Yes, security should be managed in layers, but the management of those layers should be trusted to the "right" person. A/V companies believe that they are the right person. Apple believes that they are the right person.
If you're going to levy the "security theather" attack at Apple, we could just as easily fire that shot in the direction of A/V companies. When "malware" came on the scene, A/V companies were completely unprepared. Their response was that "malware isn't a virus". From a technical perspective, they're correct, but from the user's perspective, it was a huge failing. Virtually everyone I know who has been affected by malware said the same thing: "But I pay for Kaspersky/McAfee/Avast/Whatever anti-virus, how did I get infected!?" I'm supposed to believe that these are the "right" people with whom to entrust the security of my computer?
Most operating systems allow you to install software. If the user installs software from untrustworthy sources an A/V is the only thing that can protect him/her. No it isn't perfect but few claim it to be.
My statement can't be taken without context. Stated with more context: "there is no inherent aspect of an operating system that favors A/V as protection from malicious software over any other (like code signing)." In practice, quite the opposite is true. We've proven, through practice, that A/V is a poor solution.
This has never been an inherent possibility on mobile operating systems.
The only way to run executable code on iOS is to install it via the AppStore, to get to the App Store you need to pass the review process, you have to be really clever to be able to hide your malware. Even if your malware gets trough you are still stuck in your apps sandbox and once Apple realizes what you have done, they can remotely pull all instances of your app and the police will be waiting for you outside.
So virtually it's impossible to spread virus on iOS, an anti virus would be just crapware.
Let's hope no one sets up The International Obfuscated Objective-C Code Contest; or that the code reviewers at Apple are competent and experience enough to work with advanced software security threats.
"Even if your malware gets trough you are still stuck in your apps sandbox"
Sandboxes can and are exploited. Recently [1]
"So virtually it's impossible to spread virus on iOS, an anti virus would be just crapware."
Sure on HN we all know that security is more than (anti) viruses.
[1] http://www.techspot.com/news/47731-google-rushes-out-chrome-...
Apple isn't doing thorough code reviews of every app they get. They're just checking sanity and adherence to "the rules". Even if the code was obfuscated and they wanted to review your code, they could write you an email that says "your code is unreadable, make it better or you're not going to get on the app store".
Sandboxes can and are exploited.
So are you saying that because Chrome was exploited, that the sandbox model in iOS some how makes the system less secure? The notion doesn't follow. Yes, sandboxes can be exploited, but it's NOT EASY TO DO THIS. The idea behind good, secure design is not to try to secure from every attack vector possible, but to eliminate low hanging fruit entirely, and make elaborate breaches very, very expensive to find and create. Apple is doing all of this here with the sandbox design, and it works quite well for them.
Sure on HN we all know that security is more than (anti) viruses.
Case-in-point: iOS. Quite secure, doesn't require anti-virus software to be as secure as it is.
One can still do a driveby attack where a malicious webpage kicks Safari's teeth in and finds a way through the sandbox.
(Or replace Safari with 'any app that consumes network content')
"I'm safe because I'm running McAfee/Kaspersky/Norton/what-have-you." I'd say pretty much every infected machine I've dealt with has had anti-virus software on it.
The first author[1] of a successful worm[2] did become quite famous from it, later did a startup[3] and even later went on to be one of the co-founders of this thing called YCombinator[4].
[1] http://en.wikipedia.org/wiki/Robert_Tappan_Morris
[2] http://en.wikipedia.org/wiki/Morris_worm
[3] http://en.wikipedia.org/wiki/Viaweb
[4] Seriously?
The existence of the Morris worm doesn't really say much about the resilience of today's Linux to worms or viruses.
Yes, malware can be a lot more sophisticated now than it was in 1988. In many cases it doesn't need to be though. For example, there was a comparatively recent case of a botnet infecting Linux-based routers which lacked front end authentication[2](!!).
Exploitable software runs on many platforms.
The existence of the Michelangelo virus[1] doesn't really say much about the resilience of today's Windows to worms or viruses.
[1] http://en.wikipedia.org/wiki/Michelangelo_(computer_virus)
On windows or mac, getting a virus to run is generally just a matter of packaging it with some downloadable software. In linux, pretty much everything is available through package managers. You can't just throw a virus into a torrent for photoshop or whatever.
That being said, I'm by no means an expert.
I've seen drive-by website infections, malicious software installs using social attacks, and malicious attachments infecting Office. When Flashback alone infecting the better part of one million Macs, it is now newsworthy.
What if there is a hack/mistake/bug in the appstore ?
There's no need to worry about security because the app store guarrantees that.
The Apple review process, draconian as it may be, almost certainly has a real, strong effect in actually preventing malware.
Actually average people would probably know enough to think "someone could just lie when asked - that doesn't provide protection".
To me "security theater" is doing things that seem to appear to lock something down in the eyes of an average uninformed person. Average is not Bruce Schneier or an 8 year old. It's a typical traveler who believes they are secure because they see checkpoints, have to take their shoes off, and have to have their laptop scanned.
It's exactly like the TSA having magic terrorist detectors - but you can't be told how they work because of security
The de Cartes-style demon argument is generally a weak fallback. The end game of this argument is that you can't trust anything because you can't fully trust anything.
The reality of the situation is that the app store is most definitely more secure than Windows's distribution model because it normalizes the vehicle for software delivery. Security is scrutinized and narrowed down to one place. Users become less trusting of software coming from 3rd party sources (detrimental in some cases, to a more free and open platform), but added security is definitely gained as part of the tradeoff here.
But as the recent flame worm, signed by a microsoft trusted certificate (http://isc.sans.edu/diary.html?storyid=13366) shows central security systems aren't automatically foolproof - and if you are prevented from having any sort of local control or anti-virus by that same central security system, you can be up a certain creek with a certain paddle
That was never contested by anybody. What the flame worm showed was that there needs to be stronger security around private key portions of signing certificates.
any sort of local control or anti-virus by that same central security system
Funny how that works. Anti-viral software is a central security system that uses similar distribution and signing techniques as the app store! Not to mention, anti-viral software doesn't protect you from zero-day exploits, unpatched software, and brand new malware that tends to be the thing that causes the most problem. Not to mention metamorphic and polymorphic malware, which is getting more and more common and runs circles around modern AV software.