their servers to provide extra horsepower for these applications, how much of a stretch is it to use that same horsepower to provide server-based applications, too? And I think Microsoft sees that danger now, though maybe a little too late. Their recent acquistion of virus software and announced plans to offer their own virus control software is, IMHO, an attempt to wrest control back from the ISP's and return it to desktop software that Microsoft controls.

This also is doomed to failure. Microsoft can do no better at releasing patches and updates to their virus software than they can to their OS and applications software! Many, many times in the last 5 years, I have proven to myself and the companies I work for that intelligent software at the firewall and at central mail servers can be used to protect Windows software from dangerous viruses, suspicious websites and nefarious e-mail attachments easier and faster than Microsoft's patch and update cycle. ISP's can do (and now are doing) this, also.

In short, I believe that Microsoft has done this to themselves, first through shoddy QA and then through deliberate mishandling of security issues. They should be allowed to reap exactly what they have sewn!

Not just APPS, but the OS

[nurb432]

They could start leasing the OS to you as well. And if you dont pay up, ( and stay online .. ) then your pc no longer even boots.. It just sits there, waiting for the TFTP server...

I could see a blend...

[zogger]

...of thwe two ways of computing and networking more than one or the other. For a lot of purposes, the way things are now, for special apps you only use once in awhile and are bandwith hungry, then a remotely run app. And we also have to contend with convergence in the area that is experiencing more and faster growth, and that is the cellphone/pda market. Consumers are replacing those devices a lot faster than either desktops or laptops, and the price is dropping faster when you factor in features in these small wireless computer thingees people tote around. You can't even hardly call them just phones any longer.

Next compare cost of consumer bandwith as a ratio to cost of IC chips in general. With printible ink based circuits coming online in a big way soon, it will be just so cheap to always have an advanced system that people would still want "the power" of having their own "computer" as opposed to someone elses computer.

Now, I could see your premise taking off more IF a lot of the major players and governments combine to end the "wild wild west" phase of the internet and require a good deal more in the way of identity and accountability online, and chop the internet up into subscription models a la AOL type "nets" where the consumers would pay for a package of apps, games, delivered on demand entertainment and information resources, etc., and in competetion with other nets, much like you buy a cellphone package today or satellite or cable TV package, etc. But that's a big wild card. I know they would *like* that as it would mean a guaranteed revenue stream, it remains how much lobbying and political pressure the big guys can put to it to institigate such profound changes.

Although the personal "system administration" angle is quite complex for the average user, automatic updates that can be pushed to them along with more secure design are the obvious trends now, so I see that problem getting easier in the future. People who want such systems have them now, and word of mouth and pressure from business desktop deployment will make it trickle down to the home owner level.

And there's one more thing to consider, and that is the "blue collaring" of the personal computer. They are merely little machines that take nothing more than a simple screwdriver to construct, because of this, we have the population now with millions of "shade tree mechanics" who are as comfortable with computer repair as one or two generations ago were as comfortable with a car tune up. It is no longer the leet high paid IT professional locked away in obscure academic or corporate R&D labs who can muck around with computers, either on the hardware or software side, it has become ubiqituous across the board in the general population. As computers have grown more complex and "hard", they have also become much easier. A person now can take a dozen boxes and connect them wired or wirelesssly, boot from a pre made Cd and have a mini super computer up and running in no time, and that is at the *hard* side of personal computing. One decade ago that would have required some pretty advanced skills, a lot of money and some pretty good luck to pull off. Single system admin has now become mostly a no brainer with the proper operating system and just a scosh of forethought, and it has the potential to be automated a LOT more. And with huge RAM becoming more and more common, you could see just RAM images of the OS and apps being the norm to run in,not hard drive based, and any major disaster being easily recoverable then, just poof it away and reinstall from known good, as simple as popping in a disk for a few moments, or as you point out, from the network. After playing around with various live cd based distros you can see the potential there, both in ease of use and in security and in administration.

Either he's insightful or he's not.

[billstewart]

Being a respected figure at a reputable industry source does add some credibility, mostly because editors sometimes make sure fact-checking gets done and basic writing skills get used, and a certain amount of Darwinism gets rid of many of the less capable and less insightful writers, but so what? It's an open-source news industry these days, and if some blogger says something insightful and interesting, it makes sense for Slashdot to pick it up, and if some well-respected pundit at a reputable trade rag says something lame and uninteresting, it makes sense for Slashdot to ignore it (unless somebody's writing an article about how lame most of the industry mouthpieces are.)

In this case, I don't think the article has much depth to it - the main concept is appealing, but I don't see enough thought behind it to really win. But even so, I'd mod you -1 Flamebait :-)

The Author Is Misinformed

[JesseT]

The ignorance of this news item astounds me. Thin-client/server architecture is nothing new; it's been in use for well over a decade. Such architecture allocates presentation to the end-user on the client side, and application processing and data management on the server side. This is essentially what the article is talking about. However, I find the fact that the author believes Microsoft to be completely adverse to distributed computing and remoting of applications to be somewhat offensive. The author clearly lacks a solid background on subject.

Sure, Microsoft certainly may not have been involved with the industry push towards distributed computing initially. But over the last decade, they caught onto it. They became involved with the OMG (Object Management Group) and the introduction of CORBA (Common Object Request Brokerage Architecture). Later, they created DCOM (Distributed Component Object Model). As web-services and standards like XML and SOAP began to surface, they two became supported (albeit, perhaps reluctantly at first).

The .NET Framework (of which the framework/base class library is a subset of the new WinFX API, which replaces Win32, and will be the defacto API for Window Longhorn) has integrated support for object-level remoting built directly into the language runtime. Standards like XML and SOAP are put at the forefront. Web-services aren't something just tacked on as an afterthought, they're given importance in the .NET world. And I think that Microsoft is finally learning not to twist such standards, they realize the impact it can have on developers.

Of course, only the future can tell whether or not Microsoft's Windows family of operating systems will survive the impending ubiquity of thin-clients. Or whether thin client/server architecture will be embraced by the masses.

Regardless, I don't believe Microsoft is fearful of bandwidth. I think they're embracing it. And as far as I know, GNU/Linux isn't any further ahead of Microsoft, in terms of technology, when it comes to this type of architecture.

Re:I'll believe it....

[ScrewMaster]

I agree. The reason that consoles don't screw up is because a. their code runs from read-only media and b. the quality control standards that game developers are under are just incredible. My brother works for a large independent game development house, and let me tell you, Microsoft would do well to apply some similar QC standards to Windows. If they did, complaints about Windows reliability would go away. Put it this way, when you release a game program on a cartridge or CD, there is really no effective way to update the product after the customer has bought it. You simply have to get it right the first time. Kind of like the space program. Microsoft knows that even if they screw up they can always put out a fix on WindowsUpdate.

I also agree with you about the mirroring, from a reliability perspective. However, I've set up a number of systems for individuals and recommended mirroring, and the attitude is "That's great! Now I don't have to worry about backup!" Then I have to spend a half hour explaining that a mirror is not a backup! When I'm done they still don't believe it ("But, it makes a copy of everything I save! Isn't that a backup?") Phooey. The other problem is that you generally don't want people to have to go into their BIOS to rebuild the mirror ... odds are they'll copy it the wrong way. But I'm sure that a floppy or bootable-CD solution could be developed.

Bandwidth is not the issue

[metamatic]

I'm the author the article. The "network is the computer" was a false start because the bandwidth was not there.

No, it was a false start because it was a dumb idea.

People don't want to pay subscription fees for software. If they did, we'd see a ton of software being sold month-by-month, with remote activation via Internet. There's no technical block to doing so, and there hasn't been in over a decade. The problem is that whenever someone tries it, nobody outside of the business world is interested.

People don't want to be at the mercy of the cable company or the phone company. We're talking about the two companies the average person probably hates most, and now you're offering them a way to make their entire computer system totally dependent on the whims of the corporate behemoths they hate?

People don't want ever-increasing prices. Look at how the cable company jacks up subscription rates several times a year. Who wants that for all the software they run?

Network connections aren't reliable enough. Ask DSL users if they want their entire computer to turn into a doorstop every time the DSL is slow or out.

People don't want the upgrade treadmill. If you buy your software by subscription from an ASP, you get upgrades when they decide. And of course, the upgrades may break things, make your PC slower, or even outright fail to run. That's why people don't upgrade their OS, don't install new Windows patches, and don't upgrade their applications. They've been burnt before. If it ain't broke, they don't want it fixed.

Computers aren't fast enough. Thanks to the ever-increasing bloat of software, editing a text file today is slower than it was in 1987, when my 16MHz Atari ST system could smooth-scroll (pixel by pixel) at 64 lines per second running Tempus on a large soft-wrapped text file. My Linux box can't even seem to line-scroll that fast in vim. Hence, there's always a need to make PCs faster, and given a network computer, the easiest way to make it a shitload faster is by adding a hard disk, installing the software locally, and removing the network latency delays.

In short, the minor benefits of Network Computing don't outweigh the enormous costs and liabilities. It isn't going to happen in a free market. It only happens (sometimes) in business because PHBs impose it on everyone regardless of cost/benefit analysis.

Bandwidth is probably *good* for MS

[Control-Z]

If MS keeps behaving the same way they have been, they'll welcome more bandwidth. Look at XP product activation. People have pretty much put up with it. That is a step down the slippery slope of losing control of your own personal machine.

If remote system administration is going to be a trend, I'm sure MS will be at the front. They'll either be there first or wait until a big player emerges and buy them out.

Re:I'll believe it....

[tomjen]

Just because you dont understand computers you are not stupid.

That beeing said i am kind of attracted to the idea of a internet for real geeks.

At the very least is should kill email worms.

Re:web based apps becoming very very popular/Activ

[Valdar729]

I took your approach initially as well. I went crazy building web apps and people loved it. Built it on a subscription service so everything was on my servers and not at their local office.

But now I'm reverting back to desktop based because of external hardware requirements (barcode readers, picture scanners, thermal printers, etc.). Also, responsiveness of the application and offline connectivity are major factors (some businesses run off of modems or shoddy internet connections.)

On top of this, Smart clients (not rich client or thin client, but smart clients) are much better than the web based approach. With .Net applications can auto update themselves so there is no maintenance and only a "single server" to maintain. With web services data transfer is incredibly easy and the .Net application blocks make everything even easier.

Lastly, I can cut down on my server farm expenses because it moves the processing from my server to the client for searching, processing reports, etc.

Re:I'll believe it....

[Anonymous Coward]

one aspect you seem to forget, that's ok though .. most people do :). That is the amount of luck and or experience you have with a specific OS is totally your experience. I've read hundreds of horror stories about how XP's SP2 it a total muck, That's not true in my experience. I installed it several times on several PC's and it has yet to do anything abnormal at all. On the opposite hand Win 98 and I don't get along at all, the best I've done with it is it runs for about an hour and then crashes.

Another aspect that is often forgotten is the total number of Windows PC's as opposed to Linux or Mac. With the 95%+ market share that MS has comes the vast volumes of reported issues. You need to take the percentage of problems .. not the volume. Example would be if I distributed my own OS to 100 people and only 1 has a problem. then that's 1% .. now what is MS's 1%? ... Millions.

Re:I'll believe it....

[ScrewMaster]

Well, mirroring is simple, but the problem is that when a mirror desynchronizes it's not always due to a hardware failure. A defective write cycle (even a power line transient, I've had that happen) can cause the two images to differ and one of them will be deactivated. The problem then becomes one of determining which of the two apparently good images is the bad one before running a rebuild. Both may be accessible, both may even be bootable ... but one of them has bad data. I have one machine with a Highpoint controller (I know, not the best RAID chip) but the BIOS is pretty slick ... it will recommend which way to run the rebuild but I can override it if necessary. A couple of ASUS BIOSes I ran into (Promise controllers) insisted that they knew best, even though I knew it wanted to copy a completely unreadable drive onto the good image. Basically they need to put some more thought into RAID management before it will be suitable for the masses.

A proper backup is one that is physically removed from the computer, and ideally stored offsite some distance away. A mirror is not a backup, period ... think of it as a single drive that is just more reliable. If you save a file, and your application writes corrupted data you've got a bad file on both images and are screwed. That's not a backup, let alone the fact that if, say, you suffer a catastrophic failure (lightning strike, bad power supply, you name it) you've lost everything anyway. The mirror won't save you.

But you're right, for a lot of people it's a lot better than nothing. I put together a machine for a guy at work (actually, for his 80-year-old mother) and when one of the WD's died I was able to bring her back up with a minimum of fuss. But she couldn't have handled that on her own. If one drive dies, and you don't know how to a. replace it and b. restore it from the backup image, the mirror isn't as useful as it could be.

The Computer As A Freezer

[ellem]

I say this all the time. I used to believe everyone should have a computer, then they started calling me and I quickly realized very few people should have a computer.

However, the applications a computer runs are very good and very important. Email, Web Surfing, some data/word processing are all terribly useful for the average person. If only the damned computer weren't in the way!

Palm has a really smart way to deal with that by limiting any interaction with the OS and making the App king. Plus having everything running all the time makes everything faster.

The smart money is on going BACK to mainframe type applications and computing. Java (etc) have been invented so what's the wait?

In Britain, I'll believe it when I see it ... :(

[Kaemaril]

Increasing bandwidth? Great, except that our beloved telecom chaps seem to regard the concept of "always on" computing as the spawn of satan. It seems that if you actually buy into "always on", they'd prefer it if your always on computer only used a couple of meg a day or so. Forget about backing up your PC to a remote box on a regular basis, or something like that :)

The introduction of caps on broadband (1GB a month, 15GB a month, xGB a month whatever) doesn't really gell with their advertising (yay, watch all the movies you like! Videophone your parents! Send your buddy streaming video from your wedding) and yet they will insist on it.

So I'm not too worried about increased bandwidth ... 'cos our telcos are very keen on taking it away from ya :)

Kids these days...

[dutky]

Does nobody remember the origins of the personal computer?!? Back in the mid-seventies most computing was done in business and institutions on terminals connected to large, centralized computing systems: time-shared mainframes or mini-computers. To a large degree, personal computing was a backlash against these centrally controlled systems and the managerial structures built up around them: system administrators who decreed what software would and would not be installed, billing systems that accounted for every fractional second of computing time, computer operators who controlled which users jobs would run and when, etc. Inexpensive, single-user computers (starting with mini-computers like the PDP-8 and PDP-11, but continuing with the Apple II and IBM-PC) represented a revolution for end users control of their own systems.

Every decade or so, since the rise of the personal computer, we see some attempt to re-impose the rule of centralized systems, usually under the guise of 'easing the burden on end-users' but always including an increased financial burden on those same end-users. The simple economic facts are that computer power (by any measure: instructions per second per dollar, main-memory bytes per dollar, on-line storage bytes per dollar, etc.) has become so inexpensive that all the old reasons for centralized computing systems no longer apply (and haven't applied for at least 20 years). The only reason these new centralization schemes is to find some way to extract money from existing computer users, whether or not the users actually want the sevice being provided. The idea that people will willingly give up control of their own systems and pay for the privilage may be a wet dream for companies hoping to collect the money, but it doesn't sound like a very good business plan.

The solution to the increasing administrative burden on computer users is not hire someone to do the administration: instead, we need computers that actually reduce amount of administration required or make the task of administration markedly easier. This is what personal computers did 40 years ago, and it can be done again.

Re:The big get bigger

[Shakrai]

Microsoft is as 'big box' as they come.......while there is no doubt that strategies behind operating systems and the internet will meld together I don't see it as a reason to see Microsoft to not be a prominent part of that.

Yeah but if the article is correct (disclaimer: I disagree with it; but for the sake of debate...) would Microsoft still be as heavy of a hitter as they are today? How much of their revenue comes from new OS sales again? In this area they seem to be a victim of their own success. Why do you think they released ME instead of going straight to XP (or just marketing 2000 to end-users)?

Could Microsoft adapt in time to take over a new market and continue to own the PC industry? Perhaps. But historically they have responded very slowly to new threats. The only thing that saves them is the control over the operating system and desktop.

Re:I'll believe it....

[Maestro4k]

• What should click in their brains? That they should only allow users to operate on one piece of proprietary software/hardware, and never ever allow them to upgrade? I'll set up a windows box for you, and i'm betting if I dont ever let you change it in any way - it will still be working just fine many years down the line. Just a hunch.

You're dead on there, and this is why Windows problems are more on consumer machines than well run business ones. I worked for one department within a university a few years back as the Sysadmin. I was basically it for tech support in the department, and while having to run policies through faculty committees, I basically made the rules.

As we went into the first computer upgrade cycle for staff machines since I had taken over, I got the policy changed to remove admin access from all staff (not faculty mind you, that's a lot trickier to pull off, but this was a starting point). Of course they weren't happy about it, having had admin access for years, but the policy stuck. As their new machines were configured, I went to a lot of trouble to make sure they had the apps they wanted (as long as they weren't spyware or other stuff that had no place on a business computer), and that everything worked for a non-privledged domain user login.

After about 2 weeks, the complaints dissapeared. In a month, it switched to compliments and increased productivity. Our two main department secretaries went from having to have viruses cleaned off weekly (this was even WITH Norton running managed) and daily reboots to basically never having to reboot the machines at all.

So for those that don't believe a windows machine can have good uptimes and work well, I've done the case study already, and two of the folks in the study had previously exhibited an absolutely stunning ability to fubar a machine beyond all belief. When they didn't have the privledges to do that anymore, the machine worked just fine, AND they got more work done. Everyone was happy.

Now personally I have to reboot every couple of weeks at least, and do a reinstall every year and a half or so, but I'm brutal to computers. I run on average a good 20 apps at once, switching back and forth as I need to. I do this on everything, not just Windows, and frankly Linux, the *BSDs, and even Solaris can't handle me. I'm still the only graduate of my University's CS dept. who managed to crash the Solaris lab machines without any root privledges, and I did it more than once. I even have to reboot my linux machines at least monthly.

Bottom line, no OS can handle either idiots (who continously click on the same damned E-mail attachements to get reinfected over and over) or major multi-taskers very well. They're not designed to be abused, when they are, they respond in ways that weren't anticipated.

Who Eats the Energy?

[danila]

Check out Who Eats the Energy? study. A HDD eats about 2-3 Watts when in use and 0.5 Watts when idle. A wireless card eats 1 Watt in base idle mode (less than in power saving mode) and 2-3 when transmittin/receiving. With a network computer it would be working 100% of the time, while a HDD would be mostly idle. So your assumption is not true. Also note that notebook power consumption in that study was 11-16 Watts in total, so switching to network computing mode can decrease battery life as much as 10-25%.

Re:I'll believe it....

[quistas]

PCs with two drives won't be offered because it's still cheaper to make a PC with one, and you can't convince the average user that RAID is an expense that they want to pay for.

Don't think of it from a user's perspective. If Dell can offer desktop systems that are dramatically more reliable and so reduce the number of troubleshooting and service calls, that's not only a marketing advantage ("Runs forever!") it's a support advantage.

Now, the cost per unit to put RAID-1 in is still going to exceed the reduction in support costs, but there's a point where the marketing advantage combined with that cost savings makes it economical.

hold on a sec

[mnmn]

The internal bandwidth of machines will always be faster than the externel pipes. The cpu - memory speed will always be faster than your connection to the Internet.

All that means local applications will outperform hosted apps. Given applications will always push the limits, the execution of most graphic apps, and apps that require more interaction than is possible through a terminal services screen, will always be slower from a remote station.

That and our tendancy to OWN everything onto our desktop, similar to getting satellite dishes than pulling a cable and being at the mercy of cable companies. If most desktops are laptops in the future, its hard to believe any procssing will be offloaded killing the mobolity of the laptop.