I host my own domain on a redisential SDSL line and run pop3/IMAP webmail for a handful of my friends.
Although the mail is backed up every night to another spindle in the machine I've told every one of my friends *not to* rely on me as a mail provider - that their free email address was just for convenience and that if the server ever crashed I may recover the mail - or I might just turn the server's case into a barbecue grill.
I had a heck of a time with the FC1 --> FC2 upgrade but wasn't worried because everything was still flowing into mbox files, I just couldn't get it into IMAP mailboxes. Took me three days to figure out.
During those three days I had a handful of phone calls from friends who couldn't hit the mail server and one that was particularly irate that he couldn't access his 15mb (!) of mail. I almost decided at that popint to leave the server down;-)
On another note I'm part of an IPT working on outsourcing a helpdesk for 32,000 users and one of the requirements they're trying to implement is 100% database uptime with financial disincentives to the contractor if they don't provide.
I tried to explain that 100% uptime is impossible and five nines is ridiculously expensive - but that if that's what they really wanted I'm sure they could find a contractor to bid on the uptime requirement. They're pretty set on the 100% uptime requirement - I sure hope they've got deep pockets.
Oh, wait - they do have deep pockets. The project is being funded with tax dollars.
I'm not sure this helps build study skills - which will be critical, at least in college.
I can see this for elementary school kids but I think part of the challenge in high school and college is learning to do research and to express yourself orally and in writing.
Sight unseen, there are times when learning isn't supposed to be fun, I guess. That's why they call it work;-)
Too bad the firewall software loads *last* in the startup sequence, leaving a gaping hole of anywhere from 20 seconds to two minutes (on a slow machine) when your machine is on the net and unprotected. And during the height of worm activity, that's *more than enough* time to get infected.
This has changed in SP2. The firewall loads before any network interface is activated.
Unfettered, unrestricted application access to a housekeeping system with as much clout as the registry should plain not be possible. Since it's impossible to secure the registry the registry never should have been implemented.
This isn't exactly true - we can still blame programmers;-)
Installers run in the user context - unless they follow Win2k/WinXP logo guidelines or programmers use shortcuts and force the application to run in the system context. Under NT/2k/XP, plain old users cannot write to the registry, Power Users can change keys but not create them (there is no Power User group in XP Home) and administrators can add and change keys as required.
Having done large-scale deployments of Windows 2000 (and gearing up for Windows XP) the permissions issue can be a real problem - for which a common solution is to make all users administrators - so it's not MS breaking the security model here.
In our own deployments, much as we'd like to make users just plain users, there are too many applications that flat refuse to run properly in the plain old user context - one that comes to mind is Netscape 4.x - which will throw an error every time it tries to write to the user's profile in the registry and finds it doesn't have permission.
After reading the article it's kinda clear to me the author isn't a Windows user;-)
If the registry or the filesystem gets bloated because of malfunctioning application uninstallers, how is that MS' fault? Blame the nitwits who wrote the malfunctioning application.
Every OS has security patches available - if lack of patch has been exploited that exploit would apply to *any* OS - not just to Windows. If someone decided to write malware for Linux an unpatched machine would be just as vulnerable. Windows is a big target.
The way I read the post was the *only* reason for SMP was if you reached an architecture limit - and I disagree with that.
Hell, it wouldn't be the first time I've misread a post here;-)
On multiuser systems the benefits to SMP are clear. IM frequently less than HO running two processors at 50% is considerably more efficient than running one at 100%.
The _ONLY_ reason to do this is as a last resort when you can no longer clock your existing core any higher.
Incorrect.
As the subject line says, I've been running SMP desktop PCs for years. My current home PC is a dual 1GHz P-III, my wife's is a dual 850 and my Linux web/file/mail/whatever server is a dual 700 with a 12% overclock.
You can only figure on about a 40% performance increase with a dual processor desktop PC, but being able to play Quake and burn a DVD at the same time has it's advantages;-)
As others have mentioned, multitasking is greatly enhanced - and two midrange processors are generally cheaper than one high-end processor.
Also, even though some applications aren't multithreaded, all modern desktop OS are - so you get a performance boost even running single-task applications. If you're into running Windows, Internet Explorer is multithreaded, as are all Microsoft Office applications. There's a real-world productivity boost using SMP machines.
I don't know about the US, but elsewhere Telecom has been one of the hardest hit sectors of IT. It was one of the first and most severe casualties of the bursting bubble.
Yeah - I was thinking more network infrastructure than anything but you're right. At least they can't outsource the guys pulling wires to India;-)
Just my opinion, but I think a lot of the reason the free technical support experience generally sucks moose balls is that corporations may want to provide a disincentive for end users calling the helpdesk.
Case in point: Let's assume Microsoft's profit on a copy of Windows is $40 (on an OEM version it's probably less than that). At ~$20 per helpdesk call the end user only has to call customer support twice to erase any profit from that software sale. If I were in charge I wouldn't want users calling the helpdesk either;-)
I found the article fascinating and have decided to inflict my own observations on/.;-)
As long as the cost of living (and therefore the cost of labor) is higher in the US I believe corporate America will continue to outsource. Anyone who thinks efficiency or innovation is primarily an American product probably has a bit to learn about the rest of the world.
A subject that hits close to home for me is outsourcing support operations. I work for a fair-sized federal agency and we're currently looking at consolidating Level 1 helpdesk operations from 13 separate helpdesks to one outsourced agency to serve the whole organization.
Current industry standards say that for an organization our size a helpdesk call should cost the company ~$20. Outsourcing to India would cut that cost in half, so it's easy to see where that option would be attractive to big business.
Support operations do not generate income, therefore offshore outsourcing reduces operating costs. IM frequently less than HO corporate America's first loyalty is to stockholders and unfortunately altruism doesn't increase the bottom line, so I think companies will continue to outsource until there's a financial incentive for them to quit doing it.
I don't think code written in the US is necessarily more innovative than code written in India for half the cost - so until third world IT organizations raise wages it's still gonna be more attractive to outsource.
I think the bottom line is that wages for skilled American IT workers will continue to slip unless they're in a job that cannot be outsourced - I just suggested to my son-in-law that if he wanted a job in IT the place to be was probably in networking - preferably telecom or Information Assurance. Those fields will probably remain for the most part in the good old US of A;-)
Since the big-ticket item in corporate IT is support and not software, how does running Windows under Linux provide lower TCO - since although software costs will drop, Tier 1 support costs will increase?
To me it sounds as if the group Bruce is on the board of is trying to exploit or otherwise captalize on the FUD of SCO actually winning a lawsuit.
Agreed.
They're not really offering a lot for the money anyway - on the kernel developer side they offer $25k of protection. IMO if someone was to actually *win* a lawsuit $25,000 would be just a drop in the bucket.
Also, why would I give $100k to someone who doesn't know that at least in this context 'panel' isn't a proper noun?;-)
If we're concerned about hacking why not just vote using POTS instead of computers?
One issue would be non-repudiation, but once we've figured out a way to make sure people are who they say they are voting via telephone should be a pretty simple process to implement.
I was at a shindig in Redmond last month and Steve Ballmer took this very question from the floor.
He didn't exactly have an answer, other than to say they were still looking at the problem - but from what he did say MS is acutely aware of the problem.
I think my solution would be to allow security updates only. During this trip I had a long discussion with a pile of MS executives about community and/. came up more than a couple of times in the conversation;-)
I don't think the requirement to install service packs, updates and AV software goes away with a switch to Linux - and executable attachments have been stripped out by our mail gateway for almost five years. It's been awhile since we've seen a worm inside the firewall;-)
And you're right - most companies do pay a bit more. Bigger organizations pay a bit less.
Still, support costs more than software. Let's say that the price of RedHat and the price of Windows was the same - when you switch everything on a user they tend to call the helpdesk pretty frequently. At ~$20 per call it's pretty easy to see that at least for the first year, helpdesk costs would go through the roof just because you've put something on the user's desktop that *looks* different - never mind the fact that it most likely works just as well;-)
The average user in our organization calls the helpdesk about 14 times a year. About half of those calls are desktop issues and the other half are telecom issues or problems with applications hosted remotely - those problems wouldn't go away either.
But - one point where I haven't been clear is that a switch to *any* OS would increase costs dramatically - if we substituted Windows for OS X, BSD, Solaris and so on you'd see the same cost increase and wouldn't necessarily see an increase in performance or reliability. Once an organization has settled on a desktop architecture changing that architecture is ridiculously expensive. Software acquisition and licensing is the least of our worries here.
In addition, there are smaller but still significant problems with stuff like local mail stores and address books, Office macros that need to be rewritten for OpenOffice, Access databases to be converted to something OO can understand and so on.
I don't doubt that if you were to start a new company and roll out Linux the costs would be roughly equal to Windows - maybe a bit less as users got more familiar with open source applications - but at least in a large organization it'd be pretty tough to build a business case to make the switch, mainly because even an OS upgrade for Windows or a refresh of MS Office (which has zero software cost - it's paid for through enterprise licensing) costs millions of bucks and even though the workstations work fine after deployment you generally see a huge spike in helpdesk calls because everything *looks* different and although we document upgrades pretty extensively and distribute user guides, users would rather call the helpdesk than RTFM;-)
The bottom line for me is that even if software costs were equal and I could get all the Office stuff converted for free, the support costs required for a switch to a completely different OS and office automation suite make the move pretty unattractive.
It's not that Windows is better than Linux (although there's a bit to be said for standardizing look and feel), it's that when you factor in migration and support costs it'd be tough to show any financial advantage for the first few years. After you got all the users trained, maybe it'd be less expensive - maybe not.
BTW - at home I run two Windows workstations and one Linux webserver. If I were trying to position Linux in the corporate world I'd be going after departmental servers first - IMO Windows is bloatware in this area but Linux doesn't do SMP well enough for enterprise servers - at least not yet. Windows and Unix still have a bit of an edge in four-way and eight-way enterprise boxes, I'm afraid - but the advantage here is diminshing quickly.
All I can do is use the numbers they provide, but even if enterprise pricing was a wash that doesn't negate the issue of increased support costs. Software licensing isn't really a factor in an enterprise this size - hardware and support costs are.
Actually ours isn't too bad when it comes to Tier 1 support; ~$20 per call is about average for an organization our size. That's one reason I only counted desktops - trying to keep the network stuff out of it.
But you're right - the network is a mess. Too many servers, too many network "engineers" who couldn't admin their way out of a paper bag - but at least they're expensive;-)
My job is ADP R&D - I'd like to redo the network from the ground up.
Their pricing structure doesn't incluse *any* end-user support - they're positioning themselves at Tier 2 and 3. They'll respond to calls from your helpdesk, but not from end users - at least that's what RedHat says.
I'm actually enjoying this discussion - it's nice to be able to kick things around without a buncha finger-pointing;-)
(620 * $3500 price per extension pack) / 31000 users = $70 per user
MS enterprise licensing costs less than half that.
Support: RedHat is still only providing Tier 2 and 3 support - my organization would still have to field all Tier 1 questions. From RedHat:
"Red Hat Desktop users will receive their day-to-day support from your company's Help Desk. To assist your Help Desk staff your Red Hat Desktop annual subscription entitles them to unlimited Help Desk Escalation Support, so that they have continuous access to Red Hat support resources." [boldface mine]
Again, I read the article. Doesn't say anything about support at all - other than it was available.
To be fair I went out to RedHat's website and checked pricing - apparently 30 days of phone support and one year of web support are included - but the SLA for that support leaves a bit to be desired: installation and configuration only (which would normally be performed by Tier 2 technicians, not end users).
My users don't call the help desk to find out how to install or configure Windows - they call with OS or application "how do I' calls that appear to be excluded from RedHat's support offering.
The article doesn't say what volume licensing costs - but since RedHat is targeting corporations maybe it's safe to assume that the $5 per desktop per month *is* volume licensing?
If there are different numbers I'd be happy to compare them - but again, software licensing isn't much of a factor when costing out enterprise architecture for a large organization - support costs are the single biggest expense.
Slashdot Mirror
Although the mail is backed up every night to another spindle in the machine I've told every one of my friends *not to* rely on me as a mail provider - that their free email address was just for convenience and that if the server ever crashed I may recover the mail - or I might just turn the server's case into a barbecue grill.
I had a heck of a time with the FC1 --> FC2 upgrade but wasn't worried because everything was still flowing into mbox files, I just couldn't get it into IMAP mailboxes. Took me three days to figure out.
During those three days I had a handful of phone calls from friends who couldn't hit the mail server and one that was particularly irate that he couldn't access his 15mb (!) of mail. I almost decided at that popint to leave the server down ;-)
On another note I'm part of an IPT working on outsourcing a helpdesk for 32,000 users and one of the requirements they're trying to implement is 100% database uptime with financial disincentives to the contractor if they don't provide.
I tried to explain that 100% uptime is impossible and five nines is ridiculously expensive - but that if that's what they really wanted I'm sure they could find a contractor to bid on the uptime requirement. They're pretty set on the 100% uptime requirement - I sure hope they've got deep pockets.
Oh, wait - they do have deep pockets. The project is being funded with tax dollars.
I can see this for elementary school kids but I think part of the challenge in high school and college is learning to do research and to express yourself orally and in writing.
Sight unseen, there are times when learning isn't supposed to be fun, I guess. That's why they call it work ;-)
I'd like to know the value of the following folks, please:
George W. Bush
Larry Ellison
Darl McBride
Inquiring minds want to know.
Never traded sex for geek work, though. Always figured my ass would hurt enough that I couldn't sit down to do my daytime job.
This has changed in SP2. The firewall loads before any network interface is activated.
This isn't exactly true - we can still blame programmers ;-)
Installers run in the user context - unless they follow Win2k/WinXP logo guidelines or programmers use shortcuts and force the application to run in the system context. Under NT/2k/XP, plain old users cannot write to the registry, Power Users can change keys but not create them (there is no Power User group in XP Home) and administrators can add and change keys as required.
Having done large-scale deployments of Windows 2000 (and gearing up for Windows XP) the permissions issue can be a real problem - for which a common solution is to make all users administrators - so it's not MS breaking the security model here.
In our own deployments, much as we'd like to make users just plain users, there are too many applications that flat refuse to run properly in the plain old user context - one that comes to mind is Netscape 4.x - which will throw an error every time it tries to write to the user's profile in the registry and finds it doesn't have permission.
If the registry or the filesystem gets bloated because of malfunctioning application uninstallers, how is that MS' fault? Blame the nitwits who wrote the malfunctioning application.
Every OS has security patches available - if lack of patch has been exploited that exploit would apply to *any* OS - not just to Windows. If someone decided to write malware for Linux an unpatched machine would be just as vulnerable. Windows is a big target.
The way I read the post was the *only* reason for SMP was if you reached an architecture limit - and I disagree with that.
Hell, it wouldn't be the first time I've misread a post here ;-)
On multiuser systems the benefits to SMP are clear. IM frequently less than HO running two processors at 50% is considerably more efficient than running one at 100%.
Incorrect.
As the subject line says, I've been running SMP desktop PCs for years. My current home PC is a dual 1GHz P-III, my wife's is a dual 850 and my Linux web/file/mail/whatever server is a dual 700 with a 12% overclock.
You can only figure on about a 40% performance increase with a dual processor desktop PC, but being able to play Quake and burn a DVD at the same time has it's advantages ;-)
As others have mentioned, multitasking is greatly enhanced - and two midrange processors are generally cheaper than one high-end processor.
Also, even though some applications aren't multithreaded, all modern desktop OS are - so you get a performance boost even running single-task applications. If you're into running Windows, Internet Explorer is multithreaded, as are all Microsoft Office applications. There's a real-world productivity boost using SMP machines.
Yeah - I was thinking more network infrastructure than anything but you're right. At least they can't outsource the guys pulling wires to India ;-)
Exactly.
Just my opinion, but I think a lot of the reason the free technical support experience generally sucks moose balls is that corporations may want to provide a disincentive for end users calling the helpdesk.
Case in point: Let's assume Microsoft's profit on a copy of Windows is $40 (on an OEM version it's probably less than that). At ~$20 per helpdesk call the end user only has to call customer support twice to erase any profit from that software sale. If I were in charge I wouldn't want users calling the helpdesk either ;-)
As long as the cost of living (and therefore the cost of labor) is higher in the US I believe corporate America will continue to outsource. Anyone who thinks efficiency or innovation is primarily an American product probably has a bit to learn about the rest of the world.
A subject that hits close to home for me is outsourcing support operations. I work for a fair-sized federal agency and we're currently looking at consolidating Level 1 helpdesk operations from 13 separate helpdesks to one outsourced agency to serve the whole organization.
Current industry standards say that for an organization our size a helpdesk call should cost the company ~$20. Outsourcing to India would cut that cost in half, so it's easy to see where that option would be attractive to big business.
Support operations do not generate income, therefore offshore outsourcing reduces operating costs. IM frequently less than HO corporate America's first loyalty is to stockholders and unfortunately altruism doesn't increase the bottom line, so I think companies will continue to outsource until there's a financial incentive for them to quit doing it.
I don't think code written in the US is necessarily more innovative than code written in India for half the cost - so until third world IT organizations raise wages it's still gonna be more attractive to outsource.
I think the bottom line is that wages for skilled American IT workers will continue to slip unless they're in a job that cannot be outsourced - I just suggested to my son-in-law that if he wanted a job in IT the place to be was probably in networking - preferably telecom or Information Assurance. Those fields will probably remain for the most part in the good old US of A ;-)
Since the big-ticket item in corporate IT is support and not software, how does running Windows under Linux provide lower TCO - since although software costs will drop, Tier 1 support costs will increase?
Agreed.
They're not really offering a lot for the money anyway - on the kernel developer side they offer $25k of protection. IMO if someone was to actually *win* a lawsuit $25,000 would be just a drop in the bucket.
Also, why would I give $100k to someone who doesn't know that at least in this context 'panel' isn't a proper noun? ;-)
One issue would be non-repudiation, but once we've figured out a way to make sure people are who they say they are voting via telephone should be a pretty simple process to implement.
He didn't exactly have an answer, other than to say they were still looking at the problem - but from what he did say MS is acutely aware of the problem.
I think my solution would be to allow security updates only. During this trip I had a long discussion with a pile of MS executives about community and /. came up more than a couple of times in the conversation ;-)
And you're right - most companies do pay a bit more. Bigger organizations pay a bit less.
Still, support costs more than software. Let's say that the price of RedHat and the price of Windows was the same - when you switch everything on a user they tend to call the helpdesk pretty frequently. At ~$20 per call it's pretty easy to see that at least for the first year, helpdesk costs would go through the roof just because you've put something on the user's desktop that *looks* different - never mind the fact that it most likely works just as well ;-)
The average user in our organization calls the helpdesk about 14 times a year. About half of those calls are desktop issues and the other half are telecom issues or problems with applications hosted remotely - those problems wouldn't go away either.
But - one point where I haven't been clear is that a switch to *any* OS would increase costs dramatically - if we substituted Windows for OS X, BSD, Solaris and so on you'd see the same cost increase and wouldn't necessarily see an increase in performance or reliability. Once an organization has settled on a desktop architecture changing that architecture is ridiculously expensive. Software acquisition and licensing is the least of our worries here.
In addition, there are smaller but still significant problems with stuff like local mail stores and address books, Office macros that need to be rewritten for OpenOffice, Access databases to be converted to something OO can understand and so on.
I don't doubt that if you were to start a new company and roll out Linux the costs would be roughly equal to Windows - maybe a bit less as users got more familiar with open source applications - but at least in a large organization it'd be pretty tough to build a business case to make the switch, mainly because even an OS upgrade for Windows or a refresh of MS Office (which has zero software cost - it's paid for through enterprise licensing) costs millions of bucks and even though the workstations work fine after deployment you generally see a huge spike in helpdesk calls because everything *looks* different and although we document upgrades pretty extensively and distribute user guides, users would rather call the helpdesk than RTFM ;-)
The bottom line for me is that even if software costs were equal and I could get all the Office stuff converted for free, the support costs required for a switch to a completely different OS and office automation suite make the move pretty unattractive.
It's not that Windows is better than Linux (although there's a bit to be said for standardizing look and feel), it's that when you factor in migration and support costs it'd be tough to show any financial advantage for the first few years. After you got all the users trained, maybe it'd be less expensive - maybe not.
BTW - at home I run two Windows workstations and one Linux webserver. If I were trying to position Linux in the corporate world I'd be going after departmental servers first - IMO Windows is bloatware in this area but Linux doesn't do SMP well enough for enterprise servers - at least not yet. Windows and Unix still have a bit of an edge in four-way and eight-way enterprise boxes, I'm afraid - but the advantage here is diminshing quickly.
Thanks for the discussion ;-)
Still, support costs are the most significant expense here.
All I can do is use the numbers they provide, but even if enterprise pricing was a wash that doesn't negate the issue of increased support costs. Software licensing isn't really a factor in an enterprise this size - hardware and support costs are.
But you're right - the network is a mess. Too many servers, too many network "engineers" who couldn't admin their way out of a paper bag - but at least they're expensive ;-)
My job is ADP R&D - I'd like to redo the network from the ground up.
Their pricing structure doesn't incluse *any* end-user support - they're positioning themselves at Tier 2 and 3. They'll respond to calls from your helpdesk, but not from end users - at least that's what RedHat says.
I'm actually enjoying this discussion - it's nice to be able to kick things around without a buncha finger-pointing ;-)
(31000 users / 50 users per extension pack license) = 620 licenses required.
(620 * $3500 price per extension pack) / 31000 users = $70 per user
MS enterprise licensing costs less than half that.
Support: RedHat is still only providing Tier 2 and 3 support - my organization would still have to field all Tier 1 questions. From RedHat:
"Red Hat Desktop users will receive their day-to-day support from your company's Help Desk. To assist your Help Desk staff your Red Hat Desktop annual subscription entitles them to unlimited Help Desk Escalation Support, so that they have continuous access to Red Hat support resources." [boldface mine]
To be fair I went out to RedHat's website and checked pricing - apparently 30 days of phone support and one year of web support are included - but the SLA for that support leaves a bit to be desired: installation and configuration only (which would normally be performed by Tier 2 technicians, not end users).
My users don't call the help desk to find out how to install or configure Windows - they call with OS or application "how do I' calls that appear to be excluded from RedHat's support offering.
If there are different numbers I'd be happy to compare them - but again, software licensing isn't much of a factor when costing out enterprise architecture for a large organization - support costs are the single biggest expense.