Here's a headline we never thought we'd write 20 years ago: Microsoft readies antivirus for Linux, Android Microsoft has extended its antivirus package for servers – better known the Defender Advanced Threat Protection (ATP) for servers suite – to Linux as a general availability release. Redmond said today that the ATP-for-Linux port will run directly on Red Hat Enterprise Linux, CentOS, Ubuntu, SUSE Linux Enterprise Server, Debian …
Sounds like a good idea!
Cool?
Good idea?
You're either joking or just didn't notice when your mates spiked your ale.
Fortunately, I'm not running a server or have responsibility for maintaning one.
But if I were I would not let any MS software crap near it.
Not if Hades froze over.
Twice.
Good idea indeed ...
Sure.
O.
"Linux doesn't come with any security built in."
Well maybe not the masses of add-ons needed when you want to run Windows securely but that's because Windows was never built with security in mind and has been playing catch-up ever since it was released.
Linux, the kernel, on the other hand has followed Unix in as much as security is something thought about and implemented. I'm sure Linus Torvalds had security forefront in his mind when starting the development of Linux and I don't think he is going to change now.
Linux, the kernel, on the other hand has followed Unix in as much as security is something thought about and implemented.
Which is nice, but then they go and layer a shower of shit called systemd on top of it which, no doubt, has been riddled with exploitable flaws from day one.
Even the hallowed ‘su’ has had its own exploitable problems, and not in the dim and distant past either...
To be more abstract about it, Linux is a legacy OS written in a dangerously tricky language, C. Starting again in Rust or converting over to it is probably the way to go to be sure if eliminating whole classes of bug. Fuchsia has rusty bits, Redox is all Rust, and even MS is thinking about such a move quite seriously. Linux may start looking positively antiquated before too long.
Yeah but it surely wouldn't be thrashed by a font rendering. Since you seem to know much about modern languages, you're not suited for C programming. That is reserved for real pros who know what they're doing. It is not a dangerously tricky language it just requires you to master it before actually write a single line of code. It also requires the programmer to concentrate on what he is doing instead of adding bells and whistles and rushing code into production.
The mistake you're making is to think that a real pro would stick with a language like C or C++ when there is a better alternative (and Rust is an especially useful alternative because it integrates with C nicely). C/C++ requires extensive care, review and testing to ensure memory correctness. The Rust compiler does all that dreary work for you.
You can stay there, back in the past, chiselling out code in with you little C hammer line by line, but the world will likely move on to Rust. Only an amateur would cling to the past for the sake of it.
Take a look at the Redox project. The most notable thing there is how quickly it's gone from a standing start in Aug 2016 to a running desktop only a short while later. That's pretty fast work, for a small team.
Legacy projects have to consider the possibility that they'll get left behind, and cannot afford to be dealing with memory fumbles, archaic language syntaxes and avoidable bugs whilst competing projects aren't.
... but then they go and layer a shower of shit called systemd on top of it ...
Quite so ...
But that's Debian and that Poettering DH's take on what the init system should be like ie: an abominable piece of steaming crap.
Inexplicably, quite a few distributions have followed suit even though it flies in the face of the Unix philosophy of 'doing one thing and doing it well'.
It is not part of Linux (the kernel).
O.
Poettering worked at Red Hat when he and Sievers started the systemd project, and Fedora and then RHEL were the first to adopt it, with Fedora enabling it by default. Debian would not select it as the default in 8/jesse until years later, and not without much discussion and controversy that continues to this day.
Red Hat have been the only ones to adopt systemd without some kind of massive internal controversy (that we know of).
It is not part of Linux (the kernel).
True, but:
1) Irrelevant; try using the Linux kernel without a user land on top of it or (in the case of microservices) being compiled in with it,
2) Irrelevant, as the Linux kernel's security correctness is for nothing if the available userland ecosystems that run on top of it are themselves flawed.
3) Ignores the fact that, whilst the Linux kernel has been moderately OK w.r.t. security flaws, it's been far from perfect. I chose not to mention that originally largely because I'd rather have a dig at systemd, which has been a needless, unasked for mess imposed by a few people in a position of influence on everyone else. The Linux kernel project in comparison has been run a lot more carefully, and it would be a shame if it fizzles out simply because it fails to take a bold decision at an early juncture.
Plus, a Linux kernel re-write in Rust would be an ideal opportunity to re-licesnse it at the same time; re-written modules can be licensed in anything they want; GPL3, whatever.
To be more abstract about it, Linux is a legacy OS written in a dangerously tricky language, C. Starting again in Rust or converting over to it is probably the way to go to be sure if eliminating whole classes of bug.
^^^^^^^^^^^^^^^^^ Nonsense written by somebody who never quite grasped pointers.
Starting again in rust ? No concept of just how long it takes to write code properly. I've come across this type of programmer before. Wants to rewrite everything because he doesn't understand other people's code. I once let it happen, what a disaster that was.
I'll have you know that this old AC has had a firm grasp of pointers since the 1980s, when there was no choice in the matter, and still does.
I also have a firm grasp, having been around for a long time, on how much project time gets used in review, testing, bug hunting and support, and how valuable it is to the goals of a project to use something that reduces that. Writing code in C/C++ and ensuring that it has been written properly simply takes more time than it does in Rust.
Nonsense written by somebody who never quite grasped pointers.
Listen up, youngster, it's supposed to be the hot new bloods rushing off to use new technologies and lanaguages. If you're being left behind by us old fogies, you're in for a nasty shock.
Yes, that's mostly what Linux admins would want it to do. If you have a Linux server that's handling Windows files/attachments like mail servers, web servers, or whatever and your security rules require virus scanning then you have to install antivirus on Linux.
A lot of people are paying a lot of money for commercial AV software on Linux for that very reason. The article doesn't say, but if this is free or at least much less costly than the usual suspects Microsoft is going to get a lot of converts.
This disease isn't going anywhere near my Linux systems. Disease you ask? Yes, MS is spreading almost as fast as CV-19 and is just as deadly to your data as the real thing is to your health. Once infected there will be no way to totally eradicate it other than a complete security erase and re-install.
The 'emperor' penguin says no!
Microsoft is doing a lot to embrace Linux with the new features in Windows. At some point I fully expect to see a Microsoft licensed distribution of Linux that will have all the Powershell stuff crammed in. Licensing to start with will appeal to companies that maybe have a small Linux footprint and currently either use commercial distributions (RHEL etc) or open source with no, or limited support. It will be commercially beneficial to use the MS distribution because it will be bundled with any existing MS agreements. Companies are used to dealing with Microsoft so market share will increase. In education commercial Linux distributions with support are very expensive when compared to the Campus agreements Microsoft put in place.
Once it is in, it will become increasingly difficult to get out. Microsoft are then in a win-win situation where it does not matter which OS you use, they will still get your money.
For better or worse, commercially Microsoft could end up being dominant in both Windows and Linux.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
