Usefulness of security software

Putting privacy issues aside, which scenario would you consider "safer"? -MS Windows + paid third party security suite (AV, firewall, behavioral analysis, IDS, etc) + common sense -GNU/Linux + selinux + common sense common sense entails not clicking on random links, not opening random pdf/docx/exe and other suspicious files, only downloading software from reputable sites (including pirated software), browsing with adblockers and using a VM for any activity you consider risky Now obviously, we know GNU/Linux is generally more secure than Windows, both because of the way the system is engineered and relative obscurity for desktop users which leads to fewer malware developed for GNU/Linux. Unfortunately, this also means there is no interest for maintaining security software for home users. Unless you can shell out $500+ yearly for enterprise security, any existing solution such as comodo, drweb and clamav is a bare minimum signature checker which is useless 20 year old tech. Which forces the user to run GNU/Linux without any meaningful comprehensive security solution and rely solely on hardening techniques such as selinux and apparmor. On the Windows front, you would be number 1 favorite target for malware and 0days, but at the same time there are numerous security solutions from variety of "big player" vendors such as Norton, Eset, Kaspersky and McAfee (regardless of whether you like one particular or not). Now keeping in mind that practicing common sense will protect you from most low effort malware, we are interested in real threats that utilize 0day vulnerabilities and obfuscated/encrypted/handcrafted malware that would easily bypass signature based antivirus engines. The only way to detect and deter these threats would be utilizing heuristics, behavioral analysis, firewalls and intrusion detection systems. The ultimate question is: Are the current consumer facing paid security solutions for Windows, good enough to mask Windows' bigger attack surface and provide better security compared to a hardened GNU/Linux system? While answering the question, you might be inclined to say that GNU/Linux doesn't need antivirus at all, but 3 points to keep in mind: - Consider that most software that we use on a daily basis, such as browsers, text editors, torrent clients and media players are actually cross-platform and likely to suffer from same security vulnerabilities across multiple platforms with minimal customization. - Unlike the olden days, there are more incentives for targeting more tech literate users that would use GNU/Linux now. They are more likely to be holding a cryptocurrency wallet in their computer for example, or have access to internal company servers and data. - Android OS which runs Linux kernel is a popular target for malware, despite much more restrict hardening practices compared to average GNU/Linux system. The situation is bad enough that Samsung has a deal with McAfee and preinstalls their anti-malware on all their phones in addition to Google's Play protect service. Of course, the effectiveness of such anti-malware services are limited because the real threats would use 0days on high privilege processes that the anti-malware doesn't have access to. Finally, I'd like to mention that using alternative operating systems such as OpenBSD might bring about the most security, but due to limited software and hardware support for such OSes, it's not practical to use them for daily driving. They are best suited for servers and routers at this time.

>Now obviously, we know GNU/Linux is generally more secure than Windows stopped reading here. should really have stopped at the cop-speak 'obviously'

I also forgot to mention that this is about securing yourself against general and indiscriminate malware that are targeted against desktop users. Of course we know nothing is impenetrable and highly sophisticated malware such as stuxnet can't be protected against realistically without losing a whole lot of comfort and usability, which isn't generally worth the trouble unless you are trying to protect a critical facility or something. >>21212 Please, do provide evidence for the contrary. If Windows is truly safer than GNU/Linux by default, then the whole question is moot.

>>21211 I am of the opinion that GNU/Linux is "safer" by default, however when I say "default" I mean using a distribution with reasonable security practices. This means that 1) the kernel will be reasonably configured, 2) you will use a signature-checking package manager, and 3) there isn't a significant level of proprietary or bloated software installed by default; my gold standard for multi-purpose distributions is Void Linux. When you use a package manager which verifies signatures, especially if you use open-source software, you make it far harder for you to install anything which could be unsafe; it is ingrained within the Windows software installing system that you should go out and find executables on websites to install. It seems that windows has some sort of sandboxing application [1], however Linux has Apparmor/SELinux in addition to the simple chroot, and I hadn't even heard of Windows Sandbox before now despite having used Windows for years. While security through obscurity isn't true security, it certainly does help. In the end, privacy aside, I think that while Linux is "safer" than Windows by default, 90% of risks can be mitigated by, as you say, common sense, and the general public would be better off paying for basic "security good practices" training instead of antivirus. [1] https://learn.microsoft.com/en-us/windows/security/threat-protection/windows-sandbox/windows-sandbox-overview

> Now obviously, we know GNU/Linux is generally more secure than Windows, both because of the way the system is engineered and relative obscurity for desktop users which leads to fewer malware developed for GNU/Linux. even though there's fewer malware developed for linux, the rest of this sentence is pants on head retarded windows has been repeatedly hardened, layered, and has even moved to a model where the kernel is no longer the highest authority in the system. Session isolation, process isolation, and general rbac on windows is years ahead of desktop linux. Desktop linux is a bunch of tacked together stuff that frequently has massive and embarrassing oversights, due to its lack of a unified privilege model as well as the rigorous testing and engineering that microsoft does desktop linux has no such benefit. And while there are other issues with windows, especially regarding freedom and openness, OP clearly has no idea what the fuarrrk they're talking about. actually, a bunch of fundamental claims and assertions in OP are poorly written, based on poorly understanding a concept, out of date, or just plain wrong. > we are interested in real threats that utilize 0day vulnerabilities and obfuscated/encrypted/handcrafted malware that would easily bypass signature based antivirus engines the feds dont care that youre looking at furry inflation porn

>I am of the opinion that GNU/Linux is "safer" by default, however when I say "default" I mean using a distribution with reasonable security practices. This means that 1) the kernel will be reasonably configured, 2) you will use a signature-checking package manager, and 3) there isn't a significant level of proprietary or bloated software installed by default; my gold standard for multi-purpose distributions is Void Linux. strong disagree, supply chain attacks are still quite easy to pull off, especially on the desktop linux ecosystem i think the general purpose OS closest to the ideal solution is probably macos/ios, with strong sandboxing and consent based API access, followed by modern windows apps. I wouldn't consider desktop linux to even be on the list as a serious contender if it was configured to be a general purpose os. By the way, there's still a lot of work to be done. If you told me I needed to ship computers to 20 people who needed bumpers, I'd choose either chromebooks, ipads with keyboards, or macos/windows in app store only mode. Depending on the use case. however, if you're facing a nation state adversary, it depends on the nation state and your nationality. in some cases, i would guess that your best bet is desktop linux, although the most important controls would be discipline and human based. In others, a macbook with adp enabled would be fine.

>>21218 >>21219 Your posts read like someone who is butthurt because macOS wasn't mentioned in OP. >Desktop linux is a bunch of tacked together stuff that frequently has massive and embarrassing oversights Do name such oversights. Something like EternalBlue perhaps? oh wait, that was windows. So much for rigorous testing and engineering. right? Anyway, no need to get all defensive. The whole point of this thread is to discuss different platforms and the ability to secure them. Maybe GNU/Linux isn't as good as Windows in security, but the fact it's less targeted by malware is still a point to be considered. So, you are saying that if someone were to really develop a malware for Linux they'd have much more success infiltrating systems vs Windows? That's fine. Just try to back up your claims instead of reducing your arguments to insults. >the feds dont care that youre looking at furry inflation porn I don't care what the feds care about, I care about my chances of surviving a WannaCry-like worm releasing in the open that exploited a truly unfixed 0day vulnerability (unlike WannaCry). Now a ransomware is easy to deal with, because you will immediately know you are infected with malware and restore your backup on a fresh OS. The truly dangerous kind of malware are the rootkits that stay dormant in your computer for months or years, without your knowledge. >macos/ios I'm not interested in operating systems that lock me into a specific brand of hardware. Maybe some people are, but unless you want to argue that MacOS is some godlike impenetrable OS, I see little benefit to locking myself into a brand. >windows in app store only mode That'd make windows an unusable garbage that you can't even properly secure with a third-party solution, while still being vulnerable to gazillion 0days in various services such as SMB or whatever. The point is having security while still daily driving your computer like a normal person. Things like Chatting with friends, writing some programs, browsing the web, pirating some movies and looking at furry inflation porn.

>>21220 >Your posts read like someone who is butthurt because macOS wasn't mentioned in OP. which is a serious omission, as the work apple is doing is not insignificant macos is trash and i refuse to use it as a serious os. i personally have a macbook air that i literally use as a firefox and text editing machine (the battery life is 15+ hours). >Do name such oversights. Something like EternalBlue perhaps? oh wait, that was windows. So much for rigorous testing and engineering. right? acting like strongly architected, defended, and tested platforms cant have serious vulnerabilities is so stupid. keep up as for oversights? just take a look at the dbus model - despite supposedly having the concept of individual sessions, the actual code is buggy and the model is poorly segmented. systemd has some serious issues as well actually, if you want to lol, go look at the number of high score CVE privilege escalations linux has every year. look at dirty pipe and dirty cow. desktop linux is a fuarrrking mess also, the base user/process privilege model that all current major OSes use is fundamentally broken, but that's a whole separate conversation. i really hope google gets traction fuschia/zircon and puts it under a more permissive license >I don't care what the feds care about, I care about my chances of surviving a WannaCry-like worm releasing in the open that exploited a truly unfixed 0day vulnerability (unlike that's unlikely to happen on windows again, microsoft actually learned their lesson. desktop linux doesnt have the penetration for something like WC to happen again. that era of exploiting default configs is probably over. >The truly dangerous kind of malware are the rootkits that stay dormant in your computer for months or years, without your knowledge. windows actually now has controls against this (they call it "core isolation"). linux does not by default, although a handful of weird nerd type distros and server solutions implement something similar. macos has controls against this, and their chain of trust model is really solid. in fact persisting code through boot on macos/ios is now considered to be a career achievement for researchers, apple has done a remarkably good job. >I'm not interested in operating systems that lock me into a specific brand of hardware. Maybe some people are, but unless you want to argue that MacOS is some godlike impenetrable OS, I see little benefit to locking myself into a brand. if you restrict it to app store only, it is the closest to godlike impenetrable. AS macbooks are the best laptops too. it absolutely is like kicking dead whales down the beach ass that they are so completely proprietary, and i would consider that aspect to be a massive risk >That'd make windows an unusable garbage that you can't even properly secure with a third-party solution, while still being vulnerable to gazillion 0days in various services such as SMB or whatever. see above, but "0days in various services such as SMB or whatever." hasnt been an issue for a long time, especially on windows 11 - one of the greatest leaps forward for enterprise endpoint security. too bad the ui is like kicking dead whales down the beach ass lmao >The point is having security while still daily driving your computer like a normal person. Things like Chatting with friends, writing some programs, browsing the web, pirating some movies and looking at furry inflation porn. you can do all of this with windows 11 "s mode". you just need to sign up for msdn and get your signing certs from them for the running code part or use an azure hosted environment (let's all have a LOL at microsoft's recommended solution)

>>21211 >both because of the way the system is engineered and relative obscurity for desktop users which leads to fewer malware developed for GNU/Linux. Linux is open source, GNU userland is opensource, lots of manuals around, is not obscure at all. >>21211 >Unless you can shell out $500+ yearly for enterprise security Security is not a product OP, never will be. Read NIST 800 docs and ISO 27k. Approaching it as a product is already a failure. Seems like you are converging to the fallacy of security through obscurity. Let me tell you that this never works. Unless you have full transparency, (with proof in cases of cryptography), there is no security. Seeing it as a simple secure/insecure state is also another bad argument OP. Security is implemented in layers. Focus on ownership over the whole software supply chain, good architecture and transparency. Open source gives you the first 2, but you can still make a mess in case you don't know how to put the pieces together.

>>21211 safer from what? what is your threat model? "more secure" is stupid and a fundamental misunderstanding of how security works

The more I've been thinking about it, the more I've started to feel like that loonix "security" is just a fuarrrking meme. Privacy is another thing, but modern Windows isn't nearly as bad as it used to be. That being said, mobile OS's like iOS and Android are generally much more secure in my opinion, just because they isolate applications by default. The traditional desktop model where every application you run instantly gets access to all your files is quite terrible when you really think about it. Modern Windows is pretty bad in some other ways but I don't think that security is one of them. I've honestly been wondering if it would be better to just run Windows on computers I use for work and other important stuff. I know it's arguably much worse for privacy, but I don't really want to cause trouble because my computer got hacked.

>>21231 Secure from non-state-backed malware Read the thread >>21238 That is what I was thinking as well, which is what that led me to create this thread. I still believe any self respecting malware programmer would at least make sure that he can bypass windows defender before they bother releasing their malware, so I firmly believe that a third party security suite is a must for windows. There is also the fact that if a Linux desktop is infected, the user is much less likely to notice because linux users generally aren't worried about infection. They (me included) only follow the general safety guidelines such as the ones mentioned in the OP and assume the system is secure.

>>21240 >There is also the fact that if a Linux desktop is infected, the user is much less likely to notice because linux users generally aren't worried about infection. They (me included) only follow the general safety guidelines such as the ones mentioned in the OP and assume the system is secure. This is honestly my biggest worry as well. Do I really have any way of knowing if my system is infected or not? This is an issue on Windows as well. However, there is some change that windows defender or some other antivirus you're running will catch it. These could be concerns especially if you're working for example as a sysadmin in a big organization. In that kind of a position, you could be subject to targeted attacks. Hard to decide whether to run linux or windows if these things are a concern. Linux is much better in some other ways, but imo modern windows honestly just has better security features out of the box.

Have to add that it's funny how there seems to be a billion posts online that just parrot how >Linux is much more secure than Windows! like some NPC. But almost no one of them have really actually thought it through. It's almost always some idiotic surface level soykaf like >Whenever you want to do a system-level operation on Linux you need to provide the root password. When you install Windows, it will create an administrator account and give you full access to do basically whatever you want with the operating system using the magical “Run as administrator” option In general, I've grown a distaste of the "linux community", because it seems to be full of midwits who just parrot the same false truisms about it being so much better than anything else. I use it on the desktop and on servers, but for a lot of people it just seems to be a fashion statement.

>Are the current consumer facing paid security solutions for Windows, good enough to mask Windows' bigger attack surface and provide better security compared to a hardened GNU/Linux system? No chance. Just use Qubes, obviously. It is perfectly suited to being a daily driver and no Windows setup has chance of competing.

>>21251 Noting the existence of Qubes essentially ends this thread and makes all the earlier comments seem dumb

Is Qubes that good? I would like to hear more opinions. From what I understand you need a dedicated high end machine to run it on, I guess kind of like GrapheneOS for phones is only supported on new expensive Google Pixel phones. I'm personally thinking of going the route of using Windows for every day stuff and having a bootable encrypted persistent Debian USB for cryptocurrency stuff.

>>21264 Yes. I used it as my daily driver for months with no major issues aside from my hardware being too underpowered. I highly recommend it to anyone who really wants to stay anonymous. Its not great as a programming environment, though.

>>21264 It's only okay if you really want security above everything else. A lot of functionality is sacrificed to run everything inside a vm, most important being the lack of hardware accelerated graphics. I honestly can't see it working as a daily driver for now. Maybe after a decade of development things will be smoother for it. Major hurdle at gpu acceleration is that the greedy suits in graphic companies won't let the engineers put vgpu functionality on consumer grade graphics. It has a few other minor issues as well, but I didn't use it long enough to have an opinion on them.

>>21264 > From what I understand you need a dedicated high end machine to run it on, The opposite: the community and developers recommend a few particular thinkpads; newer hardware is more likely to be untested and have unexpected bugs. I've been using it as a daily driver for 6 months and there's no way I can go back to using anything else. I just switch to a separate PC for gaming or high resolution films. It feels so insanely risky like im naked to do things like web browsing and chatting in typical insecure OSes. >>21265 >Its not great as a programming environment, though. Why? I totally disagree, it is so easy to have separate development environments that i hate to use anything else.

>>21274 It's difficult to install system libraries so you can build and test your code easily.

>>21211 >Now obviously, we know GNU/Linux is generally more secure than Windows, both because of the way the system is engineered I find Linux privesc significantly easier than Windows. Especially under a conventional home user Linux install, your current user probably has full sudo perms, and unlike with Windows UAC bypasses, you don't have AMSI and defender out of the box there to stop you from dumping the user's creds with LaZagne and then pulling some GTFOBins soykaf to get root. End user fails too like shared passwords or similar could be another thing to fuarrrk around with. There's just a lot more on a linux end user system that you could go for and likely don't have to duck and dodge around AV the whole time. >relative obscurity for desktop users which leads to fewer malware developed for GNU/Linux Your average linux user likely has wine installed, which permits a good deal of windows malware to run. Pic related for PoC. I know it's just a basic bitch meterpreter shell, but I now have a windows shell and can write anything the wine user can write, meaning I could probably get a native ELF shell aboard if I really thought about it... and because I have no real time protection out of the box on linux, I don't really even need to make it evasive or do anything complicated to get my initial access shell to pop. I wouldn't have even been able to create this file on stock windows due to defender, much less execute it. >Unless you can shell out $500+ yearly for enterprise security I think sophos has a free tier endpoint for linux. Can't remember; don't use it. Most linux users won't either because it's nonfree software. >On the Windows front, you would be number 1 favorite target for malware and 0days Most new CVEs in my experience are for IoT, webapp frameworks and network infrastructure. This soykaf is critical infrastructure for enterprise or else highly exposed, and so it gets tested the most and thus has the most vulns found for it. >The only way to detect and deter these threats would be utilizing heuristics, behavioral analysis, firewalls and intrusion detection systems. IDS is still pretty sig-based, and network artifacts < host artifacts. It's good defense in depth soykaf to have an IDS and all that, but really you need good soykaf on the endpoint to get optimal protection. And yes, that means behavioural analysis, etc. Most modern AV with real time protection and cloud delivered services can do this, though. Really new techniques can fudge behavioural analysis too, so nothing is 100%.

>>21301 >There's just a lot more on a linux end user system that you could go for and likely don't have to duck and dodge around AV the whole time. true, the consensus in this thread seems to be that Windows is in general more secure on the account of having more tools to deal with threats. I'm almost convinced to switch my main PC to windows but Windows is just so soykaf in a lot of areas. lack of proper filesystem support, fewer personalization options, requiring frequent reboots for trivial things, powershell not even remotely close enough to replace a proper terminal, not supporting some special characters in filenames, using backslash for file path and the list goes on. Such a shame. >I think sophos has a free tier endpoint for linux. It doesn't, as far as I can tell. Perhaps I didn't look hard enough but I couldn't find anything on their website. I would definitely feel a bit more safe with a proper endpoint security, even if proprietary. Overall, thanks for your insight

>>21218 >the feds dont care that youre looking at furry inflation porn for real? can I stop worrying?

>>21325 Just because you trust the current government for some reason doesn't mean you will be ok with all future administrations. Look at how hard they are trying to trans the kids, we could easily end up with a violent reactionary swing to the opposite side of the pendulum. A populist fundamentalist religious totalitarian faction could take power and it would take 5 seconds for the NSA to run a script across all your browsing history and come up with a "degeneracy score" the government and their crony corporations will then use to oppress you. Something something opsec before you need it.

>>21223 >>21218 Damn, quality posts. I would disagree with the assertion that a wormable bug for Windows is unlikely to happen again, but it is significantly harder, more valuable, and likely will be patched much faster. Maybe semantics. I entirely agree with what you've mentioned about hardening. There have been tons of strides made in the past five years alone much less the past decade. VBS+HVCI (Credguard, etc) is a fantastic solution, esp after the new update a few months ago which properly enabled and enforced the bad driver blacklist. Sadly there are an ungodly amount of soykaf drivers out there, cranked out by "developers" who don't think to secure a device object or minifilter comm port.

why are there a bunch of CIAniggers in the thread claiming that Windows and MacOS are secure operating systems?

>>21492 Do you have anything to refute any of the points made itt? I'm guessing no...

>>21494 Just your typical imageboard neckbeard schizo type soykaf I suppose. The "linux is more secure" idea mostly stems from its obscurity/lack of widespread usage to justify developing malware for it to use on a grand scale which is a pretty poor argument. MacOS and ChromeOS are actually quite secure, in the same way a prison is. Windows itself as far as I understand isn't all that great especially when you consider how you have to install software from random executables you get from the internet, but Microsoft does invest quite a bit into Defender to try to improve and maintain it because they know that their OS is the biggest target.

This thread convinced me to switch to Windows for security. Here's hoping I didn't make a mistake. I did enable all security features but it still doesn't "feel safe" somehow

>>21492 They have been organising for years, especially surrounding Daniel Micay like fanboys >>21275 what, how so?

>>21494 lmao as if windows supporters made any single point at all. they just spill soykaf about windows being more secure and don't provide you with any single evidence. just empty talk. i think security of windows is just as bad as of linux, because considering that micro$oft is one of the worst coprorations ever, i have little doubt that windows is the biggest roorkit itself. does it really matter who comprises your system? with linux, maybe you are more likely to be hacked by a chinese nerd, but the biggest security threat is cia anyway. and with linux cia at least have to spend resources to get in your computer. with windows micro$oft will get them in your system if paid enough, which is much faster and more reliable. and don't forget that linux has a huge market share in IT industry (web technologies, servers, clouds etc) so amount of security research on linux is NOT insignificant and cianiggers conveniently decided to ignore this fact itt.

>>21367 >I would disagree with the assertion that a wormable bug for Windows is unlikely to happen again, but it is significantly harder, more valuable, and likely will be patched much faster. Bit difficult to identify part of the operating system behaving like a worm, made by the developers/India.

>>21972 > linux has a huge market share in IT That's a double edged sword though. It means you are more likely to win cool soykaf if you manage to compromise a linux system compared to windows. Linux users are more likely to hold cryptos on their computer, you are more likely to find source codes and private internal software

>>21223 >AS macbooks are the best laptops too Certainly not from the standpoint of affordability. Maybe if Apple lowered their prices and went in the right direction as far as right-to-repair... Some of the things Apple has done are totally insane hardware wise, like soldering the hard drive onto the motherboard. The ports are... lackluster... to say the least, but I guess you can get an adapter. I also HATE their keyboards. I did read on the new macbook airs that they have replaceable SSDs, which is a step in the right direction. I think it depends on what you are using it for. Honestly, where you are just using it for web browsing and text editing, I think it seems way overpriced. You could do all the same things with a duo core laptop from the 2000s.

>>21972 Most people don't have nation-state level actors or intelligence agencies as their threat model. And I think he pretty much acknowledged that depending on where you live, Desktop Linux is better in such a scenario. Also, you are concerned about information gathering, not necessarily security. The CIA's primary purpose is intelligence gathering. It is not trying to recover actionable and prosecutable material. If you were really trying to avoid that level of threat actor, you would not have a static security structure. You would not operate from a home network. You would be running things on live environments off of RAM, with many many fallback devices. To protect against a state-level actor, you'd need another state-level actor's backing financially in order to maintain a static security structure. Do you even realize how expensive and labour intensive that level of security infrastructure actually is? It's not really tenable unless you are doing something of importance on behalf of another actor.

>>21211 What makes most Linux distros more secure than Windows/Mac has mostly to do with them having actually useful package managers. Getting malware in via Debian repositories is a lot harder than getting some schmuck to download a program from some low-barrier-of-entry "App Store", let alone some website. Besides, all this soykaf about "kernel hardening", "root access" or whatever is a lot less important than the data that sits in your homefolder. I don't know of any mainstream desktop OS that takes serious steps to restrict access to the homefolder. Phones OSes with their fancy "capability"-based security are frankly also useless since users are stupid enough to give apps all the permissions they ask for without giving it any thought. You can't fix stupid. In which case I strongly prefer an OS that gives me actual control (while being practical), i.e. some Linux distro. t. uses Devuan.

>>23404 I agree mostly. It, of course, depends on a lot of soykaf we both could mention but that would be too needlessly verbose. Depending on the network and configuration a Windows host may offer more opportunities for internal pivoting, but I've seen just as many opportunities on linux systems. >>21972 >ignores kernel hardening posts >no evidence just empty talk yeah sounds about right for a double spacing reddit normal person

>>21218 >windows has been repeatedly hardened, layered, and has even moved to a model where the kernel is no longer the highest authority in the system. Session isolation, process isolation, and general rbac on windows is years ahead of desktop linux. The problem is who Windows is providing with security? Is it the users'? Windows security is used to protect their corporate interests no the users. My Windows using friends tell me that their OS displays advertisements on their screen. It's probably phoning home, sending information about every URL they visit, every thing they type including every password, back to M$. This might be used for marketing purposes, send to law enforcement, and anyone who pays them enough money. Instead on my Linux installation, it's MY security that is being protected, so even if Windows were to have more advanced security technology theoretically.

>>23430 >It's probably That's certain. Last years news "Edge browser feature sends images you view back to Microsoft"

Windows may provide some heightened security features, but they go unnoticed. The OS is still bloated, ridden with vulnerabilities due to its bloatedness, it still telemeters all your data, and now it even forces advertisements upon you. Any possible security it offers is only for a corporate interest, it is to keep your data within Microsoft's sphere of control. The same goes for any of these major Big Tech companies instituting "security measures", it is a false security. A good OS should be designed in a secure manner, it should be written efficiently, it should not take up a large amount of space, it should run smoothly, it should not spy upon its user either. To ensure it is not spying upon its user it must have its source code publicly available so it can be audited. Windows might cover the first requirement, maybe it really is designed securely I wouldn't know, and I don't need to know truthfully because Windows is terribly written, it is beyond bloated, it is riddled with vulnerabilities, and it has far more attention upon it by maldevs and hackers which allows those flaws to overcome its supposedly secure design. And it spies on you, so it goes into the garbage. Linux is really not designed with security in mind, it has a lot of shortcomings but because it's not a large target for malware we don't see these flaws brought out in real life all that often. I will say that linux still provides a lot of tools to enhance your security, and in general it is just easier to use tools and develop behaviors intended for greater security, in my experience doing anything even moderately technical on Windows is a pain. Additionally, the OS does not spy on you. Anyway I'm not gonna bother reading over this entire thread again, it seems people are really just arguing because they can't just agree that linux does have flaws, and that its flaws should be addressed. Linux's flaws should be addressed, it is arguably less secure than Windows or Mac/iOs/Android in some inconsequential ways, and it could be improved upon, but instead we just use these palliative tools like SELinux or Apparmor, Bubblewrap, whatever people use, to compensate for its flaws. Instead the root of the issue should be addressed, but I guess it won't be. The real problem should be that there doesn't exist a truly great OS in terms of the secure environment it offers to its users.

>>23440 regarding tiny OS: https://kolibrios.org/en/download.htm fits on a floppy disk (2.44 mb) and still has windows like gui with file manager etc. Sets an idea how much you can do with very tiny footprint

Linux has a higher number of known exploits while Windows it takes longer for exploits to be patched due to less eyes on it. So Windows is possibly more secure against criminals while Linux would be safer from government agents assuming you are using a distro frequently updated but not so frequent that shady updates get pushed. Though on other hand Windows is usually targeted more when it comes to malware so maybe that wouldn't be the case. So it depends who you want to be safe from. Though MacOS would be preferable over Windows if you're talking proprietary since they have historically not had as many bad security holes as Windows and better privacy features which in turn is better security before thought I'm not sure if that is still the case when Hackintosh installing. Though using something like Haiku you can probably get by on just security by obscurity for a little while though this is a gamble the more time passes. But really the best practice is probably setting everything up assuming you will have active malware. For example you want keystroke encryption to prevent keyloggers from recording anything useful.

>>23440 My god, a nuanced and coherent point? In this thread?? Fr tho this is a pretty fair and reasonable assessment.

>>23440 Considering Linux is also used on servers I don't see why it should be considered not a target.

>>23440 >The real problem should be that there doesn't exist a truly great OS in terms of the secure environment it offers to its users. what is qubes

>>23458 Thanks for being so predictable lain. In case some dummy replies to me talking about VMs send this: I'm glad you brought that up, I forgot to mention VMs. Here is a reply I prepared in case someone mentioned that (I didn't wanna spam the thread as I thought my argument still stood alone): One more thing I forgot, the virtual machine usage. That is the best that people have by the way of palliative tools, just containing potential threats within virtual machines with exploitable hypervisors that are still based upon flawed OSes. Arguably one of the most secure existing OSes, one being mentioned in this thread, and one that is certainly heralded as 'the most secure operating system' is Qubes, which is really less of an OS itself than a palliative interface that attempts interact with flawed OSes in the most secure and reasonable manner possible. Are VMs secure? Sure, but not infallible, in fact they are far from infallible and I've grown to believe that people overestimate their security. Additionally the OSes they are based upon are absolutely fallible and can provide an attacker opportunity to escape a hypervisor. As I said before, the root of the issue needs to be addressed instead of people pushing palliations.

qubes more like rubes amirite the body was too short

>>23459 In the interest of pivoting the thread away from omegamind-ing innocent lainons into the shadow realm and towards more practical pontificating, I will ask the Exceedingly Predictable Question: how would you architect an OS with better security properties than what is currently available? It's an interesting idea, to bake a hopefully simpler solution into the OS rather than relying on an add-on, but it reminds me of Linus Torvalds complaining that security nerds never think about performance when they submit patches.

>>23462 And security nerds complain to no end that dev (and pure networking guys) never gives a thought to security -- all about passing packets or making the thing happen, never a thought for what works that really shouldn't be able to. That and the insult of being seen as an inconvenience, killjoy, destroyer of efficiency, etc. by the people whose assets and business continuity they were hired to protect (but only after they were forced to by regulation or our good old American way of putting the fear of God into people in the 21st century: the threat of litigation and liability). Whatever, anyway, as rather pointedly suggested by your image but for some reason not addressed in the post, the problem is indeed that the OS that came to be used by the vast majority of home PC users for many years started by normalizing just doing everything as root (it was after all not really any kind of a shared environment no matter how many local user accounts you made), and then added a networking stack. What followed was an astonishingly hard fight, now on the order of actual decades, to figure out how to get people to elevate privs as needed instead of all the time for no reason, without unduly annoying their user base, many of whom already took great relish in being endlessly annoyed by the product. Things were a real mess around 2008-2010 when I had my first "real" job doing consumer PC tech support. At the time we had whitelabel contracts with every major office supply store chain in America plus the same services directly to the public, plus... whatever the hell TigerDirect even was (literally nobody ever called in to that queue). It was an endless stream of the same type of soykaf: someone randomly found a "rogue" (bullshit) antivirus installed on their PC, invariably by way of a "driveby download" exploit requiring no particular lapse of judgment by the user except the habit of running everything (and therefore everyone's soykafty web code) as an account with no restrictions. The rogue AV would proceed to install one or two of a large rotation of obnoxious rootkits with which to protect itself by creating hooks to intercept various OS functions that would otherwise be easily used to disable it again. Thus shielded against amateur attempts to run SuperAntiSpyware or Malwarebytes, it would flail around, slowing everything down and popping up annoying messages pretending to have found "viruses" (not always spelled correctly) and soliciting money to "remove" them. We would then -- over remote desktop sessions -- when ATT Uverse FTTNode VDSL was pretty new, and FTTPrem wasn't even a thing yet unless you were lucky enough to live in a part of Verizon territory where they were rolling out FiOS -- very often spend hours fighting (slowly) to circumvent the rootkits using obscure tools like gmer that none of us fully understood, and trying random hail-mary procedures that frequently involved a lot of waiting for a PC to reboot into safe mode, walking old fuarrrks through steps over the phone, or repeating appallingly long malware scans in hopes that some random thing removed by the last one would turn out to have cleared the way to detect and remove another. Despite the very cowboyish operation, or maybe because of it, it had a great ingenuity and tenacity to it all. We rarely had to give up and tell someone to reformat and reinstall, and in fact many of us took great pleasure in making each ticket as unprofitable as we could manage while still doing our jobs, especially after a rather insulting raise of twenty-five cents an ho-- Well, anyway, everybody quite predictably learned nothing from any of this, and continued fighting all but to the death against any infringement on their divinely bestowed, unalienable right to run everything as admin and to pursue life, liberty, and the feeling of being a Power User who does not need training wheels or water wings. This flaw is inherent only to that lineage of OS and has been the essential cause of the vast majority of consumer-sector security concerns.

>>23462 I feel bad and regret adding that snippet to the top of my post as it was just gratuitous but I did get annoyed with the poster. Anyway it just jeopardized the positive direction of the thread. Truthfully I don't feel I immediately have a great deal of ideas, and hopefully other people can offer some more. I mean I don't have the power to influence anything anyway. I think the other poster already addressed everything I was gonna say. The all or nothing root privilege is a major issue which is just not tenable. If there was instead a framework to make things so that one user could not just have complete access to a system, it'd make a big difference in altering the security of linux. Regulating what data applications have access to would be good, but instead they are given complete access to all user data pertaining to an invoked user account. All of the aforementioned palliations are dedicated in part to compensating for these problems, when instead the roots could be addressed, then these utilities could exist simply to bolster our security rather than compensate for flawed design. There are still methods of attempting to ameliorate these two issues, particularly found among enterprises, but they're always convoluted and vulnerable to exploitation. I mean the thing is that this 'flawed design' as I call it is really due to the fact that linux is based upon an old foundation, during the time the design didn't pose much of an issue, and to be fair if one is extremely exacting with how they conduct their activities then troubles can be avoided for the most part. Like I use all of the 'tools' I mentioned already, I'm using Qubes right now, I have very strict rules for myself on how I execute things. Maybe I'm completely misguided and being unreasonable, since SELinux will serve to deal with the application permissions problem on some level. Other OSes do regulate these things by default, like Android, which could leave them with a superior design if you weren't to consider all of the other flaws of those OSes and the devices running them. One problem that fills me with dread is the bloat issue. A lot of bloat comes from the provision of compatibility. Once-minimal distros will want to provide compatibility for video games and the applications used for running them, and will want to maintain support for soykafty web-browsers that are configured to support HTML5 garbage, non-soykafty, minimalist web-browsers suffer from great pressure from users desiring the ability to use big tech websites like youtube, that inability hailing from the lack of support for html5 garbage for instance, and so the developers will offer compatibility and watch as the browser turns into a bloated mess, a shell of its former self. It happens with everything. The online infrastructure that many people depend upon which is handed down by big tech is a bloated, malicious mess. If one does not have the ability to interface with it, then one may be unable to live their life, if they are required to say conduct bank activities online through a web browser, otherwise they'll need to use a client to do it (but the client won't support linux/bsd and will be a bloated mess if it does, and to even support the bloated mess requires linux/bsd to bloat itself). It is a conundrum people are faced with. I don't think most people consider the bloat to be an issue, and it is somewhat manageable on one's own. But people do tend to bring up bloat when considering security, but never take the time to question how their behavior contributes to the existence and perceived necessity of bloat. I don't think people's behavior in this aspect will change, so the issue will remain.

>>23459 Nonsense response. if you're going to outright reject the concept of defense-in-depth then make an actual argument.

>>23462 Interesting points have been brought up here, but I think Daniel Micay (someone who I’d call a leading expert in this particular field if nothing else) provided a better brief: https://www.reddit.com/r/GrapheneOS/comments/bddq5u/os_security_ios_vs_grapheneos_vs_stock_android/ Anyway, my answer to your question would be some combination of spectrum (https://spectrum-os.org/) and redox (https://www.redox-os.org/). Oh, and grapheneOS, if that wasn't clear already. Maybe take a look at the goals of these projects, and it’ll give you an idea of current weaknesses in mobile computing security. The obvious problem is that these projects are immature, but that'll always be the case when you start something new from scratch. The point is that by doing so, you can architect with security by design, like with what Wayland, rust, redox, spectrum, graphene, chromeOS, Titan M, Fuchsia, macOS, Windows Core OS, and so on attempt to do. This post about replacing qubes with a modern stack gives a user’s perspective: https://roscidus.com/blog/blog/2021/03/07/qubes-lite-with-kvm-and-wayland/ And it’s not always just building new sandcastles. In the interim, we get things like flatpak, selinux, qubes, whonix, Intel CET, arm MTE, and so on.

>corporations are not malicious actors! lol

>>23467 Remember that time corporations making anti-cheat DRM for game companies ended up exploiting the kernel side malware for mining bitcoin?

>>23466 I'd always heard that microkernels had too many performance issues to work, so it's cool seeing something like redox. Plus, having an OS written in rust is itself an exciting thing. I suppose there's really two ideas to investigate here, which is that of kernel security and the userland security model. Or "applications security model"? Whatever. Kernel protects system from getting rooted, userland should stop random programs from dumping passwords from firefox. >>23465 I came to the same critique, although I think what that poster was trying to get at was that there should be a stronger security model by default in Linux, rather than putting it into extensions. Pic related in >>23462, also reference >>21301; both consider what a typical install looks like. It would simplify the system to have One Security Model to Rule Them All, but maybe capabilities, cgroups, etc. came about as compromise; I don't know that history.

>>23473 That's historically been the zeitgeist, but microkernels have advanced. If you're really interested, I find this course to be educational: https://www.youtube.com/watch?v=IvqM2pmApSY >Kernel protects system from getting rooted, userland should stop random programs from dumping passwords from firefox. Not sure if I’d describe it that way, but the key point is that general security principles such that of least privilege, compartmentalization, etc should hold no matter what you're doing. There is controversy when these principles are taken to extremes, like with all the babble around rootless systems. I think it was Micay who said something about rootless being a non-issue if you have good APIs, but that argument is not one I want to get into. Personally I like the macOS approach with SIP. In my experience it works quite well, and you can disable it if you really want to. However, since you can’t do the same with iPhones, I can see why people are scared that rootless means they lose control of the device. Even more generally, I think good security and user control aren’t mutually exclusive. Alas, the groups that have the resources to make their products secure tend to be profit-driven and therefore control freaks, with few exceptions. This is why grapheneOS currently only supports Pixel devices. Unfortunately, this has resulted in a lot of FUD about security principles being antithetical to user freedom, which just isn’t the case if implemented correctly.

>>23474 >Kernel protects system And thats why you have bios malware that starts before the kernel now. Interestingly game cheats became sophisticated enough to use it for their own purposes too. Kinda funny how far cheats escalated.

>>23486 That sort of thing is mitigated with proper hardware attestation, like Intel Hardware Shield or what GrapheneOS does. Cheaters obviously avoid that soykaf at all costs, same reason you wouldn't want a shadow stack if you're using ROP. Most of them will be fuarrrked when anticheats start enforcing TPM, VBS, IOMMU, etc usage :D.

>>21213 >Please, do provide evidence for the contrary. He wasn't contesting that fact, he was criticizing the way you worded that like "Don't worry guys, OBVIOUSLY we know GNU/Linux is impenetrable, so feel free to let your guard down =P"

>>23502 GNU/Linux is impenetrable, xz-utils backdoor targeted GNU through non-GNU component systemd, it's also likely that it could've worked through PAM and SELinux the exact same way, because all of these three are not present on my machine, I was in fact not affected even if I installed backdoored liblzma. Moral of the story is don't use malware that is systemd, and don't fall for soycurity meme by installing 50 different things that are actually opposite of secure.