Trojan.Milicenso: Another Reason to Avoid That Other OS

M$’s OS is still falling for a malware that has been active since 2010. This thing finds multiple ways to hide itself and in the process of setting up, creates huge print jobs that may drain hundreds of printers of paper. How many such horror stories do we have to read to know there is a better way to do IT than to accept any involvement in IT by M$?

This malware is not just malware. It exploits many intimate details of that other OS to make life difficult for everyone. It spams. It floods print-queues. It encrypts itself multiple ways. It uses files that should be data as executables. The horror that is that other OS just accepts such malware with open arms.

I recommend Debian GNU/Linux. It won’t execute data-files. It won’t open your PC like Pandora’s Box to a nightmare that could shut you down indefinitely. It won’t compromise your performance to pretend to “fight” malware while welcoming malware as an opportunity to sell more licences.

see Trojan.Milicenso: A Paper Salesman’s Dream Come True | Symantec Connect Community.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology. Bookmark the permalink.

62 Responses to Trojan.Milicenso: Another Reason to Avoid That Other OS

  1. oldman says:

    “I’m just going by memory. I’m not going to waste my time to nail down precisely what happened for your benefit.”

    So you make a statement and refuse to do the work to back it up. You are funny Mr. K.

  2. ch says:

    “That’s basically what I said, isn’t it?”

    No, it isn’t. That’s the problem: You’re unable to read, unable to write, and don’t know a thing about security. I suggest that you learn reading first.

  3. kozmcrae says:

    ch wrote:

    “Oh, and about your contention that “There is not enough of the VMS heritage in Windows NT”, that’s bullocks as well:”

    I’ve used both, they are nothing alike. If XP has NT at it’s heart, where is the security? Not in XP, not in NT. Where is the security now? Oh, that’s right, there’s a multi-billion dollar security industry supporting Windows. That’s the security. Where’s the security in Linux. It’s built in, that’s where.

    “Later, he worked on targeting Windows NT to Digital’s 64-bit Alpha architecture…”

    That’s basically what I said, isn’t it? He left to do other things. Things that went nowhere and accomplished nothing. Meanwhile, NT went on without him. I’m not Googling this, I’m just going by memory. I’m not going to waste my time to nail down precisely what happened for your benefit.

  4. ch says:

    “Not sure why you even care.”

    I thought it fitting to show how you make up stuff, that’s all. Quote from WP:
    Cutler left Digital for Microsoft in October 1988 and led the development of Windows NT. Later, he worked on targeting Windows NT to Digital’s 64-bit Alpha architecture (itself based on the Prism design),[citation needed] then on Windows 2000. After the demise of Windows on Alpha (and the demise of DEC), Cutler was instrumental in porting Windows to AMD’s new 64-bit AMD64 architecture. He was officially involved with the Windows XP Pro 64-bit and Windows Server 2003 SP1 64-bit releases.
    (endquote)

    Where’s your evidence that he moved on after a short time? Oh, and about your contention that “There is not enough of the VMS heritage in Windows NT”, that’s bullocks as well:
    http://www.anonymous-insider.net/windows/research/1998/12.html

    (Mark Russionvich does actually know Windows.)

    So I suggest you stop wasting your and our time by making up stuff.

  5. kozmcrae says:

    ch wrote:

    “He was still working on Windows versions until 2006…”

    Not on NT. He moved on after a short time in that project. Not sure why you even care. It doesn’t make a bit of difference one way or the other. NT was still a piece of she it. Compared to Linux that is.

  6. Linux Apostate says:

    “False logic. You have to do rootkit hunting and other attack prevention on Linux. Linux servers are very regularly attacked by malware writers.”

    Oh, I agree. But then, unlike kernel.org, all my machines are behind a firewall with no externally accessible services. An attack on this Linux VM would have to be delivered by a Firefox exploit (Flash is disabled, naturally), or by some sort of social engineering attack. I know it’s possible, but it is unlikely, and since no malware author has yet bothered to mount a drive-by download attack against a Linux web browser, I am also safe from the majority of automated attacks. For now, anyway. I am sure that if a drive-by attack is ever carried out on Linux outside of a security research lab or hacking competition, Slashdot will tell me…

    On the subject of NT, I do agree with you that it’s the outer layers of user interfaces, services, and compatibility code that are responsible for the security problems. Linux mirrors the same thing. Maybe it is better, maybe not, but there’s no doubt we would be much better off if applications could be properly isolated from each other in secure sandboxes, so that an exploit in Firefox goes no further than Firefox.

  7. oiaohm pulled a fast one with, “Remember the kernel.org and linux.com breaches by malware. Linux is not past attack. That infection today would not work against upto date distrobutions. Linux out evolves its attackers.”

    Kernel.org promised a report which they have not yet delivered. Apparently official investigations were mounted and until concluded they do not want to reveal anything. I think that’s a mistake because the bad guys knew what they did but we, the good guys, do not. The issue seems to have been weak/compromised passwords, not necessarily malware or a vulnerability. Any distribution with an open SSH port is open to attack by humans, not malware, if the passwords are known/weak. What is not known for sure is whether root was allowed to log in by SSH or a user had to escalate privileges.

    Given that kernel.org had a bunch of users with accounts and did not run lots of PHP etc., I suspect that the problem was limited to weak/compromised passwords and not necessarily anything that is worrisome to others. My site which has lots of haters and runs PHP should have been rewritten dozens of times if GNU/Linux servers were terribly weak. OTOH I have very few accounts and strong, very strong, passwords.

    It’s not as if a script-kiddie can compromise 10K GNU/Linux servers on an afternoon without really trying, the way they can with that other OS.

  8. ch says:

    Mr Pogson,

    I am still assuming that you are an adult, so please refrain from posting links to childish stuff like that in order to prove anything. I mean, seriously:
    “A 64 bit upgrade to a 32-bit patch for a 16-bit GUI shell running on top of an 8-bit operating system …” He got the “64-bit upgrade” kind of allmost right, the rest is just nonsense.

    Need more examples? “NTFS (New Technology Filing System) which is of course a Windows 2000 invention” is just as wrong, NTFS was introduced with WinNT3.1 (the very first version).

    “Before you check your bank account to see if you even have the money for Windows XP”

    Great, you’ve found another penny pincher.

    “Find out which OS is still standing after this Windows XP vs. Windows 98 and Me fight club.”

    A submachinegun-vs-pocketknife contest? The good news: The link is broken.

    Please try to find stuff that could at least be remotely considered serious.

  9. oiaohm says:

    Linux Apostate “Support for legacy apps was added later.”

    The process of adding the legacy support ruined NT. So to someone who did not see NT 3.5 and before saying what Robert Pogson I understand. Because adding the legacy support added lots and lots of the 9x series defects to NT.

    “DOS which knew no users/just a single user” This bad habits were imported into nt in the process of adding legacy support. So to anyone who knows the bad habits of 9x looking at NT after 3.5 sees them really clearly. To the point that the layers of junk of the compatibility layers hides the secuirty designed kernel from them.

    “And actually I do use Linux in Virtualbox because malware authors don’t target it.”

    False logic. You have to do rootkit hunting and other attack prevention on Linux. Linux servers are very regularly attacked by malware writers. Linux due to have more modern secuirty proper setup Linux is more resistant. Remember the kernel.org and linux.com breaches by malware. Linux is not past attack. That infection today would not work against upto date distrobutions. Linux out evolves its attackers.

    Compared to current Linux networking stack the current nt network core is dumb very dumb. Lacking massive ammounts of logic. Like lacking zone/group/jail virtual networks.

    So nt current network core is dumb compare to current secuirty ware Linux, BSD and Unix systems.

    The big critical point NT core is design to be secure. The MS implementations on top of that core are crap.

    “Dave Neil Cutler” did not have final say.

    Linux did not start inherent security. But it has evolved over the years progressively getting more and more secure and more and more resistant to attack.

    NT started with a head start because it was formally designed secure and has fallen behind because it has failed to keep its secuirty systems evolving. This could be a problem of the limited size of Windows Development teams not getting enough ideas compared to BSD or Linux.

    Remember I said secuirty is sum of parts problem Linux Apostate. One good part does not make up for being stuck with 100 bad ones. This is the NT problem its got a few 1000 bad ones it picked up. win16 got from from 64 bit windows was a good bit of crap removal. Problem is there is still a lot more crap that needs cutting out.

    A Linux install today out is inherently more secure than a Windows 7 or 8 install out box on enabled and correctly set-up secuirty features.

    Linux has inherited its secuirty from all the years of work.

  10. ch revised history with, “Somebody tricked you there: If it needed daily reboots, it wasn’t XP. Trust me on that, for I have actually used XP.”

    Here’s a review from 2001:
    “Oh, you heard that Windows XP won’t crash? You heard wrong. The architecture of Win XP is a big improvement over previous version to be sure. If you experience, say, 5 crashes a day, then most likely Windows XP will only crash about twice per week. An improvement for sure. But if you are thinking that Windows XP is crash proof, ah….ya really need to rethink that.”

    And that was in the days when I ran GNU/Linux non-stop on a cluster of PCs in my classroom for six months… Instability was why I left Wintel.

  11. ch says:

    > Viktor 1993 NT did not support running dos applications. When that support was added a few extra holes were added to NT. Also 1993 NT did not support running win16 applications.

    > As support for OS/2, dos and win3.11 get added (…) you watch the secuirty quality drop very quickly

    In fact, WinNT supported DOS, OS/2 1.x textmode and Win16 apps (as long as they didn’t try to do “illegal” hardware access) from the very first release. And as far as security is concerned, look at the link you posted. Quote:

    > MS-DOS is completely unsecure. (…) Windows NT solves this problem by protecting itself from VDM’s and by running each MS-DOS application in its own VDM (so that they’re protected from each other).

    However, I’ll give you that the OS/2 subsystem was later dropped because of security concerns (and of course because it was no longer necessary).

  12. ch says:

    > They did. He didn’t stick around too long

    His name is Dave Neil Cutler. He was still working on Windows versions until 2006, then went on to other work within MS. It looks like you pulled your “facts” out of thin air.

  13. ch says:

    Mr Pogson,

    another post got swallowed by the spam filter. Could you please … ? Thank you.

  14. ch says:

    > You have security on the radar when you imagine multiple users and multiple processes running on a system.

    Not when those multiple users are your _friends_, you don’t. Unix was developed in a very friendly environment, and so they could easily live with a “security design” that amounted to a “Don’t disturb” sign on the door (instead of locking it). I was a concious design decision that was right at the time. Of course, if such a “friendly” system ever gets into a hostile environment …

    > The innocent user could not mess up another innocent user by accident.

    On a _Personal_ Computer, there _is_ no other user. At least it most definitively wasn’t in 1980/81 on a non-networked 8088 with 64k of memory and two FDDs. So making DOS single-user was a conscious design decission that was right at the time. Of course, if that PC ever gets connected to a hostile environment … But not only did nobody in 1980 have any idea that 15 years later the internet might become a mass phenomenon, I’m sure nobody even believed that DOS would still be around in 15 years.

    BTW, MS even built – and sold – a unixoid OS for the PC, but it wasn’t very successful.

    > Indeed, M$ did not even try to get multiple users going

    Xenix was multi-user, so MS did have a multi-user product out in 1981 – but it wasn’t successful on the market.

    In the PC world of the late 1980s, multiple users were best served with – big surprise – a server, like – for example – OS/2 LAN Manager from Microsoft. WinNT followed in 1993.

    > M$ kept distributing the totally insecure stuff until ~1999

    … because of people like you, who couldn’t be bothered to get the more secure stuff.

    > they surely did not hurry

    Huh ? Again, WinNT was out in 1993! YOU were obviously not in a hurry if you didn’t bother to even look at NT until 2002!

    > I used Lose 3.1 from 199x until 2001 when I went to GNU/Linux and I first saw NT4 in the workplace in 2002

    Now we all have a better insight into your competence: You still used 3.1 in 2001 ??? You never even looked at NT (or at least Win9x), although you had so much problems with 3.1?

    Curious question: When you set up a network in a school, prior to 2001, which server OS did you use?

    > XP needed daily reboots when I first began to use that.

    Somebody tricked you there: If it needed daily reboots, it wasn’t XP. Trust me on that, for I have actually used XP.

  15. Linux Apostate wrote, “It’s just not true for the NT family.”

    M$ imported lots of bad stuff from the old “Lose” line which gave malware writers another decade of laughs. Even “7” which is supposed to be absolutely wonderfully secure has bugs descended from Lose ‘9x.

  16. Linux Apostate says:

    oiaohm’s post is informative. Notice that he confirms that NT was not based on DOS/Win16, but was instead designed to be secure, multiuser, etc. etc. Support for legacy apps was added later.

    I hope that this brings some sort of end to the FUD that has been posted on this subject. Let’s not hear more about how Linux has inherent security “from the first line of code” but NT “has single user, network dumb code at its core,” or “That other OS was designed on DOS which knew no users/just a single user”. It’s just not true for the NT family.

    That doesn’t mean there aren’t good reasons to use Linux. And actually I do use Linux in Virtualbox because malware authors don’t target it. I know I’m not safe from malware because I use Linux, but the risk is lower for people willing to put up with Linux. I’ve been a Linux user for a very long time; I gave up the zealotry, not the OS.

  17. oiaohm says:

    Linux Apostate problem here I know NT design inside and out.

    The simple problem its not good enough any more. It needs major work. Has needed major work for the past 10 years. UAC does not solve the base problem. NT is designed to look at a system as a stack of users.

    Secure current day secure OS’s are required to look at a system as a stack of applications being run by a stack of users. With applications having there own secuirty limitations on top of user secuirty limitations to slow down assault.

    The design being outdated and obsolete is one of the problem with NT based systems. Next is that is poorly configured out box.

    Linux today from a secuirty stand point is many times stronger than Linux 5 years ago. Until the year 2000 Linux was very much following path of Unix. At 2000 Linux started going its own path on secuirty.

    Microsoft upto and include XP was making a mistake unix systems had taken out there install systems in 1987 that Linux systems took out there install in 1997. This is the fact that installing those old Unixs it was mandatory to set up a limited user account on install. Vista was the first time this appeared in Windows. 2006. Basically 10 years late to the party compared to linux and 20 years late compared to Unix.

    Instruction of a built in firewall. This appears in Linux 2.0. This is 1996. Unix’s built in firewall appears over 10 years before Linux even appears.

    Half functional built in firewall appears in XP 2002 and full function you have to wait until Vista.

    As you keep on charting secuirty features when they are added to each OS. The features that did not exist in NT base design are being added on average 10 years behind Linux. Please note firewall was in BSD when it was release open source 1991 it was also in BSD long time before that.

    Yes Linux compared to BSD systems has also been very late to the party of security and has had to play catch up. Thing is that Linux has caught up.

    The most annoying part for me is the primary designer was a down right futurist and got the base design of NT right just today it needs some updates to be modern again.

    http://technet.microsoft.com/en-us/library/cc767884

    The key problem is documented right here.
    “If one subsystem does crash, the others are unaffected (unless it’s Win32, which handles keyboard and mouse input and screen output for all subsystems).”

    The base NT design does not have Win32 running keyboard mouse and screen. That was a modification done for performance not secuirty. If NT design worked as it was designed windows would have no problem sandboxing applications. Just run each application in its own private instance of the subsystem.

    Linux with cgroups and Solaris with Zones are implementing what the great designer of nt Dave Cutler designed yet his fellow assistants stuffed up.

    Even BSD implements its own version of subsystems called jails. This is the problem Dave Cutler was and is brilliant problem the people who have worked around him are not brilliant so a lot of bad things have been done to his design making it a secuirty nightmare.

    Object based secuirty in NT this appears in Linux Secuirty Modules and Solaris Trusted Extensions basically 7 years latter. When you know this NT could be leading the pack in secuirty design but its not.

    Worst part is object base secuirty out box in Windows even that the framework is there is incorrectly configured to be effective. Under NT there is no reason for any service to be running as user SYSTEM none at all. NT was design that it was completely not required. Linux, BSD and Solaris have all three had to alter there core design to support this.

    This is the shocking part NT OS’s had a lot of secuirty firsts. Even that the kernel has a lot of secuirty first out box today they are not implemented. So leaving Microsoft provided NT OS’s many times more open to assault than they should be. This is why Microsoft needs a serous kick up ass. Please provided us with a NT based OS that is configured correctly should not be a hard thing to request really.

    Basically open up task manger in windows show all process and every one showing user system bar idle is a configuration error. Yes you don’t have to look hard to see current day NT based OS’s are poorly configured.

    Viktor 1993 NT did not support running dos applications. When that support was added a few extra holes were added to NT. Also 1993 NT did not support running win16 applications. It was the Dave Cutler design back then and it worked and it was secure even the services had individual users. That is NT at one of of its secuirty peaks and its an area that Linux distributions are only starting to fully catch up on with systemd.

    As support for OS/2, dos and win3.11 get added what was developers that had been working on OS/2 and Windows 3.11 added to the NT development team you watch the secuirty quality drop very quickly. Problem is lot of those drops have never recovered.

    Both of your victor and Linux Apostate are argueing with emotion. Not thinking for one moment that NT has been stuffed up by configuration and some of the alterations to support Dos and win16. Yes the win16 subsystem was removed from 64 bit windows to address some of the secuirty nightmares it added under the lie it could not be made work. Wine has no problems running win16 applications on 64 bit systems.

    Robert Pogson you are a little bit shallow on your knowledge. By the time you saw NT 4.0 it was already ruined. There is a great core to NT based OS’s. Problem is secuirty is a sum of parts problem. Its like having a motor bike with the world best quality engine but the frame is rusted badly so as soon as you apply force to it the bike will fail then a person comes along and coats in in spray paint so it looks new.

    Linux all part have not be great but they have been all operational from a secuirty point of view due to being configured correctly.

    As some people say GNU/Linux is a Tank of OS’s. Tough but gives a rough ride and can go almost anywhere without much worry other than the personal pain. Linux is not a nice looking thing. Most people don’t want a tank parked out front.

    Windows is the lemon car that looks good don’t hit any pot holes causing all its hidden defects behind the paint to show themselves mostly because people have forgot to do up bolts and other important things that should be holding the car together.

    Apple is that expensive sports car that is fine unless you want to go down a road with some speed bumps.

    What we really need is one truly functional car. Android is closer to the one truly functional car but in some places a few bolts have been untightened too far.

  18. kozmcrae says:

    Linux Apostate wrote:

    “I understand they even recruited an ex-VMS OS designer from DEC to help build it.”

    They did. He didn’t stick around too long and the project was taken over by Microsoft old-timers. There is not enough of the VAX heritage in Windows to say so. It’s just a story to make the Microsoft faithful feel good.

  19. I understand very well that NT was an attempt at redesign but the salesmen running the ship felt ordinary folk would never stand for a buttoned-down OS so they watered it down and we got XP, the greatest vector of malware ever devised by man. What happened to all that security in NT 5.1? From whence came the default firewall, autorun, executable images and all that crap that was not in NT4? They made insecurity backwards-compatible…

  20. Linux Apostate says:

    You misunderstand. NT *was* the redesign; a complete break from Windows 3.1 and WFW, with compatibility only at the application layer. The new kernel was designed to be competitive with Unix in terms of stability and security. I understand they even recruited an ex-VMS OS designer from DEC to help build it.

  21. kozmcrae says:

    Linux Apostate wrote:

    “I think that both Mr Pogson and kosmcrae fail to appreciate the extent of the NT redesign.”

    I’ve worked with both VAX and Data General systems. That’s where I learned to appreciate what a real multi-user, security built-in system was. Like you said in so many words, NT was playing catch up.

    Microsoft didn’t start out making heavy metal systems. They came at it from the single user, network dumb mindset. They built NT up in layers. That’s the way with all of Microsoft’s products. It’s the way the company is set up. They don’t know how to do it any other way.

    Microsoft’s code is a mess. Don’t take my word for it. Their own developers have said as much. They don’t know what will happen when they change a few lines of code.

    You say NT is “redesigned”? I say your definition of “redesigned” needs a little work. Microsoft didn’t start from scratch and create a new redesigned NT. They just added more code. More layers. More crud. How many lines of code is Windows up to now? Vista was at 40 million. I think they cut it down a little with ‘7’. What’s up with ‘8’. More crud or less?

  22. Linux Apostate wrote, “NT was never intended to be single-user and it’s just FUD to make out otherwise.”

    I know very well M$ tried to inject new DNA into TOOS but they surely did not hurry. I used Lose 3.1 from 199x until 2001 when I went to GNU/Linux and I first saw NT4 in the workplace in 2002. It was scarcely better. We were forever having to reboot things and only the IT guy was allowed so we often went without when he was not at home base. Printing was a crapshoot. XP needed daily reboots when I first began to use that. There are valid reasons why I tried GNU/Linux. I was on the Wintel treadmill far too long.

  23. Viktor wrote, “Whenever you write something, Bob, one can be sure that the cold, hard facts will set the record straight. You have such an amazing capacity for self-deceit, Bob.”

    Pfft was even more insulting, “You are on some seriously good drugs, Mr. Pogson.”

    Let’s set the record straight. I don’t lie and I don’t do drugs. I don’t even drink wine/beer. I don’t imagine myself great/wonderful.

    SO, Goodbye to both of you. It has not been fun…

  24. Viktor says:

    That other OS was designed on DOS which knew no users/just a single user.

    Wikipedia says:

    Windows NT is a family of operating systems produced by Microsoft, the first version of which was released in July 1993. It was a powerful high-level-language-based, processor-independent, multiprocessing, multiuser operating system with features comparable to Unix. It was intended to complement consumer versions of Windows that were based on MS-DOS.

    What can we learn? Windows NT was not based on MS-DOS.

    Whenever you write something, Bob, one can be sure that the cold, hard facts will set the record straight. You have such an amazing capacity for self-deceit, Bob.

  25. Linux Apostate says:

    UAC is on and my user account type is “Standard User”. I have another account for admin. This is the configuration recommended by Microsoft during installation and within the Control Panel. If I try to do something that requires root, the machine asks for a password like gtksudo does.

    I think that both Mr Pogson and kosmcrae fail to appreciate the extent of the NT redesign. It was intended to support multiple users and multiple processes from the very beginning. And as Mr Pogson said –

    “You have security on the radar when you imagine multiple users and multiple processes running on a system”

    Before Citrix and Terminal Services, it was intended that the multiple users would use thick clients (workstations) and authenticate centrally on the NT domain (later Active Directory). By authenticating they would gain access to shared resources such as folders and printers. These facilities were in the very first versions of NT. Users could also roam, because their login profile was held by the server (“in the cloud”, we would now say), and multiple users could share a single workstation (“hotdesking”). I know it is hard to believe that Microsoft was doing all this twenty years ago, but history is history, and to be honest they were playing catch-up to Unix and VMS at the time.

    So please, stop with all of this nonsense about users changing system files and the whole thing being based on DOS. NT was never intended to be single-user and it’s just FUD to make out otherwise. And we’re all against FUD, right?

  26. kozmcrae says:

    Linux Apostate, since you are not a hypothetical user, do you have the UAC on or off? And, do you run as Admin or just a normal user?

    Please don’t embarrass yourself by throwing a bunch of words at me and try to weasel out of it. Just answer the questions. They are simple, legitimate questions too. You should back up what you write with what you do.

    I only execute root to run single commands or applications. And that’s the difference between the UAC and root, and Admin and root. Did you catch that? UAC/Admin vs root. Linux is simpler and more direct. No BS layered crap.

  27. ch writing of UNIX operating systems, wrote, “However, it was _not_ conceived as a _secure_ OS – that only came later, essentially as an afterthought.”

    You have security on the radar when you imagine multiple users and multiple processes running on a system. There was essentially no security in the hardware of the time but UNIX OS was a start. The innocent user could not mess up another innocent user by accident. That happened all the time with that other OS. Indeed, M$ did not even try to get multiple users going until Citrix made a business out of it.

    The idea of shadow passwords was included in UNIX operating systems a decade before a similar idea was popular with M$. NT had it from the beginning but M$ kept distributing the totally insecure stuff until ~1999, a decade after UNIX OS had the concept. Imagine the world operating for a decade with a monopoly pushing insecure crap.

  28. Linux Apostate wrote, “Notice that this is about design, not implementation.” as if he rules.

    Software is data-structures and algorithms. Once designed, software is as good as written. M$ designed their software in the Lose ‘9x era with little or no security. It was not even on their radar and when they did add security, they put in so much backwards compatibility that much security was bypassed. M$ wanted software to be easy to install and routinely allowed installers to tweak everything… That was by design, not some bug.

  29. oiaohm says:

    ch suid bit is counted by capabilities and Linux secuirty modules.

    Yes secuirty in Unix has been evolution. The problem is the security in Unix and Linux world is 2 to 3 generations ahead of what MS is offering at the min.

    Year 2000 the introduction of selinux was partly to address the suid bit problem.

    linux capabilities 2004 another reduction method to that problem.

    Finally cgroups user namespace even more limitation to that path. Since using this user id 0 to the application might not be.

    About time you read some more upto date books ch.

    There is a suid bit equal problem with user started services inside windows even worse is some ole sevices.

    The Cuckoo’s Egg problem has been address in the Unix and Linux world. There are a lot of places inside windows where it has not been addressed. That book is 1989 the tech has moved on a lot since then in the Unix world. Current windows is very 1989.

    /etc/passwd was first based on the secuirty flawed idea that hashes could not be have collisions falsely make. This was address quite a long time ago with the introduction of the shadow file.

    Mind you its possible under windows from a limited user account to to extract the hash of all users on the machine from the sam file. Yes the /etc/passwd problem still exists in Windows 7.

    How is really funny is how. ntbackup. Limited user can trigger a ntbackup so make a back file containing the sam. Now extract that in there own directory and override permissions hello one free copy of the sam when you are a limited user. Same method can be used to take the disc copy of the active directory out a windows server.

    About time you step back and attempt todo the old Unix style attacks against windows. You start finding MS has a lot of work todo and some really funny and stupid bugs.

    Linux is simply a few generations ahead of where windows is. History of problems with Unix and its problems puts Linux ahead in the security game.

  30. Linux Apostate says:

    “NT has had huge numbers of vulnerabilities, many of which relate to code written in the DOS-era. e.g. animated cursor and executable images, bypassing any goodness inserted in NT. How about that auto-run? Good stuff, eh? If you’re malware”

    Ah, but the question I asked was “Are you claiming that Windows has been designed to allow an unprivileged user to modify system files?”

    Notice that this is about design, not implementation. Was the system *designed* to be insecure? No. Was it *implemented* insecurely? Yes. But this is a different matter.

    Windows NT was designed to enforce its access control. Aside from implementation errors, it does. And the implementation errors are fixed when they are found. What more do you want? This is the best you can expect from Linux as well – Linux is also designed to be secure, but in practice there are vulnerabilities. (Google “linux kernel privilege escalation”.) It’s just FUD to make out that it’s insecure by design.

    And it’s outright false to claim that, by design, an unprivileged user can modify system files on Windows.

    (Another fun thing to type into Google – “USB Autorun Attacks Against Linux”. Is this insecure by design, or just a UI bug in a Gnome subsystem?)

  31. ch says:

    > UNIX (…) was conceived as a general-purpose, multi-user, and multi-tasking OS.

    Right. However, it was _not_ conceived as a _secure_ OS – that only came later, essentially as an afterthought. (Check out the history of /etc/passwd)

    A still remaining bit of trouble is the SUID bit, BTW. Have you read “The Cuckoo’s Egg” ?

    > That other OS was designed on DOS which knew no users/just a single user.

    Oh, now I understand. You have necer worked with anything better than Win9x – no WinNT, no Win2k, no WinXP. That explains a lot.

    > Backwards-compatibility

    As Linus has recently discovered, that’s a good thing.

    > XP SP2 was forced on us one night and apps were broken.

    Because they did things a security-conscious OS couldn’t allow them to do? Good.

    @Victor:
    > I recommend Paulaner Salvator

    Nope, it’s Augustiner Edelstoff that you want 😉

  32. Viktor says:

    I like this trojan. It’s a friendly reminder to pursue a paperless office.

    I recommend Paulaner Salvator. A great beer. You might even enjoy GNU/Linux when you’re drunk. Once you’ve sobered up you’ll remove it from your PC as fast as you’d kick out an ugly one-night stand.

  33. oiaohm says:

    Linux Apostate read mine please. User does not have to enable UAC to admin. All user has to be doing is running an application with a bad upgrade service from there the attacker can take the system.

    All Linux package management forbids packages overlapping files. This is why .d directories exist. To contain extensions to service settings. One of the big problems in windows is applications are allowed to play with other applications settings. So stuff like flame running lua scripting can download a script to search for paths to get deeper into the system and find a lot of doors.

    A flame like viruses have existed for Linux even Unix before Linux existed. This is what Linux Security Modules are designed to slow down or find viruses that at core are a remotely controlled scripting engine that can change there signatures and attack methods instantly. Notice anti-virus companies say they cannot really do anything about flame style infections. Yet Linux has been successful at neutralisation of this kind of infection. You need application secuirty to detect something like a flame infection without signatures.

    You are still only talking about user level secuirty breaching. Not like the tech selinux implemented in the year 2000 where application to application is restricted.

    NT design is old and outdated in major need of updates. Linux Apostate. But the true killer is that the Linux system has had to take in a feature of NT. NT has ids unique to kernel not exposed to userspace. To allow subsystems to operate side by side. Linux has had to implement that recently to allow cgroups to isolate services.

    NT design with a major overhaul could be made decent again. First thing would be allowing multi win32 and win64 subsystems to run side by side with each other. Then services that user does not need to interact with being placed in a sandbox isolated from user. Then registry altered to be per application registries so that applications cannot alter settings of other applications without explicit permission.

    Linux Apostate NT design was ahead of its time. Problem is its not longer. User to User secuirty is an out dated model. Application to Application secuirty inside User to User secuirty is the modern model inside some form of watch dog is modern.

    Linux has been in the for ground with solarais designing secuirty frameworks of the modern type.

    Problem is NT core design also got broken to support old code and due to lazyness. Like the fact the posix system in windows could not stand on its own too feet.

    If you dig out the early coding instructions they only interlink point between subsystems should be ntdll.dll. Problem its not. So cross talk between subsystems are required for operation these days.

    On Linux interlinks between subsystems is not there. Lot more stuff is self contained so simpler to secure. Surface area problems.

    Linux Apostate your example against Linux should not work if apparmor or selinux was left on. Linux does suffer from the dumb user who turns the secuirty system off.

    systemd addresses some of the problem of dumb user who turns the secuirty system off. Sections of the secuirty system that with systemd you turn off you cannot login. Problem solved.

    This is the thing. Application asks for root user. Under Linux I can place that application in cgroup it will be told it has id 0 what is root. Reality its has permission of some other user of my choosing.

    This is the big thing about Linux current day secuirty. Linux can lie to applications. What ever information you could possible probe that tells you that you are running as root user. Linux system can fake. Attacker does not know if there program has succeeded to privilege exploit or if they have been just directed into a honey pot trap.

    UAC Linux Apostate this would be Linux sudo. Please take a look at policykit sudo replacement. Yes not everything on linux has migrated to policykit yet. This is why I say its a generation behind. Tech MS implemented was out of date when they implemented.

    Even running as root. Running a webbrowser and visiting and infected site may not see a Linux system toast. Why because if selinux is enabled webbrowser could be auto sandboxed and as soon as the attack attempts todo something the browser was terminated. This is not dependant on the coder of the application to implement the sandbox. Linux takes responsibility for Sandboxing the application it runs from its Linux Secuirty Modules in current distributions. Newer using systemd is cgroups and Linux secuirty modules sandboxing. Cgroups can automatically be placed around programs like webbrowsers using ulatencyd in systemd to prevent access to files user wants protected. To the webbrowser any hidden directory simple does not exist.

    Running as Administrator should not be enough to give 100 percent certainty of taking out a system. Other layers of secuirty on the application like Linux secuirty modules should kick in. Windows lacks this backup layer. Windows depends on the application itself to implement sandboxing. Linux has been weak in application self implemented sandbox great in system wide implemented sandboxing.

    seccomp filter that coming makes Linux great at application self implemented sandboxing as well. This is the last area in secuirty design that Windows had any advantage left.

    Linux Apostate so my question why in hell running as administrator under windows absolute sure disaster when running as root under linux is only a high risk of disaster. For root to be as high risk as administrator under windows the Linux Secuirty Module has to be disabled. Basically the equal to running without an anti-virus under windows is running without a Linux Secuirty Module to Linux people. Thing is Linux Secuirty Modules do not depend on signature to detect threat. They know what applications should be doing and if an application does something not right reacts normally by termination.

    This is what is lacking from windows. Means for windows to detect infections without signature of infection from real-time monitoring. To be correct Linux is using white list signatures what is a signature of the application that should be running and how it should act. Where anti-virus uses black list signatures trying to find what should not be there.

    Correct secuirty requires both black and white lists. Linux has access to both list types. Windows only has black lists. Yes there are reasons why windows is a sitting duck. Windows need a lot of work. trusted extentions in Unix systems are based off the work for selinux that implement white list checking against applications. Application do something that is not on its white list terminate it. This is a massive attack surface area reduction.

  34. Linux Apostate wrote, “The whole point of the NT redesign was to go from single user to multiuser, by introducing proper access controls and VMS-like permissions.”

    The whole point of Wintel is to make the user comfortable with what they are using so they won’t notice its flaws. NT has had huge numbers of vulnerabilities, many of which relate to code written in the DOS-era. e.g. animated cursor and executable images, bypassing any goodness inserted in NT. How about that auto-run? Good stuff, eh? If you’re malware…

  35. Linux Apostate says:

    kosmcrae initially told us all that “[Linux’s]… security is built in, inherent would be the correct word to describe it. Microsoft’s OS isn’t. It still has single user, network dumb code at its core.”

    After being challenged to justify this, he has now retreated to the argument that Windows is insecure if the user disables UAC and runs as administrator, which (he claims) Windows users do, and Linux users do not.

    Well, anything is insecure if you switch off the security features. koszmcrae’s hypothetical Windows user is smart enough to disable UAC and promote himself to Administrator, but stupid enough to ignore the warnings about doing this. The only question is – will he be infected with malware, or will he win a Darwin Award first?

    Meanwhile Mr Pogson chimes in to tell us that Windows is too much like DOS to be rid of the DOS legacy of no access control. For his information, no version of Windows NT has ever been *designed* to permit a user to modify system files. The whole point of the NT redesign was to go from single user to multiuser, by introducing proper access controls and VMS-like permissions.

  36. Linux Apostate wrote, “Are you claiming that Windows has been designed to allow an unprivileged user to modify system files?”

    Yes. That other OS was designed on DOS which knew no users/just a single user. Backwards-compatibility brought that ethos forward many years. Malware loved it. I remember in 2004 having our lab quit because XP SP2 was forced on us one night and apps were broken. A decade or more of that laxity is hard to put back in the bottle and a lot of spaghetti code results trying to cover the holes.

  37. oiaohm says:

    Ted on cert security count for Linux includes all applications in the distribution. Where windows only is windows. So the bias is back the other way. Yes you do have to watch the bug count numbers carefully.

    Linux world also complains about bloat. Bloat links to memory usage.

    SYSTEM account you become by accessing a service. There are too many services running as system.

    “There are also some system folders an admin user is locked out of, as they are only accessed by the SYSTEM account.”
    There are tones of ways of getting system account.
    http://www.pcper.com/news/General-Tech/Firefox-12-will-be-able-bypass-UAC-and-possibly-corporate-security-settings

    This is a major surface area problem on windows 7. How many programs that bypass UAC just to be able to update self. Each adds a extra bug points.

    Package management in Linux has been battled between distrobutions but its required for secuirty. I hope UAC hole making would reduce with windows 8 but windows 8 still lacks a means for applications to register there own repositories.

    ted
    “The same way installing malware on Android is “social engineering”, but on Windows it’s the fault of the OS.”
    Lie. Android has a rule applications cannot access other applications settings without direct operator permissions. This prevents side to side spreading as simply. Applications under android are not allowed to touch each other configuration files or binaries and unless given permission not allowed to even touch user data.

    Android social engineering is harder. There is more grants to get application in than windows. Of course you miss the fact that downloaded data under android is forbin to be an application. Windows data in the user directory can be a .exe and run without any extra approval. Then do assessment on system and find a hole.

    Android is running per application secuirty not per user secuirty. Per application secuirty is also what selinux does on Linux systems so a webbrowser altering $HOME/.profile by Linux Apostate should fail on Linux running a proper selinux setup or android since .profile is not a file webbrowser should touch.

    This is why windows is weaker. Per user secuirty vs per application security. Per user secuirty is crap against the current malware problem leaves attacker option of using every path installed on the system.

    Seccomp filters and systemd will hardden Linux even more. Linux is already a generation a head in security designs due to LSM sandboxing around applications. By the time of windows 8. MS will be 2 generations behind. Why because double layer sandboxing of services will be in most Linux systems by the time windows 8 releases. One layer by the LSM and one layer by cgroups done by systemd. Then there will be a third coming in the form of Seccomp filters that has to be coded into applications.

    Each sandbox a attacker has to get past to get into the system. Sorry to say windows does not put that many hurdles in the way. Under windows all you have todo is defeat user secuirty. You don’t have any application secuirty to defeat.

  38. kozmcrae says:

    Linux Apostate wrote:

    “Are you claiming that Windows has been designed to allow an unprivileged user to modify system files?”

    Quite the opposite. Can’t you read? They have separate accounts. Didn’t I say that? So why do so many people ignore the running as Admin rule? It’s not a problem with GNU/Linux. So there is some difference in the way they are handled.

    The UAC? How many people leave that running? Do you? The UAC is just another way in which Microsoft inserts themselves between you, your hardware and your data. I’ve been accosted by the UAC. It’s another unexpected interruption Microsoft suffers the user with. The root account is nothing like that. It responds to your command, not the other way around so you always know when to expect it.

    Are there any other points that sailed over your head? I’ll help you understand them if so. You seem to have trouble understanding software freedom. Too bad.

  39. Michael Rudas says:

    “(B)eing infected by “Visiting a website” is not a bug in the OS either. Unless you are going to tell us all that the browser is part of the OS?”

    Actually, didn’t MS actually try to make that argument in open court? Part of the functionality of Windows, in fact, does rely on Internet Explorer code.
    So far as I can tell, even the “IE unbundled” Windows 7 simply moved much of the IE code to the OS itself; it looks like the Win 7 version of IE is now mostly a front end to the code Win 7 itself contains.

  40. kozmcrae says:

    Ted wrote:

    “The same way “root” can change any file? And as “root” is the first user created on a Linux system, therefore it’s the default if there no others are created. You had a point?”

    The root account is special, similar to the Windows Administrative account. But when I set up a non-Ubuntu based system I am asked to create a user account *after* assigning a password for the root account. You cannot go on without creating at least one user account. From the very first it is drilled into every GNU/Linux user *not* to use the root account for everyday use. To the best of my knowledge people adhere to that rule.

    On Ubuntu based systems the user is asked to supply their password for any action that will affect the entire system. That is another way of warning the user that what they are doing goes beyond just their user space.

    For whatever the reasons, Windows users are not encouraged to operate as just a user. I suspect because it’s to much of a pain in the ass not to. I’ve used both Windows and GNU/Linux and I have no problem with the separation of root and user accounts. I know before hand when and action will require the root password. With Windows the user is, unfortunately, conditioned to expect the unexpected. They are too eager to click next, click next, click next…

    The point, apparently went over your head. There may be similarities between the Windows Admin account and the GNU/Linux root account but GNU/Linux users don’t operate as root. That’s just another major difference.

  41. Linux Apostate says:

    Are you claiming that Windows has been designed to allow an unprivileged user to modify system files?

    When did you last use Windows?

  42. Linux Apostate wrote, “even if you are using the Administrator account, you still have to go via UAC if you want to modify a system file, because not even Administrator has permission to do that.”

    So, how did the malware in question routinely modify the system folder when all the user did was visit a site?

  43. Linux Apostate says:

    Adding to Ted’s comment, even if you are using the Administrator account, you still have to go via UAC if you want to modify a system file, because not even Administrator has permission to do that.

  44. Michael Rudas wrote, “Actually, didn’t MS actually try to make that argument in open court? Part of the functionality of Windows, in fact, does rely on Internet Explorer code.”

    Yes. See Court’s Findings of Fact in US DOJ v M$: “155. In contrast to other operating system vendors, Microsoft both refused to license its operating system without a browser and imposed restrictions — at first contractual and later technical — on OEMs’ and end users’ ability to remove its browser from its operating system. As its internal contemporaneous documents and licensing practices reveal, Microsoft decided to bind Internet Explorer to Windows in order to prevent Navigator from weakening the applications barrier to entry, rather than for any pro-competitive purpose.”

  45. Linux Apostate says:

    “And, you know the default Windows user can change any file on the system.”

    Do you really think so?

    When did you last use Windows, anyway?

  46. Ted says:

    “And, you know the default Windows user can change any file on the system.”

    The same way “root” can change any file? And as “root” is the first user created on a Linux system, therefore it’s the default if there no others are created. You had a point?

    As a matter of fact, the admin users in Windows cannot access other user profiles on the computer by default. There are also some system folders an admin user is locked out of, as they are only accessed by the SYSTEM account. The admin user has to be added to the ACL of these folders first. You might want to try researching how Windows actually works before mouthing off next time.

    Admin/Administrator is just the FIRST account to be created, same as root. It’s not a “default”. The first user on any newly installed system has to be an admin/root user. How else would other accounts get on the system?

    This “default user is admin on Windows, and that’s bad” thing is just a double-standard.

    The same way installing malware on Android is “social engineering”, but on Windows it’s the fault of the OS.

    The same way Linux marketing/advertising is just getting the message out there, but Microsoft marketing is always propaganda or astroturfing.

    There’s no such thing as a fair comparison between Windows and Linux from the FOSS crowd. No fact shall remain untwisted, no semantic loophole left unused. And apples must never be compared to apples.

    When it comes to features; Windows only has a browser, media player and simple utilities. Which is bloated.
    Linux has Office suites, mail servers, databases, web servers, IDEs and DVDs worth of applications. And it’s slimline.

    However, when it’s bug count, Windows suddenly also includes Office, Exchange, Visual Studio, two instances of SQL Server, IIS and Photoshop. Linux is just a kernel.

    Add a feature to Linux? Best feature ever, can’t live without it.
    Add a feature to Windows? Bloat! Copied from Apple! Anti-trust!

    Any feature not in Windows? Why isn’t this feature in?
    Any feature not in Linux? You didn’t need it anyway!

    Windows Aero? Resource hungry, useless eye-candy.
    Compiz? Absolute must-have! Spinning cubes!

    Benchmarks from established and trusted sources, with reproducible methodology that show Windows is better – shills, FUD, lies, MS paid for results, corrupt.
    Benchmark from random blog with questionable (if any) methodology showing Linux is better – holy gospel truth.

    Windows bluescreens several times a day for no reason.
    Kernel panics do not exist.

    Any Windows-only program = unoriginal, buggy, useless rubbish.
    Linux programs – Original, innovative, not in fact, clones of those Windows apps.

    Copying interface changes or features from Windows apps into the OSS clones is OK, even though they’re much maligned in the Windows version… If they’re so crap, why clone them?

    It’s OK to infringe copyrights of music and films, but don’t dare violate the GPL. It’s copyright, you know.

  47. kozmcrae says:

    Linux Apostate wrote:

    “Tell me that a user’s applications should not be able to modify files in the user’s home directory.”

    You know the Linux user can only administer/affect his own files. And, you know the default Windows user can change any file on the system. So, what are you trying to say?

  48. Linux Apostate says:

    Some parts of the registry are like $HOME, other parts are like /etc. Which is this? You seem to assume it’s like /etc. I say it is probably more like $HOME.

    Tell me that a user’s applications should not be able to modify files in the user’s home directory.

  49. Michael Rudas says:

    “(B)eing infected by “Visiting a website” is not a bug in the OS either. Unless you are going to tell us all that the browser is part of the OS?”

    Actually, didn’t MS actually try to make that argument in open court? Part of the functionality of Windows, in fact, does rely on Internet Explorer code.
    As far as I can tell, even the “IE unbundled” Windows 7 simply moved much of the IE code to the OS itself; it looks like the Win7 version of IE is now mostly a front end to the code Win7 itself contains.

  50. Linux Apostate wrote, “A second error is that you assume malware on Windows can immediately privilege-escalate and write to system folders. “

    The malware under discussion does write to the registry… It is picked up merely by visiting a website… That is the OS’ problem.

  51. Linux Apostate says:

    Well, being infected by “Visiting a website” is not a bug in the OS either. Unless you are going to tell us all that the browser is part of the OS?

    The basic inconsistency here is that if an Android user is tricked into installing malware, that’s social engineering. But if a Windows user is tricked into installing malware, that’s because Windows is insecure and it’s all Microsoft’s fault because they designed it to run programs opened by the user. Or something.

    A second error is that you assume malware on Windows can immediately privilege-escalate and write to system folders. Actually, the last time I saw any malware it was not able to do this, and just modified the Windows equivalent of $HOME/.profile. Perhaps you are imagining we are still in an XP world of “everyone runs as Administrator”? Privilege escalation bugs do exist, Linux has had plenty of them, but malware authors may not be aware of the ones that still work on up-to-date systems.

  52. Linux Apostate says:

    “Bad argument. The OS is meant to set the rules.”

    I know. But it’s Pogson’s argument.

    You see, if someone is tricked into installing malware on their Android, that’s social engineering. If someone is tricked into installing malware on their PC, that’s Microsoft’s fault because Windows is so insecure.

    Notice that he assumes that the malware gets System privileges straight away. Does it? The last time I saw any malware, which was more than a year ago, it did not escape user mode and instead did the Windows equivalent of adding itself to $HOME/.profile. Malware is so rare for me that I am always interested to see how it works and spent a while with tcpdump, processexplorer etc., watching this beast in operation.

    Notice also that he ignores all of the following: “Trojan.Milicenso may arrive on a compromised computer by various means, such as malicious email attachments… We have also encountered quite a large number of samples that appear to be packaged as a fake codec” and concentrates instead on “visiting websites hosting malicious scripts”. But this would be a web browser security flaw, i.e. an application bug, wouldn’t it? I thought this was about the OS?

  53. Linux Apostate wrote, “the malware you linked to spreads by social engineering: email, IM, pirate software downloads.”

    “Trojan.Milicenso may arrive on a compromised computer by various means, such as malicious email attachments or visiting websites hosting malicious scripts. The latter often unintentionally occurs when a user clicks a link in an unsolicited email. We have also encountered quite a large number of samples that appear to be packaged as a fake codec.

    The Trojan creates and executes a dropper executable, which in turn creates a DLL file in the %System% folder. “

    Visiting a website is not exactly social engineering. It’s HTTP. How the Hell does a user application write to the system folder???

  54. oiaohm says:

    Linux Apostate thing is the Linux world is responding to those privilege escalation bugs in a process of on going hardening.

    Windows has a lot more privilege escalation in drivers. Why one web cam chip can have 50 different drivers under windows because it has different usb ID codes where Linux has 1 driver for the lot of them. Clean driver supply stack reduces weak points.

    Its a simple problem of surface area there is more attack surface area in the windows kernel space.

    http://travisaltman.com/windows-privilege-escalation-via-weak-service-permissions/

    Its not like the kernel is the only way in either. Service busting under windows is in fact the most common attack used by malware to gain privilege under windows. UAC under windows made this worse.

    There is a difference to android malware. Most android malware does not make it to a higher privilege on the system.

    “By allowing the user’s processes to proceed? That’s what Windows is supposed to do.”
    Bad argument. The OS is meant to set the rules. Microsoft failed to. Not true on Android at all.

    Linux LSM systems allow applications to proceed but also know what the application is not approved todo so any stepping away from what it should be doing causes the system to drop on it like a tone of bricks.

    http://developer.android.com/guide/topics/security/permissions.html

    Android is very interesting. Each application is its own user isolated from every other application were able.

    Also end user data and application data is split. You cannot run a program from androids equal to the users my documents.

    So android is lot harder target than windows Linux Apostate. There is even options in android to lie to applications that they have been granted privilege when they have not.

    Attacker is facing black box in Android where every bit of information they can be getting from the OS can be a lie.

    The android difference is how successful the malware is. Android malware is far less successful at acquiring data than windows malware. Also Android malware has shorter life span in the field.

    http://lwn.net/Articles/494252/ With seccomp filters attack surface area will reduce even more.

    Linux Apostate openbsd was good but they need todo a lot of new work in attack surface area reduction.

    seccomp filters is just the one of the latest steps in the many levels of hardening the Linux kernel is doing. The amount of hardening on linux far exceeds Windows. Linux is starting to exceed all the BSD’s for the hardening. BSD’s still exceed for code auditing.

  55. Linux Apostate says:

    It’s a myth. I’ve seen too many privilege escalation bugs in the kernel to believe there are no more. Next time you get a new kernel package from security.debian.org, have a look at the changelog and see how many they fixed in the last months. Most likely, there are several of these holes on your machine right now.

    If you want a kernel that really was designed to be inherently secure, how about OpenBSD.

    By the way, the malware you linked to spreads by social engineering: email, IM, pirate software downloads. So, it is no different from Android malware, which you do not regard as a problem with the Android/Linux platform. How did Windows succumb to the malware? By allowing the user’s processes to proceed? That’s what Windows is supposed to do.

  56. Linux Apostate wrote, ” Linux is not inherently secure, for instance, there is plenty of malware on the Android/Linux platform,”.

    Yeah, sure. How did Linux succumb to the malware? By allowing the user’s processes to proceed? That’s what Linux is supposed to do. Android is more like an application in Android/Linux and is much more complex than an OS, so it can be flawed seriously. Linux, on its own is very secure.

  57. I doubt the malware is a direct threat to the forests. Interestingly, I did have a configuration error in CUPs (to retry failed jobs…) once that did the same. Fortunately, the person in charge pulled out the paper-tray… It was a very fast printer. 😉

  58. kozmcrae wrote, “Linux is inherently secure”.

    That’s true. UNIX, upon which GNU/Linux is modelled, was conceived as a general-purpose, multi-user, and multi-tasking OS. The idea that one user and his processes should be protected from others without permission was there from the beginning. M$ did not think of that until about 2000 for all its OS… after malware began to take over the world’s IT. Instead of doing things right, they added more layers making an already over-complex and porous OS more complex and porous. Instead of stopping malware, they only required the writers of malware to adapt by changing tactics from time to time. The target was huge per PC running that other OS and the monoculture that was Wintel made networked malware a growth industry.

    It’s interesting to consider the priorities of GNU/Linux and that other OS. Vulnerabilities are discovered from time to time but M$ tends to wait until it is actively exploited to fix things. GNU/Linux fixes things in hours. There are many reasons for that. Complexity is built into that other OS which makes it harder to fix anything without breaking something. It’s spaghetti-code. GNU/Linux is nicely modular and each fix tends to require few files to be changed. As well, the fix is propagated ASAP. No waiting for the world and its dog to test the fix. Just do it. Because the source code is available, source-code patches can be distributed to the world ASAP. Every distro can incorporate the patch and rebuild their particular kernels and it’s done. That other OS does a fair job of updating binaries but they leave the applications alone which is a huge part of malware/exploits today. That’s precisely because of the secrecy/closed-mindedness of M$ and its “partners”.

  59. kozmcrae says:

    Linux Apostate wrote:

    “It’s always essential to balance arguments such as the one above by pointing out that Linux is not inherently secure…”

    Actually Linux is inherently secure. I think what you meant to say is that Linux isn’t 100% secure. No operating system is, but Linux was built from the first line of code to be safely connected to the rest of the world. Its security is built in, inherent would be the correct word to describe it. Microsoft’s OS isn’t. It still has single user, network dumb code at its core.

    The malware you speak of infecting the Android/Linux (So now Android *is* Linux after all?) platform is, so far, of the social engineering type. It needs to be downloaded and executed by the user to infect. Compare that to getting infected by simply visiting a rogue Web site with Microsoft’s OS.

    So there really is no comparison. Microsoft’s record on security is that they have wrought a multi-billion dollar industry into existence just to service it. There is malware for Android like you said but so far it’s more scamware than malware. Google needs to do a better job of maintaining their app store. There’s no reason why it can’t be as secure as any distro repository. There has been only one known case of malware getting into the repository system so far in the history of Linux.

    And Android is only one branch of Linux. There’s desktop Linux, servers, Navy drones, super computers and stock markets. All of these uses need to be protected from malware. Freedom from malware is just one reason to migrate to GNU/Linux.

  60. Mats Hagglund says:

    So there is still some hope for Canadian, Swedish and Finnish pulp and paper industry? 🙂

  61. Linux Apostate says:

    I’m often impressed by the quality of modern malware. Evil, but clever. Stuxnet was particularly special in that regard.

    It’s always essential to balance arguments such as the one above by pointing out that Linux is not inherently secure, for instance, there is plenty of malware on the Android/Linux platform, and many GNU/Linux servers have been compromised. Security fixes are regularly released for Debian to close security holes in the kernel, libraries, daemons and applications.

    I was puzzled by “It uses files that should be data as executables”… such as?

Leave a Reply