Slashdot Mirror
6 Months On, Vista Security Still Besting Linux
Martin writes "Great report on security vulnerabilities for MS/Linux/OS X. This is a revised version of the one Jeff Jones did back on March 21: Windows Vista — 90 Day Vulnerability Report. This time he did what the Linux community had asked. Everyone complained that he did the report based on a full Linux distro including optional components, not on just a base OS install. So this time he did both; Vista still came out on top. I was shocked that Apple was even on the list as I believed all those Mac commercials!"
Sure, if EVERY action you do prompts a "You are clicking your mouse, cancel or allow", or some other message, sure that is security, but then you are left with a crappy user experience. I think Linux and Mac have got a better balance between allowing actions in user mode without authorization and actions requiring authorization.
So what are you waiting for exactly? You could fix them today and then prove the author wrong. Oh wait, maybe you couldn't...
This game will waste your life. Don't clicky!
This is stupid, Linux as a distro is a complete solution from A-Z ... Vista is a bit of a solution as its just an operating system with limited services. Why did he do it to Vista anyway? shouldn't he be doing it to a server edition of Windows?
When i see a windows system and linux system that do exactly the same things have the same purpose software installed on them i can see the viability of the test.
Further, malware runs rampet in Windows, nearly 50% of Vista's vulns were not patched, where regardless of how many Linux has they get fixed when found. More secure? You tell me is a nightclub more secure when the bouncer only kicks out half the troublemakers whole a tougher and meaner club down the street deals with all of them?
Look, Everybody! A company is trying to use statistics to make themselves look good, when that's not necessarily the case!
Nothing to see here, please move along...
On the back of recent news that less than half of Vista "issues" have been patched, yet alone publicly announced, we get another article touting the merits of two things that can't be directly compared.
Sometimes I see Open Source kicking itself in the face with all the transparency it offers, yet I'm overwhelmed with a sense of pride and happiness that communities can develop such a transparent process in the public eye.
Discovering problems and exploiting them in a closed source product is quite a daunting task - I'd say almost 4 times as much work as exploiting a system where you can compile debug symbols into the binary, and nothing short of 1000 times harder than if you had the source code. What these "reports" and discoveries show is that layers of obfuscation act to confuse people as to the actual level of vulnerability you're exposed to.
There are many vulnerability hunters out there, now, employed by governments across the world simply to "dive in" at a deepend of closed applications looking for exploitable code - closed source simply means that only wealthy, bigger teams will be successful. Open Source means that anyone can help thwart these hunters, makes vulnerability research fair game, and most importantly, accepts community involvement into the fixing and pre-emptive policy that makes OS software better software.
Matt
One canard trotted out by MS defenders *used* to be "Windows has more vulnerabilities discovered because it's so popular, everyone attacks it!". Watch for that line to be modified in the coming months as more MS proponents switch to "it's more secure by design". Keeping the "only more vulnerabilities discovered because it's so widely installed" would imply that Vista is not widely installed/used, which is not good PR.
So, when Linux had fewer vulnerabilities, it was because it was obscure. When Vista has fewer vulnerabilities, it's because it's fundamentally more secure. I'm not trying to be sarcastic here - it may very well be *true*. It's just something to keep in mind as you watch the never-ending stream of these 'vulnerability/exploit' reports come out every few months.
creation science book
There are still a lot of problems with this 'comparison'. For instance:
- The 'reduced feature set' used for the comparison still contains a lot of software not include with windows
- All information is based on what the company behind the software discloses. I believe that not all holes in Vista that MS knows about are disclosed. It is also not unlikely that what Microsoft calls 'critical' is not the same as what Canonical calls 'critical'. In any case, different measures are used for the different OS's, and you can't compare things that are measured in different ways.
- The usual 'less known holes != safer' discussion...
I personally don't know which OS is safer, but based on these numbers, I am not going to draw any conclusions.
Jan
Why should I care whether or not people run Linux, or Windows, or *BSD, or Mac OSX, or Novell, or freakin' Amigas? At home.
Run whatever the fuck you want.
The key to the enjoyment of pop music is to replace any instance of "love" with "C.H.U.D."
He's not comparing vulnerabilities - he's comparing vulnerability disclosures.
It's not a measure of how secure the OSes are - it's a measure of how secretive the makers of the OSes are.
Wait, assuming both assumptions here are true (i.e. Windows has fewer vulnerabilities and you would fix all security problems brought to you in Linux), you would still rather _personally_ fix a lot of bugs over having a more secure platform (again, big assumption there)?
I approach this as someone who does not know a tremendous amount about how to measure security flaws, or what various security flaws really mean...
But the survey listed also shows Windows XP as the second most secure operating system of the ones surveyed.
I can believe that Microsoft improved their security with Vista. But if they also tell me their security was great with Windows XP, I have to conclude that they're fudging the numbers.
Philip Sandifer's academic website
Since Open Source rigorously discloses every flaw known in it, what is the value of comparisons of one Vendor's chosen disclosures versus that which is 100% transparent?
None
Microsoft only discloses what it has to and is often at odds with security researchers about problems only to be proven wrong later. One claim from a blog was that Vista shipped with 60,000 bugs. How many of those are documented for the public?
I can say that on my test certified Vista machine, brand new from Dell, I've already seen the network card totally disappear from the system only to reappear again an hour later. The Broadcom diagnostic tool reported no hardware issues. The Explorer shell still crashes/stalls frequently. Files get locked with no way aside from a reboot to unlock them. Wifi fails to reconnect to the same network it was previously connected to when sspi broadcast for that network is disabled. I just tried restoring a hibernated laptop, previously connected to a domain. Black screen & hard reboot.
Beyond that, on this brand new machine, specced for Vista. Vista is SLOW.
MS, concentrate on making Vista better instead of having people do useless studies. kthnxbye
Do not spread "09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0" over the internet, thank you.
https://209.34.241.68/user/Profile.aspx?UserID=780 3
No wonder Windows Vista is best in his review.
I am not convinced, next please Mr Jones.
-- "Genius is 1% inspiration and 99% perspiration" - TAE --
I suspect you've fallen into the falacy that just because people can look at the source, people actually do.
It's a fallacy? Shit. I guess that all these years that I have been working on open source software, fixing bugs, adding features, has actually been a big long dream. I'll wake up and finally see that I've been living in the Matrix, and finally see Bill G in his true Borg form hanging over me grinning...
Of course not EVERYONE looks at the source for every app, but collectively there are a HUGE number of people looking at and working with the source for just about every app out there. Unfortunately, not everyone working on open source is a qualified professional, and we do see some horrible code out there, but it's no worse than a lot of the commercial code I've seen over the years.
But back to the report. It's a shell game. Microsoft, having a closed development model, may have HUNDREDS of high threat level flaws that are UNDISCLOSED but may be known about by black-hat hackers. Open source by nature is ALWAYS disclosed. MS also has a habit of rating their flaws at a lower threat level than third party security researchers rated it. Yep, just goes to show that you can prove anything with statistics.
Here is a statistic for you... 99%+ of all the probing I get on the external side of the corp network are from windows boxes according to fingerprint analysis. Since most probing is done via compromised machines (botnet), and that windows has less than a 99% market share, that leaves me with one conclusion. The numbers are similar for spam.
How many vulnerabilities are known about and fixed in a certain time frame is meaningless. What would be meaningful, but an impossible statistic to gather, is exactly what percentage of installed Linux and Windows machines are currently compromised and being actively exploited (member of a botnet.) I've heard estimates that up to 50% of all windows machines are infected with serious malware of some sort or another...
Whenever someone says, "I can't improve the code for open-source program X, because I don't know the computer language it's written in", I tell them, "well, then you're only slightly less qualified than someone who does know the language".
Open source programs are typically not well-commented and searchable enough for a capable outsider to improve upon without significant investment of time.
Apology to Ubuntu forum.
Well problem is most common users don't want to type out a command , they , I want to say are too lazy , but more commonly they are too computer illiterate to get the syntax right and/or the spelling of the package they want. Thats why tools like windows update and synaptics (spelling?) on ubuntu are good tools for common users. And why it makes ubuntu a good choice for folks.
./configure , then make clean (in case they tried and failed before with old packages) , then make and wait for it to compile (sometimes hours), heaven forbid you have them strip and pack the binariy when they are done. to create a small binary to save size. It is really to much to ask of even a common user.
as far as source code. so many people want to pick a package install and use in minutes , thats why we have broadband and binaries. expecting a new user or even a semi experienced user to
But power users are adept to it. With issues like the above I guess I can see why Linux would be less secure then Windows Vista. Vista took the idiot out of idiot users to the best of any ones ability. Hopefully we don't see this on Linux , I enjoy my freedom on it.
This package Does Not Contain a Winner
So, how does he account for all the silent patching that Microsoft is doing?. (Link complements of Groklaw.)
More on Google.
Honestly, how can one really compare Windows against Linux when Microsoft is patching things silently? It's not a fair comparison to any vendor because you don't know what got fixed; let alone what was actually problematic. When you have one community disclosing every bug, and another disclosing only those that become high-profile for them - or likely to become high-profile since they were disclosed by others or something like that - you will not get a fair comparison.
So, if he really wants to do a fair comparison, he should get internal reports from Microsoft about their bugs, security and otherwise. Yes, CVE and similar hold the security vulnerability bugs; and you can do a comparison iff you get the security bugs that Microsoft found internally and didn't bother to report - then you would have a level set of reports.
Truth is like the sun. You can shut it out for a time, but it ain't goin' away. - Elvis Presley (source: imdb.com)
Goddammit, Sir, why did you have to post after I used all my mod points? You have provided, not only for the OSS world but developers in general, the single most important point when it comes to maintainability.
I run several servers and desktop systems. Some open, some closed. I have tons of source code, some for open systems, some for closed systems where I participate as a maintainer, developer, or reviewer. Much of the OSS stuff is unusuable except by the team that developed it. Yes, an outsider can come in, look at the code and study it but he/she is going to spend a ton of time "getting up to speed". The only batches of code that I've been able to instantly access and work with are those from projects/developers who decided that they would rather take 3 months to turn out well commented and tested code rather than take 3 weeks to churn and burn crap code that is only marginally better than old BASIC spaghetti code.
Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong fix.
The fundamental failure with the phrase "Vista is still more secure..." starts with the incontrovertible fact that Windows is shipped as a black box.
The temporary absence of security issues with Vista means nothing because neither the scope nor the scale of exploits is known. That is commonly described by the phrase "security through obscurity."
History has shown that Microsoft's approach to security is to talk a good game. Period. While I do not doubt Microsoft has hired excellent security programmers, their contributions don't make it through the management gauntlet.
Another way to highlight my point:
When you buy a windows-equipped box will you:
1: Use email on win32 without an antivirus application?
2. Go on the internet on win32 without a firewall?
3. Run win32 without a NAT?
I propose the following experiment instead:
Computer 1: Linux desktop distro immediately after install with no firewall script.
Computer 2: Vista equipped PC straight out of the box with the windows supplied firewall disabled.
Computer 3: Mac OSX straight out of the box.
Run tripwire on all three machines and put them directly on the internet. (aka no NAT)
That might be a better way to compare default security of OS's.
http://www.maxineudall.com/2010/02/should-economists-be-sued-for-malpractice.html
I believe that was mainly the point, yes.
Most end-users wouldn't recognise a security issue if it was walking in the middle of the street, naked, waving a huge vlag that had "Security Issue" printed in bold on it.
News about the Kettle Open Source project: on my blog
I inherited a whole pile of complicated PHP code recently. Not only is it uncommented, the errors it spits out are few and far between (most exceptions are uncaught and failures that are caught are merely ignored because the test is encapsulated in an if( ) statement with no corresponding else ), and when there are errors, they're merely echo statements that don't say WHAT happened, only WHERE it happened (even in those cases when the horribly designed PHP language bothers to actually have a function that returns a meaningful error).
Most of this is primarily the result of belligerent users who continually demanded unjustifiable changes or refused to provide a complete set of specifications from the outset. Fortunately, this particular way of thinking is changing and future projects will not be held captive by ego-maniacal users with unrealistic or constantly changing demands.
However, to the point, I highly doubt that any sort of management or user pressure is the reason behind the terrible state of so many open source code bases, so I have to wonder if the amateur hacking that is the nature of so many of the bit players and niche pieces of the third party support system isn't a bigger problem than the commercial issues it attempts to solve. Does adding more people really make a difference when most of those people are not even remotely close to being skilled enough to understand what role they're needed in?
I like me some BSD and OpenOffice and PostgreSQL, but those are all major projects that are designed following a very structured, traditional development paradigm. I have to seriously doubt the efficacy of designing outside of those constraints and, thus, have to seriously question just what value open source development really adds. Is it really a way to develop a better system, or is it just a "nice thing" on the side after the more traditional development processes have created a stable product that the more traditional development processes will continue to provide the bulk of support for?
I'm going to have to go with the latter. Open source is nice, open source is not really better. I think the notion that "more eyes" are on it is dubious at best, given that only a very tiny number of those eyes are going to be truly skilled and dedicated enough to actually understand what they're looking at.
You seem to be pointing the finger squarely at the developer. Most often that is not where the blame should reside.
I would point out that if you are on a deadline for delivery, things get cut. Its just business. Managers fully support good documentation, well planned naming conventions, well structured code, etc... Just so long as it doesn't interfere with getting the product out the door on time.
And... FWIW... I also have tons of source (both open and closed source) to maintain, modify, w/e...
Opinion:=TMyOpinion.Create(Me);
And if he was a Linux guy and said Linux was more secure? Let me guess, you wouldn't question it a bit.....I love how all people do is gripe about how vulnerable XP is. Well MS listened (for once) and came out with a decently secure OS. Why does the programming community insist on acting like spoiled children at every turn. Fix the bugs and stop griping! Windows guy, "My OS is better than your OS." Linux guy, "No, My OS is better!" Windows guy, "No, it's not! I have a new one and it's better!" Linux guy, "Mommy! That Windows guy said his OS is better!"
There are several fundamental flaws in the arguments in this article:
- He compares OS vulnerabilities of the first 90 days since first release. This doesn't tell us which OS is the most secure at this moment. Merely, it tells that more recent OS's have undergone more testing prior to release.
- He notes 125 known issues with RHEL prior to release compared to 0 for Windows Vista, but of course no vulnerabilities are known prior to release as Vista is closed source and has not been available for public scrutiny, while RHEL is built on available open source code.
But that's not all, differences in how bugs are classified may make some OS's appear more secure - it is known that Microsoft has classified vulnerabilities as bugs thus reducing the "official vulnerability number". Without a strictly uniform and independent classification scheme for bugs, there is simply no data to compare.
A reasonable comparison would compare the OS's vulnerability issues the past 90 days, that is with fully patched systems. Known issues that have not yet been patched should not be included as this simply is caused by the longer time for scrutiny of older OS's. Secondly, bugs must be classified in a coherent manner: Remote root, remote user, local root, local user, DOS etc...
This document is useless in the discussion of which OS is the most secure to run as of today. There is no way that a conclusion can me made in favour of any OS on the list.
It appears that OpenBSD remains the most secure system, and I bet FreeBSD is a strong contender.
Too many of these comparisons are apples and oranges things. If you run you Ubuntu box as root, you are heading for trouble. Running Windows as an administrator also exposes the user to significantly enhanced risk. If you are concerned with this risk, run as a normal user. I do. Your risk will be much lower. Vista makes it much easier to run as a normal user. My wife and kids have normal user accounts on our modern machine. I will be trying to "upgrade" my old XP box (an older Win ME box I upgraded to XP with an additional 512 MB of RAM 3 years ago) to Vista home basic for the improved security support.
Luckily, those are not the people who are relied upon to fix problems. Being able to look at the source code means that the people with the know-how to look for and fix the problems can find and fix them. One of the problems with a closed-source OS or program is that you rely on the developer of that OS or program to find and fix the problems and that means this research is going to be constrained by budget, ego, and a number of other factors.
I'd be willing to bet that OSS has a lot more competent eyes looking for issues in the code than any closed-source software, regardless of whether Joe User can read the source or not.
Why? That argument never flew with Macs on Slashdot, so why should it now?? Oh wait. Now it's Microsoft. How could I forget...
"But this one goes to 11!"
What the heck are you smoking? I'm running Vista with normal user rights (before Vista I did same with XP) and last time when I needed elevated rights was when I installed SQL Server 2005 Express some month ago. UAC prompted for administrator password, I entered it, installer continued and so on. In no way I was using higher privileges that I needed. Or do you have some magical way to install system wide components with normal user rights?
If Vista is asking admin password every other minute then you are doing some seriously wrong! There's no need for after initial configurations to elevate to admin privileges unless you are doing some system wide stuff. And if you turned off UAC go back to your Linux or whatever you like and have a nice day.
You don't know what you don't know.
This report is seriously misleading. The conclusions made do not follow from facts presented without employing logical fallacies. The data presented in the report measures amount of fixes made. The basic fallacy involves the assumption that just because a fix is not made, there no critical need for one. As a matter fact, a lesser number of fixes may indicate failure to find, report, and fix problems rather than absence of problems.
Since the Linux effort is open, all issues are reported and fixed in the open, with an effort made to report and fix as much as possible, which ensures software quality. Since proprietary systems are not open, their issues are not reported and fixed in the open. As a matter fact, a fewer number of fixes does not in itself indicate a lesser number of problems, or better software quality. On the contrary, a lesser number of fixes may indicate a lesser percent of problems being found, reported and fixed, which implies a lesser quality of software. A fewer number of fixes can be as much due to failing to fix vulnerabilities due to not finding them, or not having them reported.
Therefore, data presented in this report indirectly suggests that the open-source process is better at ensuring software quality.
it's not the holes but the number of active exploits that we should be counting.
I agree. The trouble is nobody wants to point fingers because they might get slapped. Read any of the news articles regarding the millions of bots in botnets. Every one of them I could find said "PCs". Not one article mentioned an operating system or version that was compromised. I searched Google, Yahoo, and anyplace else I could to find out if the bots had something in common such as Firefox, AIM, Flash 9, or a paticular OS. The details were sparse. If anything was mentioned it was Internet Explorer exploits and compromised websites. A search on the compromised websites gave the same generic results. About the only commonality was SQL with no mention of what flavor such as My-SQL or MS-SQL There was no mention of OS, web server or anything else. I hate thin articles when I am trying to avoid common exploits. If I can't use one SQL, can I use the other and which is which?
From the articles, I get the feeling I can't use a PC as a client with IM and I can't use an SQL enabled webserver. Other than that, there is very little hard data on botnets in the news.
The truth shall set you free!
Perhaps a part of the problem is:
When you are attacked by a bot-net, all you know is the packets you have received. Any application or OS information that these contain could easily be forged.
It *would* be nice to know. This doesn't mean that reliable knowledge is available.
I think we've pushed this "anyone can grow up to be president" thing too far.