Now's the time for the network computer

Now's the time for the network computer

Summary: Just build the best business and data-viewer browser software now, give it away, and outsource the hardware to Apple.

TOPICS: Oracle

What a week in IT. Sky-high Skype, Oracle-a-buy-buy,, Microsoft Office eye candy, MSNing around with AOL. Hard to call IT a has-been, commoditized industry with all of this action. It got me thinking of a unifying implication or common thread between these seemingly distinct points.

And then it came to me. I think that it is now, finally, time for the network computer. I don't care about the hardware, as long as it's x86 and wirelessly networked. I'd like to see it come standard with 60BG of disk space, and 2GB of memory, and lots of USB 2 slots. Operating system can be Windows, Linux, Mac OS X, I don't care as long as it runs a speedy JVM and has some stock reader applications for accessing and playing most document and media formats. I'm thinking of the Apple Mini or equivalent in about a year, that sort of thing. Under $500 loaded. Maybe $900 for the notebook version.

But what I'd really like to see as soon as possible is the next-generation thin-client, catch-all software application to run on this hardware, and Oracle is just the company to produce it. Yep, Oracle. After all, Larry Ellison was a driving force behind the network computer concept going on eight years ago (or more). And now that Oracle is number two in business apps, and by no means as cozy with (or dependent on) Microsoft as SAP, it's time for Larry to hammer this puppy home. How about a thin front-end for those business apps that has nothing to do with Microsoft, eh, Larry? That was the dream, right? Just build the best business and data-viewer browser software now, give it away, and outsource the hardware to Apple.

Start with the Firefox browser from and make it a Web developer's primary target; the object of an AJAX scripter's dream; a parser's parser; the rose on a Web services vine; the easiest endpoint for object, relational, and XML data connectivity. Make it open, fast, and rich -- but not too rich -- in UI niceties. Make it secure and easily up-dateable. Let me use a headset for VOIP calls. Let me synch to my cell phone and/or iPod through it, too. This software client, which Oracle, of course, would donate back to the appropriate open source communities with a no strings-attached and unfettered license, would be what Internet Explorer could have been but never will be.

For enterprises, put an Oracle data cache on the client, if you must, for those odd offline times and speedier performance requirements. And, oh, well ... fine, optimize it to perform best with Oracle RAC applications and stripped grid data servers. Fine. I can live with that. Just let me live in the thing as a user across all my REST/POX, SOAP, SIP, VOIP, and DHTML productivity needs. A truly unified thin client. Ahhhh.

But don't wait. The newest Microsoft Office 12 clients-as-rich-front-ends stuff won't arrive for 14 months, and not enter the market in a meaningful way for probably 20 months, so now's the time to get the new definition of thin-client computing out the door. That way the ISVs, corporate developers, on-demand services providers, SOA architects, and mobile carriers will see the true future, the one of choice, freedom, and the services fabric. Hell, even Benioff will like it, even if software is ended.

Oh, and if Oracle is not into this right now, too busy with M&A details, for example, how about you, IBM? Whattaya say? We could call it Wicked WorkPlace, or something catchy like that. Market it as Multi-Protocol, Client-Enabled Portal-Optimized Middleware (WebSphere MPCE-POM), as you are wont to do, if you must. Just get it done.

After a week like this, you have to dream.

Topic: Oracle

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • I'll believe it when I see it (nt)

    I love listening to broken records, lol.
  • Why?

    I mean tell me what advantage you see to this?
    • Power, cooling, maintenance

      And in some location, isolated acces to multiple networks. Just mundane business stuff. At least, that's what has us evaluating them. Interestingly, so far there's little evident downside.
      • I think that he was suggesting it also for home use

        Practical in an office environment, perhaps.

        Practical in the home environment? Not even on a good day.
  • Nothing has changed

    Web-based UIs are still only good for applications which are only used occasionally. If you use an app heavily, a browser will get tiresome.
    But, it would be great if an independent company did it. Last time, there was no opportunity to profit from the failure of this idea. The losses were buried in profitable, large companies like Oracle and Sun. If a small company did it, and went public quickly, we could all sell their stock short and retire...
    • Browsers are OK, but ...

      ... the network computer as suggested would give lots of other development options. AJAX or Java would be great as long as the local VMs etc. have good performance and plenty of available RAM. (in my experience a Java app needs about 6 times the RAM of an equivalent C++ app.)

      The Mono open-source implementation of Microsoft .Net framework should be on NCs too, don't you think?

      Anyway, keep all clients thin and above all, completely centralized management, including firmware updates! (no more visits to the desktop for software glitches.)
  • Nice idea

    I dunno it it'll ever happen - But I expect it'll come from google and be AJAX-delux, altho if you skip AJAX you get much more client independence.

    AJAX forces you into certain things on the client that a "client-server" web app won't require, altho certainly you'd use AJAX to spiff up the UI.

    As with all things, moderation is the key :).
  • Why hasn't client side Java caught on?

    A few years ago, there was much talk about client side Java. Once in a while I see a Java Applet running in my browser. Did MS kill it?

    Ariste Software
    • client side Java?

      It did not catch on because it is slow, slow, slow and clunky, clunky, clunky.
      • Java is slow... like chucko is slow...

        Chucko I'm sure is actually billiant and very quick in his realm. He is however EXTREMELY out of date and misinformed right now.

        The only reason Chucko believes that Java is slow is because Chucko never tried well written Java applications and bases his understand on older Java AWT applets. You see, Java did not stop growing when Microsoft introduced .Net the way the Northwest hoped. What started as hotspot has branched into the type of industry growth and revolt that only open development environments can foster. If Chucko understood Java development and knew what to ask for, his Java GUIs would scream and please his every desire as a click happy user. Yes... Java has Eclipsed its former self and continues to shine brighter. We simply needed to move the Sun dial a bit. Our dear Chucko was blinded by the buzz happy IT department misconception that "Java is slow". For shame Chucko... for shame.

        But Chucko is smart and can be brought up to date. Here is what Chucko is googling right now:

        "SWT and JFace and Eclipse and performance"

        In about 30 minutes (assuming that he actually reads web pages and doesn't just read the headings) Chucko can disconver that JFace/SWT applications are responsive and graphic rich on Windows, MAC, and Linux. He discovers that he actually can expect a Java application to respond quickly and still have all the great things of the past like serialized objects on every platform from mainframe to small server to desktop to set top boxes, to mobile phones.

        As a side, Chucko is not totally gone. You see Chucko does know that Oracle actually can optimze Java well and get things to work when it needs to (need is hard to prove to them at times..). So, he is rethinking his comment in light of the fact that Java has changed and Oracle can be smart about this stuff sometimes.
        • Java is the hope, not only for thin client, but RICH THIN CLIENT

          Yes, I strongly agree with jgruber. The first thing that come across most people's mind when we talked about Java is the speed: "Java is slow". But until they actually meet a well-written Java application...

          Java Applet is never dead. Best thing about using an applet as a rich internet application over Flash or AJAX or any other things is that, it's PURE JAVA. No extra learning, no extra nothing. It's pretty hard to find good developer who's a good graphics designer too, and it's even harder to find good graphics designer who knows programming. In the end, you need two groups of people with two different skill sets to write an application.

          I've come across this company (, which has a technology called Facado, providing not only thin client experience, but RICH THIN CLIENT experience, in a browser. RICH over here stands for rich user experience, such all the window desktop experience that you will have in your OS. It's actually a smart applet technology (not WebStart, yeah!), which provides a good framework, both for the users and the developers.

          Check out the demo. I think it's really good.
  • Yes, it's a broken record

    I also agree - it's fun to listen to broken records.

    Can you say "dumb terminal". This industry has "been there and done that" a few times before and most folks have enough common sense to realize that this is the dumbest idea since the invasion of Iraq.

    There are a (very) few area where something like this might have (some) merit, but for most situations, it's a square peg in a round hole.

    In other words, it's the FEMA version of a PC.
  • Why not go one step further

    Simplified BIOS workbook.

    Chuck out the x86 processor. Swap it for a RISC.
    No OS except BOOTP over Wireless. Machine downloads a simple Windows CE (or small Linux) style OS directly from network, with payed for apps already configured to go (similar to the model).
    All the usual ports (printing ability, download to USB key or similar etc). At the end of the day turn the machine off. Nothing is stored locally (all storage done on the apps server). If the machine gets stolen, no security headache.

    An OS-less machine. Wouldn't that be scary for Gates hehe.
    • Risc vs x86

      The processor is not important. There is no reason why such a device would not be practical with a x86.

      If you are going to rely on something like BOOTP to boot your device you are going to need a network that is alot more reliable and faster than is available in homes today.
      • Re: Risc vs x86

        Well heat and power are a couple of good reasons not to use an x86 processor, however this is an arguement for another day. I only suggested RISC because that is what came to mind at the time.

        As for BOOTP, that was the idea of using a Windows CE or mini-linux based OS, or even a QNX based system. Whatever the choice, providing it is:
        1) very small
        2) fast3
        3) Internet capable.
        then it should do the job nicely. Couple that up with a WIMAX network and the possibility is there.

        The only problem I can see at the moment is how to ensure secure delivery of the OS (ie hacker / tamper-proof).
        • Whoops...

          [i]however this is an arguement for another day[/i]

          should read:

          however this is an argument for another day.
  • Right, but what about adding Trillian or JXTA

    The P2P of JXTA for file sharing, on the edge of the network, and for synching the offline data with others including a central application repository. Trillian for IM gateway to whomever you know on whatever IM network.

    And what about Eclipse as the Rich client with FireFox running in one of the views?

    Talk about Ahhhhh?
  • "Blame it on my ISP"

    Anyone with computers at home knows you can't completely rely on on-line connectivity. How often is DSL or Cable down? Ever have a game server go down? People would just love it if their provider goes down and they can't access their word processing documents.

    Furthermore, a big technology driving PCs these days is gaming. Gaming requires beefy hardware. Yet, part of the reason for the popularity of the PC and Windows is the common interface for work and for home (i.e., gaming).

    Network terminals? Ha!
    • I would assume gaming would drive this...

      I'm confused by your comment. Network computing is a gaming platform.

      I would think that morphing gaming consoles with great graphics hardware would be where to start. As much as Apple would make it sexy, Sony makes much better consumer electronics. Their gaming console is still tops. You want this to be accepted, make it your HDTV tunner(OTA and working with cable/dish providers), PVR, and VOIP phone (or registration server) too.

      As for corporate use, Wiki like coporate portals where the work is less about word doc format as about content has already taken hold in the Internet space. Make IM/VOIP/Email one experience through the service portal (make SOX compliance for companies easier too... with now real ability to pull mail/docs locally.. you can police them much easier). The trick is to secure it and make DRM (as much as we hate it) work correctly. All possible in 2005/6 not 2010 (when is Vista going to make a dent?)

      Come on... Even diamond Bill (who is VERY sharp) is now preaching Exchange as a service. LCS as a VOIP technology is right here..right now. Do you really think Microsoft innovates in this space...please. They rest are just too disorganized to take them out with better technology.

      When the net is down... fine... watch recorded tv, play a game. Just make it all one device and everyone wins all the time. In Texas where I live, the electric is out more than my cable modem. Sounds like where you live the arena is ready for a great 3G cell network. Either way, I would bet on the network reliability growing more than it going away. VOIP plays alone will drive this. (If you live in country, you're getting your reward now and should not complain about bad network access...move into town if you want that...else..go feed a cow or something)
  • thin client?

    >60BG of disk space, and 2GB of memory

    This is more computer than most PC's sold today. I have seen Linux servers running with much less than your specs. You call that thin?

    Produced by Oracle? Since when has anything that Oracle produced anything that was anywhere close to being thin?

    >I don't care as long as it runs a speedy JVM and has some stock reader applications for accessing and playing most document and media formats

    That is the same sort of BS that Netscape was threatening before MS took them out. Remember the Java version of Word Perfect? I had a copy of that beta and it was a bloated hunk of garbage, no wonder it was killed off. How many times have I heard from the *nix crowd complain about Windows being bloated and complaining that it requires the fastest, most expensive hardware available any here you propose something that would be even more bloated than Windows. Your thin client would need things that are both not common today and cost much more than your target price. Primarily the largest flaw of your 'dream' are:

    Reliable network connections with sufficient bandwidth. Dana, if you had a realistic view of the world you would not think like someone that never leaves the server room where you work. A network computer like the one that you fantasize about requires a reliable connection, that means NOT DSL and NOT Cable modems because the have no QoS guarantees, it requires MUTIPLE T1's for bandwidth and reliability and that costs well over 1,000 USD per month for 2, at least 4 T1's would be required for all of the bandwidth that your dream will need for a satisfiying user experience. I don't think that this fits into your 500 USD price range. For those of you that do not know, 4 T1's is only 4 megabits, that is less than half of 10baseT. And where in your dream did you mention the cost of the server utilization? Did you think that some company would volunteer the CPU time and disk space for free? Silly you, you have been hanging out with the free source people too long. It is quite probable that if your fantacy network was reliable and inexpensive enough to make your dream practical that the cost of the server side of your dream over a period of say even 1 year would EASILY cost much more than purchasing an eMachine from Best Buy. This is not a step forward.

    This is nothing more than the usual Anti-MS BS that I normally read here, not much reason for it other than the author's religious beliefs.