Discussion Locked

Windows and security

0 helpful
13 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

I just read a news summary on a MS Vista advertising blurb and it got me to wondering if anyone out there can answer, in a realistic manner, exactly what and how security was improved in each of the past Windows releases. Looking back I can not identify any.

Over the years I have attended many IT shows etc and always visit the MS stand and each new OS claimed to have great improvements in security.

Win NT was twice as secure as Win 3.1
Win NT 4 was 50% more secure than Win NT 3.51
Win 2K was twice as secure as Win NT
Win XP has much more security than Win 2K
And now Win Vista is much more secure than Win XP

In each case they claimed the new one was more secure than anything else.

Now have they been bullshitting all along or was security in past system really that bad?

13 replies

by neil higgins

is er,dont connect to the internet :)

One problem is MSCAPI, which trusts lots of keys for code signing. Microsoft's model focuses on providing one build of a product that can enable weak or strong encryption at once. Although modules are not all signed by one key,as MSCAPI "allows" a large number of root certifying authorities, and "allows" multiple keys for code signing, it only takes one key to be compromised to make the whole system vulnerable to attack.
Also, the dreaded Active X,which could allow untrusted data from outside the system and can cause the authorization of arbitrary code with untrusted data.
Nothing of course is 100% secure,or fool-proof,but they could do a lot better.Vista has already had it's code "tweaked" due to a vulnerability.Oh dear!

by deadly ernest

Why have a computer at home if you are not going to connect to the Internet, no internet then buy a bloody Nintendo they are cheaper if all you are going to do now is play games.

by neil higgins

anyone remember the Spectrum + ? If we all stuck to Chukie Egg,my job,and thousands of others would be far easier.Scrap the net.
(he he)

by entawanabi

First it is assumed that you are useing a microsoft suite of things, of course then you are securer than all else,Advertising of course. What they did was find out how to be more "able" to incorporate more of the stuff from Microsoft Access into each new version.

I Don't like the answer either!

by deadly ernest

Speaking to a friend they reminded me that security has increased in each version. With each one it has become harder and harder for legitimate users to log in and access what they need of the system.

by apotheon

That's an amusing answer, and great rhetoric, but unfortunately not an accurate assessment of the situation. If it was, the discussion would be much shorter, alas.

The truth is that things are much more complex, as I'll explain in summary. By the way, I'm not entirely sure about the chronology of this. A lot of it was a long time ago, and my memory gets spotty in my dotage.

Windows 3.11 was the last separate interface install over DOS, and constituted the Snowball architecture. Snowball was in many ways more secure than anything that came after it, because it was not an integral part of the operating system in any way. Aside from that, however, security was almost not an applicable word. It just wasn't ever used in a world where malware and security crackers knew it existed.

[b]Daytona[/b] was the codename for the system architecture on which Windows NT 3.5 was built. I'm afraid I don't know much about it, and never used it. I do know that it included the partial Win32 specification that existed at that time.

[b]Chicago[/b] was the name of the Windows 95 architecture. Security-wise, it was a disaster. It was, generally, a bad idea all around. This was really the root of Windows security woes. It was a single-user OS in a multi-user world far more than anything else that came before or since. There was initially meant to be an OS based on the [b]Panther[/b] architecture (a port of NT Win32 core to run on the [b]Cougar[/b] architecture DOS 32-bit protected mode kernel), but the Win32 core was abandoned in the creation of Chicago because it was "too hard" to incorporate the superior stability and security of Win32 into a home user OS.

[b]Cairo[/b] was the NT 4 architecture. This is where the proud tradition of grand (and actually good) plans for a stable and secure OS that got whittled down to a broken small-percentage implementation of the most superficial features by release time got started. Cairo actually promised to be a really good OS architecture, on par perhaps with unix in some ways, and possibly even exceeding it in others. By the time it was released as NT 4, however, most of it was scrapped in the lab, and what we ended up with was a bucket of broken promises. It was still the best "modern" OS that carried the Windows name until the advent of Win2k, though -- and it could still be called a "modern" OS by today's standards, if a somewhat outdated one. The Cairo name continued to be in use after the release of NT 4 as sort of a sandbox for experimental technologies that would never be fully implemented in a Windows release. NT 4, meanwhile, was rather stable and secure for a Windows version at the time, until you started adding Service Packs. These, of course, started breaking stuff -- they were, for the most part, Windows 98 features painted over the top of NT 4, adding Memphis instability to partial-Cairo stability.

[b]Detroit[/b] was an update of Windows 95 for feature additions to make it more network-usable and more stable: this was Win95 OSR 2. The original Chicago architecture was flaky as a good baklava.

[b]Memphis[/b] was the architecture code name for Windows 98. It was a huge feature upgrade to Chicago, but was still basically Chicago underneath the bells and whistles. No real security improvements here, though its feature set was so increased in complexity that it introduced whole new vectors for exploit and compromise. Internet Explorer got integrated with the GUI in some surprising ways at this point, which was the beginning of a new era in unsecurable computing with Windows.

[b]Neptune[/b] was the planned successor to Memphis, but it never happened.

[b]NT 5[/b], also known as the long-awaited Cairo implementation that never really happened with NT 4, and also famously a final implementation of Win32 functionality that hadn't gotten incorporated previously, was Windows 2000. With the final implementation of all the most important features of Win32 and more than half of the good stuff in Cairo, we had the pinnacle of Windows achievement thus far. This was the beginning of true user privilege separation in Windows (though only a beginning) and other handy security-useful features. It was marketed as the merging-point of the home and business lines of Windows very briefly, but home users found all that security and stability stuff too "hard" to use, so they dropped that spiel in a hurry. One of the nice things about Win2k, ironically, was how easy (compared to other Windows versions before and since) it was to [i]turn off features[/i] that you didn't want. This was the real coup for Win2k security-wise, ultimately, since the other major security enhancements were easily balanced out by feature additions that were major security problems. Unlike NT 4, though, you didn't need to add a bunch of Service Packs that destroyed system stability to make it into a highly network-functional OS. That was a nice plus about it.

[b]Georgia[/b] was basically a posthumous honorific applied to what we laughingly call the "architecture" of Windows ME. The only reason Millenium Edition was ever conceived was because Microsoft decided they had to come up with something in a huge stinkin' hurry to market to home users as an upgrade from Windows 98, since Windows 2000 didn't fulfill that role as brilliantly as they'd hoped it would. What ended up happening was this: Microsoft skimmed some eye-candy demonstration cruft off the top of Neptune and glued it onto Memphis, called it Georgia, and marketed the whole termite-eaten outhouse as Windows Millenium Edition. Microsoft now tries to pretend it never existed, and I don't blame those executives one bit.

[b]Odyssey[/b] was the planned successor to NT 5 but, like Neptune, it never really happened.

[b]Whistler[/b] was what happened when the half-baked Odyssey and the twice-baked Neptune were, as I put it elsewhere, smacked together repeatedly and with great force until a homogenous gravel was produced. To this was added water, and the whole morass was stirred until it dissolved into a thick mud, which was baked again to produce Windows XP. It's truly a triumph of marketing over mechanics. Microsoft would have you believe it is a security and stability coup as well as a stellar advancement in high-performance entertainment computing: the truth is that Neptune had a lot more influence on it than Odyssey, perhaps because Odyssey wasn't as well-developed at the time, but more likely because the marketroids had their fingers deeply in the kettle while this one was being cooked up. The much-vaunted security features of Windows XP are just that: "features". They're application-level superficial security "features" that make it more difficult for people to accidentally stumble on ways to screw up the system without actually addressing core architectural security issues. Remember when I said Windows had a [i]start[/i] on becoming a true multi-user OS with user privilege separation? Well, this didn't get it any closer. It just made user accounts [i]look different[/i] to users so that they would seem to be more separated (for security, of course!). Rather than actually segregating memory spaces appropriately, Microsoft decided to stick sandbox kludges into the system to attempt to provide VM-like protected memory spaces, which only works as long as the software you're using follows "recommended practice" -- it doesn't actually protect against rogue applications at all. And so on.

[b]Longhorn[/b], of course, was Cairo redux: a lot of very good ideas were there, but they started throwing them out until what they eventually ended up with was a "feature upgrade". We now have Windows Vista betas floating around, basically looking like a new eye-candy confection layer painted onto Windows XP, with a rearrangement of functionality locations to make it look like a more fundamental architectural upgrade.

[b]Blackcomb[/b] is the new home of a bunch of stuff that was intended to be part of Longhorn. I'm sure the attrition rate will be no less impressive.

Hopefully, that clears up some of the questions.

by jaqui

They have one that is Rock solid
absolutely secure
and it's also a combination of 3 prior version..

The light footprint of CE
the glitter of windows ME
the salability of windows NT

it's name?

CEMENT

advertising logo:

http://www.geocities.com/rcwoolley/

by deadly ernest

Mostly horrible. You forgot that, for reasons never understood, they stick some NT5 ideas and code into Memphis to created Win 98 SE.

I still have copies of DOS 6.2, Win 3.11 and W32 that I often put on older hardware for people who want just basic Internet access. The performance and stability is incredible, but you can't use modern perihperal as you can't get drivers that work with it. But for that I can get better performance on a Intel 486 25 Mhz than you get on XP on a P4 2.4 Ghz

BTW W32 was an add on package that converted many Win 3x 16 bit service to 32 bit service thus improving performance.

by deadly ernest

When MS released NT4 they started on NT5 originally due in late 1998. Some of the ideas and stuff for NT5 got put into Win 98 and released as Win 98SE.

Meanwhile the actual delivery dates for NT5 slipped so much they changed its name when they released it to Win 2000. It was Win NT 5.1

Win XP is built on from that and thus Win NT 5.2, my latest copy has a build number of 3790, 3,790 builds of NT5.

This discussion is closed to new replies.