That was disproven with the different rates of infection wrt apache vs. iis. The study showed Apache had by far the larger market share, but IIS had by far the most vulnerabilities. According to YOUR illogic, Apache, not IIS, should have had the most vulnerabilities.
Nothing of the sort was "disproven". In more ways than one you are comparing apples (successful exploits against an entire platform) to oranges (vulnerabilities in a single application).
Also, most crackers would prefer to p0wn one unix-type box over a dozen Windows boxes.
No, they wouldn't. A "unix-type box" will almost certainly be run by a competent, attentive and often professional administrator who will notice and clean up any issues in a short period of time. Of the dozen Windows boxes, it's unlikely even a single user would even notice something was wrong, let alone try to fix it.
Windows are the low-hanging fruit because the OS is pretty crappy by design. Microsoft refuses to make a clean break with the buggy code from the past, because they know that if they do, a lot of their customers are no longer "locked in", so it is in fact insecure by design.
um since Windows doesn't properly support multiple real time users, (3-4 people can log into a single linux machine and run the same application at the same time, on windows things to hit snags when you do that) without proper multi user support and the fact that MSFT won't force developers to actually code for proper multi-users you run into a problem trying to do just that.
*Windows* supports this just fine, and always has. Many *applications*, do not.
15 years of negligence by MSFT in this regard and it will take at least 2 more major releases from them before it is fully fixed to standards that every *nix used 15 years ago.
There's no "neglect" in ignoring something relevant to only a tiny minority of customers, to focus on other issues. Indeed, even the current resurgence of the dumb terminal is focussed on client-server applications like web browers, not mainframe-esque interactive logins to a single machine.
Please remember that the vast majority of hardware and peripherals are designed from the ground up to work with Windows and that most computers are sold with Windows preinstalled and preconfigured.
How do you design a piece of hardware "from the ground up" to work with a particular OS ?
Why on Earth not ? The bulk of Windows "viruses" are, in fact, trojans.
Install Linux on your Windows box and you do NOT need any antivirus (unless you boot into the Windows side), provided you're not stupid enough to run an executable from an untrusted source.
I've spent nearly 15 years running Windows using this principle, without an AV problem, and - unsuprisingly - have yet to be infected by anything.
The TCP overhead on iSCSI is too great when you start approaching gigE speeds. So now you're looking at purchasing a TOE (TCP Offload Engine) card to deal with that, or a dedicated iSCSI adapter.
Firstly, the overhead is insignificant on any remotely modern CPU.
Secondly, are anything but even the cheapest, nastiest, NICs even available without TOE these days ?
"Server automation" excels at bringing up systems up in a known state. If your environment is unusually disciplined, it might even be able to bring a new server to configuration parity with an existing one. But you're still need to going to test the new server before putting it into production and you're still going to have to take an outage to changeover. Both of these require more manpower and expertise, and are more disruptive than a VMotion.
And that's assuming an ideal situation and a relatively trivial application. Out in the real world, where servers have been running complicated applications like Oracle Appserver, or multiple applications on a single machine, essentially unmaintained and continually tweaked for 3-5 years, "server automation" simply doesn't work because the reality doesn't match the documentation.
I'd certainly expect that. I don't quite get what people are so desperate to de-duplicate anyway. A stripped VM os image is less than a gigabyte, you can fit 150 of them on a drive that costs less than $100.
Firstly, because dedup gives you the space savings without the hassle of "stripping" the VM image.
Secondly, because dedup also delivers other advantages by reducing physical disk IOs, improving cache efficiency and reducing replication traffic.
Thirdly, because enterprise storage costs a lot more than that, especially once you account for backups.
I can't really see many situations where the extra complexity and cost would end up actually saving money.
NetApp have quite a few white papers and blogs. The most high profile winner is virtualisation, of course, but things like SAN-booted OS images, mailboxes, backups and data replication also see huge benefits.
In fact, since hackintoshes are almost certainly eating into Apple's hardware sales (maybe not by much, but they must), [...]
Not true.
Indeed, that the majority of Hackintoshes seem to be for market segments Apple has no presence, or are explicitly refusing user demand, in, then it's hard to see how anyone could argue they "must" be "eating into Apple's hardware sales".
Yes, and server automation lets you duplicate a known good OS+app combination on a whim. What's the difference again?
The difference is that "server automation" *doesn't* do that in the real world, and even in ideal circumstances does not - and can not - achieve the same functionality as virtualisation.
Ignoring the obvious troll, I fail to see how VMware helps all with the scenario you describe. How does OS virtualization reduce interdependencies or complexity?
Because virtualisation lets you pick up your existing, known good OS+app image and drop it piecemeal onto another system, with a few seconds work and zero downtime.
Yes, it helps, but it really only helps with under-utilized hardware (and this is really only a problem in Microsoft shops).
This is not true at all. We have dozens of Linux servers with CPU usage that rarely rises above single digits, and my experience is that there's nothing even slightly unusual about that.
De-duping can and does work well on file, web and email servers. It doesn't work quite so well for SANs that present multiple LUNs to multiple servers:D
With propper server automation tools (such as puppet+kickstart), I can deploy any configuration on any supported hardware (virtual or not) in minutes from scratch.
All this says to me is that you've never had to migrate a service outside of trivial cases.
If you have a relatively simple app, you might be able to migrate it from one server to another in a day or two of total time, with downtime measured in minutes or seconds. For anything complicated or interdependent, do you need weeks or months - sometimes even years - to do it properly, and even then you might be looking at downtime measured in tens of minutes or hours.
Heck, if you had a long-running core service (let's say some sort of home-grown user authentication and authorisation tool), you'd want to take *at least* a month or two just to assess what interacted with, and depended on it, before even getting into detailed planning for the actual change, let alone the actual execution.
Live migration of VMs is nice, but not essential.
Live migration turns a high-risk operation that can take weeks or months of planning, testing and execution, into something mundane a junior admin can complete in a few mouse clicks. If you cannot grasp the massive benefit that represents, you're in the wrong business.
This is before even beginning to scratch the surface of the availability and disaster recovery benefits virtualisation delivers.
So you have maybe $10,000 in 5 to 6 servers needed to run a small to mid-size company and spend tens of thousands to put them on one super-server running a complex setup of virtualized servers...oh no, the motherboard died and the entire biz is offline.
Tens of thousands ? For well under ten grand you'll get a server that will do all of that with loads of capacity to spare - a vastly superior solution than 5-6 individual servers all idling along with a utilisation in the the single-digit-percentages.
Virtualization has it's place, but only at the larger companies.
The benefits of virtualisation are so clear, and its cost so low (essentially free), that in today's work it should be the default scenario - only not used in extraordinary circumstances. Heck, there's little reason not to virtualise a single app on a single server.
If you're using VMware to hide the fact that your app doesn't work on the same operating system on different hardware, you're doing something decidedly stupid.
Or you think that being able to migrate your known, working OS+application to a completely different server nearly instantly, and transparently to users, is a vastly superior proposition than building and testing a new server+OS+application over a period of days, weeks or months.
talking about signature, hasn't the mac pro mini come out already in this iteration?
The latest Mac Mini is not meaningfully different to the previous Mac Mini (or the one before that, or the one before that, et cetera back to the original).
If that's your measure of "innovation", then it's hardly unreasonable that Microsoft hasn't "innovated" - neither has nearly everyone else, because most "innovations" in computing were over and done with 30-50 years ago, before they were in any position to.
Microsoft has no equivalent; they have mostly bought out other companies that did innovate and claimed they did the innovation instead; or stole an idea from someone else that did innovate.
Just like everyone else, then ?
What the GP was asking for was for a specific example of just one innovation that actually came out of Microsoft - a single original idea from Microsoft. (Note: Event Clippy and MS Bob were stolen ideas that Microsoft implemented.)
With the bar set so high, then no, Microsoft hasn't "innovated" - but that's hardly a stinging criticism in that context, since neither have Apple, Google or, indeed, pretty much anyone in the last few decades.
But these are limitations imposed by the applications, not the OS. OSes support what you want quite well, and have for a very long time, but if applications don't make use of those features, there's not much the OS can do about it.
What about getting RID of the file selector, and just using the normal file views + drag and drop to open and save files? Drag a file to your word processor, and it opens. Drag the tab from the word processor to a disk, and you save. Drag a section of a file, and you save that section. Drag that section to the desktop, and you save a cut buffer, and you can have as many cut buffers as you want.
Most of these things have been in MacOS and/or Windows (and/or others, like OS/2) since the mid-90s, if not longer.
Doesn't matter if you are computer savvy or not, if you use Microsoft long enough and use the internet more than going to one or two trusted sites, you WILL get hacked.
I've been waiting 15 years, when is this going to happen ? The suspense is killing me...
Slashdot Mirror
That was disproven with the different rates of infection wrt apache vs. iis. The study showed Apache had by far the larger market share, but IIS had by far the most vulnerabilities. According to YOUR illogic, Apache, not IIS, should have had the most vulnerabilities.
Nothing of the sort was "disproven". In more ways than one you are comparing apples (successful exploits against an entire platform) to oranges (vulnerabilities in a single application).
Also, most crackers would prefer to p0wn one unix-type box over a dozen Windows boxes.
No, they wouldn't. A "unix-type box" will almost certainly be run by a competent, attentive and often professional administrator who will notice and clean up any issues in a short period of time. Of the dozen Windows boxes, it's unlikely even a single user would even notice something was wrong, let alone try to fix it.
Windows are the low-hanging fruit because the OS is pretty crappy by design. Microsoft refuses to make a clean break with the buggy code from the past, because they know that if they do, a lot of their customers are no longer "locked in", so it is in fact insecure by design.
Please detail these "design" problems.
um since Windows doesn't properly support multiple real time users, (3-4 people can log into a single linux machine and run the same application at the same time, on windows things to hit snags when you do that) without proper multi user support and the fact that MSFT won't force developers to actually code for proper multi-users you run into a problem trying to do just that.
*Windows* supports this just fine, and always has. Many *applications*, do not.
15 years of negligence by MSFT in this regard and it will take at least 2 more major releases from them before it is fully fixed to standards that every *nix used 15 years ago.
There's no "neglect" in ignoring something relevant to only a tiny minority of customers, to focus on other issues. Indeed, even the current resurgence of the dumb terminal is focussed on client-server applications like web browers, not mainframe-esque interactive logins to a single machine.
Try installing Windows 7 or Vista on hardware that's 7 years old. Good luck, now try to install Linux on it. More than likely Linux can be installed.
Oldest PC I've personally installed Vista on dated from early 2000. Worked fine (albeit a bit slow - though a $30 video card fixed that).
Please remember that the vast majority of hardware and peripherals are designed from the ground up to work with Windows and that most computers are sold with Windows preinstalled and preconfigured.
How do you design a piece of hardware "from the ground up" to work with a particular OS ?
Bullshit. Microsoft made the same claim when they made the switch from 16-bit to 32-bit - "Viruses will be a thing of the past."
They did ? Do you have a cite ?
Trojans don't count
Why on Earth not ? The bulk of Windows "viruses" are, in fact, trojans.
Install Linux on your Windows box and you do NOT need any antivirus (unless you boot into the Windows side), provided you're not stupid enough to run an executable from an untrusted source.
I've spent nearly 15 years running Windows using this principle, without an AV problem, and - unsuprisingly - have yet to be infected by anything.
The problem is not the OS.
Most (all now?) graphics cards are hardware accelerated for the Windows GDI, for things like drawing fonts, arcs, ellipses, fills, etc.
All *now* ?! They'd stopped selling video cards without 2D acceleration back when clock speeds were still measured in double-digit _Mhz_.
The TCP overhead on iSCSI is too great when you start approaching gigE speeds. So now you're looking at purchasing a TOE (TCP Offload Engine) card to deal with that, or a dedicated iSCSI adapter.
Firstly, the overhead is insignificant on any remotely modern CPU.
Secondly, are anything but even the cheapest, nastiest, NICs even available without TOE these days ?
I'll counter your anecdote with my own.
What anecdote ?
"Server automation" excels at bringing up systems up in a known state. If your environment is unusually disciplined, it might even be able to bring a new server to configuration parity with an existing one. But you're still need to going to test the new server before putting it into production and you're still going to have to take an outage to changeover. Both of these require more manpower and expertise, and are more disruptive than a VMotion.
And that's assuming an ideal situation and a relatively trivial application. Out in the real world, where servers have been running complicated applications like Oracle Appserver, or multiple applications on a single machine, essentially unmaintained and continually tweaked for 3-5 years, "server automation" simply doesn't work because the reality doesn't match the documentation.
It simply has to be. There must be people who are doing it because it's cheaper than a real Mac.
Which suggests they wouldn't otherwise have bought a Mac because it was too expensive.
I'd certainly expect that. I don't quite get what people are so desperate to de-duplicate anyway. A stripped VM os image is less than a gigabyte, you can fit 150 of them on a drive that costs less than $100.
Firstly, because dedup gives you the space savings without the hassle of "stripping" the VM image.
Secondly, because dedup also delivers other advantages by reducing physical disk IOs, improving cache efficiency and reducing replication traffic.
Thirdly, because enterprise storage costs a lot more than that, especially once you account for backups.
I can't really see many situations where the extra complexity and cost would end up actually saving money.
NetApp have quite a few white papers and blogs. The most high profile winner is virtualisation, of course, but things like SAN-booted OS images, mailboxes, backups and data replication also see huge benefits.
In fact, since hackintoshes are almost certainly eating into Apple's hardware sales (maybe not by much, but they must), [...]
Not true.
Indeed, that the majority of Hackintoshes seem to be for market segments Apple has no presence, or are explicitly refusing user demand, in, then it's hard to see how anyone could argue they "must" be "eating into Apple's hardware sales".
Yes, and server automation lets you duplicate a known good OS+app combination on a whim. What's the difference again?
The difference is that "server automation" *doesn't* do that in the real world, and even in ideal circumstances does not - and can not - achieve the same functionality as virtualisation.
Ignoring the obvious troll, I fail to see how VMware helps all with the scenario you describe. How does OS virtualization reduce interdependencies or complexity?
Because virtualisation lets you pick up your existing, known good OS+app image and drop it piecemeal onto another system, with a few seconds work and zero downtime.
Yes, it helps, but it really only helps with under-utilized hardware (and this is really only a problem in Microsoft shops).
This is not true at all. We have dozens of Linux servers with CPU usage that rarely rises above single digits, and my experience is that there's nothing even slightly unusual about that.
De-duping can and does work well on file, web and email servers. It doesn't work quite so well for SANs that present multiple LUNs to multiple servers :D
NetApp would argue with that.
With propper server automation tools (such as puppet+kickstart), I can deploy any configuration on any supported hardware (virtual or not) in minutes from scratch.
All this says to me is that you've never had to migrate a service outside of trivial cases.
If you have a relatively simple app, you might be able to migrate it from one server to another in a day or two of total time, with downtime measured in minutes or seconds. For anything complicated or interdependent, do you need weeks or months - sometimes even years - to do it properly, and even then you might be looking at downtime measured in tens of minutes or hours.
Heck, if you had a long-running core service (let's say some sort of home-grown user authentication and authorisation tool), you'd want to take *at least* a month or two just to assess what interacted with, and depended on it, before even getting into detailed planning for the actual change, let alone the actual execution.
Live migration of VMs is nice, but not essential.
Live migration turns a high-risk operation that can take weeks or months of planning, testing and execution, into something mundane a junior admin can complete in a few mouse clicks. If you cannot grasp the massive benefit that represents, you're in the wrong business.
This is before even beginning to scratch the surface of the availability and disaster recovery benefits virtualisation delivers.
So you have maybe $10,000 in 5 to 6 servers needed to run a small to mid-size company and spend tens of thousands to put them on one super-server running a complex setup of virtualized servers...oh no, the motherboard died and the entire biz is offline.
Tens of thousands ? For well under ten grand you'll get a server that will do all of that with loads of capacity to spare - a vastly superior solution than 5-6 individual servers all idling along with a utilisation in the the single-digit-percentages.
Virtualization has it's place, but only at the larger companies.
The benefits of virtualisation are so clear, and its cost so low (essentially free), that in today's work it should be the default scenario - only not used in extraordinary circumstances. Heck, there's little reason not to virtualise a single app on a single server.
If you're using VMware to hide the fact that your app doesn't work on the same operating system on different hardware, you're doing something decidedly stupid.
Or you think that being able to migrate your known, working OS+application to a completely different server nearly instantly, and transparently to users, is a vastly superior proposition than building and testing a new server+OS+application over a period of days, weeks or months.
talking about signature, hasn't the mac pro mini come out already in this iteration?
The latest Mac Mini is not meaningfully different to the previous Mac Mini (or the one before that, or the one before that, et cetera back to the original).
For example, Xerox PARC's innovation was the GUI.
If that's your measure of "innovation", then it's hardly unreasonable that Microsoft hasn't "innovated" - neither has nearly everyone else, because most "innovations" in computing were over and done with 30-50 years ago, before they were in any position to.
Microsoft has no equivalent; they have mostly bought out other companies that did innovate and claimed they did the innovation instead; or stole an idea from someone else that did innovate.
Just like everyone else, then ?
What the GP was asking for was for a specific example of just one innovation that actually came out of Microsoft - a single original idea from Microsoft. (Note: Event Clippy and MS Bob were stolen ideas that Microsoft implemented.)
With the bar set so high, then no, Microsoft hasn't "innovated" - but that's hardly a stinging criticism in that context, since neither have Apple, Google or, indeed, pretty much anyone in the last few decades.
True, but only in very limited ways.
But these are limitations imposed by the applications, not the OS. OSes support what you want quite well, and have for a very long time, but if applications don't make use of those features, there's not much the OS can do about it.
Two trojans and a coding bug that could affect anyone ?
What about getting RID of the file selector, and just using the normal file views + drag and drop to open and save files? Drag a file to your word processor, and it opens. Drag the tab from the word processor to a disk, and you save. Drag a section of a file, and you save that section. Drag that section to the desktop, and you save a cut buffer, and you can have as many cut buffers as you want.
Most of these things have been in MacOS and/or Windows (and/or others, like OS/2) since the mid-90s, if not longer.
Doesn't matter if you are computer savvy or not, if you use Microsoft long enough and use the internet more than going to one or two trusted sites, you WILL get hacked.
I've been waiting 15 years, when is this going to happen ? The suspense is killing me...