Software and computer security has always been a rather important topic. As our systems become more interdependent and connected- and we expose ourselves and our important information more and more, it is becoming even more important. Operating System and Software updates are issued to address security problems, and these security problems are given the utmost importance as users are urged to install them as soon as possible. Many Operating Systems- such as Windows 10, disable or restrict the ability to prevent updates (it seems to require Pro to adjust the settings to update only when the user initiates it, for example). This is considered by many to be a positive change; the idea being that this will prevent systems from being compromised through those security exploits.
And, certainly, that is true. Installing security patches will, obviously, prevent the exploits that they resolve from being exploited for malicious purposes. However I think the impact that those exploits have in terms of your typical end user have been overstated.
Based on my own experiences, I an animate that the vast majority of end-user malware infections are not perpetuated or contributed to in any notable way by the sort of issues resolved by security updates. Those updates are more applicable to servers, data centers, and corporate environments. For end-user PCs, it is seldom heard of to find a malware infection that was not caused in some way by trojan horse malware — something which the user explicitly downloaded and ran themselves which had the unintended side effect of releasing malware onto their system. Pirated software; game mods, “keygens”, and so on and so forth. Screensavers, greeting card executables, applications that disguise as images. Something as seemingly innocuous as an aftermarket Windows Theme could very easily contain an unwanted payload, and it won’t matter if the system is fully up to date or not if you allow it to install.
The Security Circus
I call the general concept of overstating those concerns the “security circus”. It infects certain styles of thinking and sort of makes itself a self-perpetuating concept over time. As an example scenario, a user may come to an IT repairperson with issues on their PC; it may be determined that those issues are caused by malware. the “security circus” contribution to this scenario could be that the repair person discovers that the system is out of date and missing a number of critical security updates. Because they have learned, over time, that security updates are critical and prevent infections, they may — and very often do — assume that the malware made it’s way onto the PC via that infection vector. Over time these occurrences pile up and that particular IT staff can state, without any intention of lying, that they have seen plenty of systems that were compromised by vulnerabilities, even though, realistically, they don’t actually know if the vulnerabilities were even responsible.
The “Acts” of this security circus seem to largely stem around Fear, Uncertainty, and Doubt being spread and manipulated. Coincidentally, I notice that oftentimes these efforts work in favour of a number of corporate interests. Forced OS Updates for example benefit the OS manufacturer, particularly as updates may very well provide any number of other pieces of software which provide “diagnostic” information which can be used by that company for marketing efforts. Security updates benefit security firms and security software vendors, who’s products are used to “prevent” the problems that are caused until they receive that critical patch to fix the issue, or who release security “scanners” which analyze and report whether a system is susceptible to the vulnerability.
Some recent security scares come to mind when I think about the “security circus”.
The Wannacry ransomware provides some good examples of the operation of this “security circus”. Articles and postings on the issue are often decidedly vague about the extent of the vulnerability that causes it, and often overstate it’s capability; users are urged to update as soon as possible, and in some cases I’ve seen it argued that the vulnerability allows the malware to be installed over the Internet.
The reality, however, is that Wannacry had a distribution method that could exploit a vulnerability in SMBv1 within a LAN in order to spread to another system that was accessible on a LAN from the infected system. This means that a network that has systems that is vulnerable will have those vulnerable systems spread the infection if one get’s infected, however, that “patient zero” cannot be infected remotely. Wannacry would still only be installed and infect a “patient zero” LAN system through some other infection vector. and that infection vector was almost certainly through trojan-horse malware of some description.
Which is not, of course, to understate that that is certainly a concern. If Little Jimmy runs an infected game mod installer, and their system get’s infected, Other vulnerable computers on the same network would eventually be compromised. However, I think the critical thing is not the security updates, but, in that scenario, the user education to avoid installing malicious software to begin with. In the scenario, for example, Why did Little Jimmy Trust the game mod installer? Should Little Jimmy even have user permissions to install software? What sort of education can be provided to allow users that have “vulnerable” habits to adjust those habits to avoid problems? Installing security Updates, Security software, firewall’s etc is, IMO, largely a way of avoiding that question, and unfortunately it pairs poorly because a user with “vulnerable” habits is often the sort who will happily disable their Anti-virus when say a game modification installer says it is a “false positive”, or who will happily give administrator permissions — if they can — to applications based in promised functionality.
Game “cheat” software often takes that approach, making a promise and then requesting elevation with that promise in mind is enough to convince some users to “take a chance”; thewse same “vulnerable” users are also susceptible to phishing scams or other things such as software programs stealing login information from say online accounts. A specific example of that would be for example simple applications which claim to “give op” to a Minecraft player. All you need to do is give your username and password! But of course it does not give you OP. Instead it simply E-mails your login information to a specified account. It doesn’t work, the user deletes the program, but perhaps never thinks about changing their login information because as far as they know the program just “didn’t work”. Until one day they cannot log in. Or, For MMOs, they suddenly find their character is poorly equipped or perhaps banned for activity because it was used for some nefarious in-game activity.
Speaking for myself, aside from the occasional Malwarebytes scan, I don’t run any sort of background AV software or firewall. In fact, I disable the built-in Windows software that provides those features. To my knowledge, I’ve not been infected in over 10 years. And even then, what I was infected with, Virut/Sality, wasn’t being picked up by any Security software, even fully updated. Since then I’ve had systems that have lacked important security updates magically not be infected in the ways that the aforementioned “security circus” would have me believe. It seems — at least from where I am standing — that the implications of security vulnerabilities to your typical end-user system are vastly overstated, and the focus on those as a means to prevent people from getting infected may be a focus in the wrong area. Instead, Users should receive education such that their “vulnerable” habits can be eliminated or at the very least they can be made aware of them. Better education for computer systems in general can help as well; knowing the difference between an svchost.exe where it should be and an svchost.exe where it isn’t can make it possible to identify unwanted software even if any installed security software isn’t picking it up.
Another topic of some interest that has taken the security world by storm is the Meltdown and Spectre prefetch cache security problems found in many Microprocessors (Meltdown being specific to Intel chips). These security concerns relate to an ability to access memory that would be otherwise unavailable by exploiting speculative cache behaviour. Meltdown involves writing carefully crafted machine language instructions in order to trick the speculative execution into providing access to small pockets of memory that would not otherwise be accessible; there are kernel-mode pages that are part of that applications virtual address space. Spectre functions similarly but requires those carefully crafted machine code instructions in order to try to perform various memory operations and carefully measure them in order to guess at the data found in certain areas of those kernel-mode pages within that processes virtual address space.
But, the reality hardly sells articles, and certainly doesn’t sell security software- which, I might add, by sheer coincidence tends to be either a sponsor or major advertiser for many of the wider publicized security publications. Coincidence, no doubt.