Hi Joe <SNIP> >This has nothing to do with Java. It has to do with web serving. Not the >same thing by a long shot, and maybe that's part of the problem. > > > > Back to the original question, then: how much does 100M of disk > > cost? > >8GB AS/400 disk: $1400 >20GB IDE disk: $88 >This doesn't take into account labor/installation charge. Also it should be >noted that the AS/400 disk is 10,000RPM as opposed to the 7200RPM of the >IDE disk. > > > > But in my limited experience, I don't know too > > many businesses that can claim to be making any sound business > > judgments if they worry about $1000. > >I disagree. If everybody could afford to add GB to their machine, they >would. Most can't, especially in these economic times. In some cases, it's >because they've maxed their box, in some cases, it's because they can't >justify it based on the numbers above. Their CFOs start thinking "$1000 >here, $1000 there, pretty soon you're talking about real money." And they'd >be wrong not to. And yet.. and yet we see them do this over and over again with PC solutions throwing good money after bad. Once a project of any description gets started its hard to stop it. Far better that we say to them, this will cost a few bucks to do properly, let's do it right from the start. NT solutions start out looking cheap but end up being a false economy. Or perhaps you are proposing we do this on a linux box ? > > I'm not saying that well. We're talking do you support a new platform and > > an added layer of management? > >The "new platform, new layer of management" argument doesn't apply to a >simple web server. A web server is an appliance, not the object of a >masters thesis. > >Have you ever set up a web server? An 850MHz box with 256MB of RAM and a >20GB drive, assembled, costs $500, plus a monitor (another $150 if you shell >out for a new one). You buy NetMax for $50, put in the CD, answer the >questions, and you're up and running in under an hour for about $700. You >transfer images using Explorer to drag and drop, or by FTP. Your TCP/IP >person opens up one port in your firewall (preferably using NAT to an >internal 10. address) and you're done. Not a whole lot of management >required. Backups are the only management issue, and as I alluded to in a >previous post, if you can't back up a single PC, your IT department is in >serious trouble. Except that its a "PC". And probably the "NT guy" is looking after it. I don't know about you, but I've seen these guys have unreliable backups on many occasions - and those are the ones that actually do their backups ! I won't bore you with the story of the guy that told me during a recovery of a server "oh, that lights always on" (referring to the light on the tape drive telling him his tape drive needed cleaning and wasn't backing up - and hadn't been for two years) :) >You can have two redundant 20GB web serving platforms up and running for the >same price as 8GB of disk on an AS/400. Less cost, no load on your AS/400, >the only additional ongoing overhead being a regular PC backup. How is this >a bad thing? Because you are giving away control of a piece of the business and IT that you shouldn't give away. Before long the PC guy or the web guy is in with the marketing manager or the CEO telling them how if they added IIS, Apache, Perl, PHP, Zope or whatever turns his particular wheel that they could serve dynamic data such as up-to-date price lists instead of just static pages. Next thing you know you are being asked to supply the web guy with data on a daily basis.. or (and this is where this all started) you are being instructed to set up an FTP feed to the NT box to keep the data on it current. And when the PC guys says it needs to be FTP (or worse still a file share), management believe him because after all, he's been doing this stuff, right ? My view is that when you start giving away control of crucial business technology to people that don't understand how to manage this stuff properly, then you are fighting a losing battle. Once it's on the NT box you have to fight to get it back, and fight every inch of the way to stop new function (and investment, and jobs, and budget) being allocated to the harmless little web server in the corner. > > I'm sure some businesses make a decision of that magnitude along those > > lines, but I consider that more of a decision based on which way > > the wind is blowing. > >Anybody who spends $4000, or even $1000, based on the way the wind is >blowing is eventually going to find themselves the proud owner of a sock >puppet and a chapter 11 filing. > >Anyway, there's no arguing this point. Maybe it's because I grew up with >microcomputers as well as midranges, or maybe it's because I think most >operations staff can handle a PC that sits in a corner and doesn't even need >to be rebooted. But unless I have no load problems and no space problems, >and don't see any in the future, it doesn't make sense not to at least >explore the option of a dedicated server, especially for static data. By and large I don't disagree with what you say Joe, in particular how difficult the economics are to justify, but I see this as retaining a strategic service. It is far easier to start configuring your AS/400 HTTP server and the other bits and pieces using static pages to understand how to do web serving, than it is to argue the case for JSP's or websphere or whatever once the webserving is being done from the little box in the corner (or however many the NT guys has managed to grow it into. Evan Don't-give-it-away-coz-you-won't-get-it-back Harris
As an Amazon Associate we earn from qualifying purchases.
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.