Microsoft Research Builds 'BrowserShield'

SteelyBen writes "Researchers at Microsoft have completed work on a prototype framework called BrowserShield that promises to intercept and remove, on the fly, malicious code hidden on Web pages, instead showing users safe equivalents of those pages. The BrowserShield project, an outgrowth of the company's 'Shield' initiative, could one day even become Microsoft's answer to zero-day browser exploits such as the WMF (Windows Metafile) attack that spread like wildfire in December 2005."

Just what we need

More complexity on top of bloated and horribly obscure software. That'll help security, really.

Re:Just what we need

I was ambivalent on this until I read the word "Intercept". So basically this new layer sits between the browser and the Intratubes and rewrites pages according to some predetermined criteria. Now there are two options: either they ship the signatures of new exploits to me (via an update) or the layer is on their side (like a proxy).

In the first case: why not ship the actual updates? Otherwise, how would they guarantee that Grandma will update the signatures? Maybe they will need another layer between the new layer and the Tubes, so that the new new layer will rewrite the pages in case the old new layer is not updated. This is not very sensible...

On the other hand, if they host the layer on their side, clearly I am not interested in sharing this information with MS. Either way, I don't see how it will work.

Re:Just what we need

why not ship the actual updates?

Sometimes, in the short term, fixing a bug is harder than making sure that it won't be exploited 95%+ of the time. This could be due to architecture/legacy issues, not having resource(s) who know that code base, or the fixer not knowing the code. By using signatures, you're seperating the person that writes the signature from knowing any of the code for the underlying product. Its probably much quicker since they don't have a steep learning curve, can rapidly generate signatures, and its both a cheaper and faster solution. That's not to say its good long term, but considering why IE is slow to fix bugs (MS had haulted development) this has the benefit of being independant and much easier to maintain.

On implementation, Vista will have auto-updates on be default. From their work towards making Windows far more modular, they can probably now stop services, patch, and restart them seemlessly instead of requiring a reboot. If it was proxy based, any browser could use it and we'd likely see a Google proxy too, since the data would be quite valuable and power users would naively trust Google more than Microsoft.

Re:Just what we need

AdMuncher (the single best Windows ad-blocking program, for whatever browser, bar none (IMnsHO)) already provides this functionality out of the box. Any 'sploits are easily blocked, once they're identified, as are damn near all the ads. As a bonus, when people are doing things like not properly recognizing IE7's flash handling (and I'm looking at you, SWFObject), you can easily rewrite the fairly recognizable line of code in a standard library as it comes down the pipe and fix a whole ton of "broken" webpages.

"Correct"? Probably not. Convenient? Absolutely.

AM is one of the very few modern shareware programs I not only paid for, but paid early and promote often. And no, I'm not involved, just an incredibly happy customer (and boy do the Mac blockers have a long way to go in comparison).

That's not even the real danger...

Researchers at Microsoft have completed work on a prototype framework called BrowserShield that promises to intercept and remove, on the fly, malicious code hidden on Web pages, instead showing users safe equivalents of those pages.

What happens when you mix this with Digital Restrictions Management that goes down to the hardware level? What I'm getting at is, what if it's not malicious code that is being replaced by a "safe equivalent", but perhaps a controversial story on a news website, or an important email between governments?

In the future, he who controls the computers controls the world. Digital Restrictions Management will one day give just a few computer companies control over every internet-connected computer in the world.

Some people will respond to this with "ahh.. I'll just use a firewall". Those people do not realise that firewalls will contain DRM, too.

Re:That's not even the real danger...

Hmm, so I should trust MS to know what content I want them to "clean"? No thanks. MS has a lot to do to earn back my trust and I would bet the same goes for many other computer users.

What can I say except I'd hate to live in your isolated little made up "omg MS is coming to get me" world.

No matter what the society turns to be, there'll be always people to build inexplicably complex and ridiculous conspiracy theories that all link to the same "ultimate" enemy. Does it make you feel smarter that you saw this intricate plan of Microsoft to ban your blog noone gives a damn about either way?

But it's really not that cool to throw unsubstantiated FUD around as it used to be. We call it trolling, and it's mostly unwelcome.

zero-day browser exploits

... Will just get a new name: zero-day browser-sheild exploits.

Re:zero-day browser exploits

Perhaps you joke, but it really isn't that clear whether this will work or not. TFA says

The research group tested BrowserShield against eight IE patches released in 2005 and found that BrowserShield--when used in tandem with standard anti-virus and HTTP filtering--would have provided the same protection as the software patches in every case

There were far more than 8 patches in 2005. How were these 8 selected? Were they of a specific type? Without such details, it's hard to form an opinion about this 'BrowserShield' thingie. For all we know, they selected the most convenient 8 to prove their point.

I made a similar product once.

Unfortunately, I wrote it directly into my program without giving it another name, since I didn't realize I could sell the security separate from the program.

Innovation at its finest I suppose.

Solve the problem, don't patch it

How will this even help? Will the browser shield require signatures and/or heuristics like virus scanners, and thus get outdated? If manpower needs to be invested in this technology, wouldn't the same manpower be better invested in solving the problem, rather than patching it?

Hold on a second...

... so their answer to poorly written software that is security-hole ridden is to layer more software written by the same people on top of it? Wouldn't it be easier to just write good software in the first place then actually fix, in a timely manner, anything that crops up? I'm failing to see how more bloat is going to help.

Re:Hold on a second...

... so their answer to poorly written software that is security-hole ridden is to layer more software written by the same people on top of it?

Yes, much like you would run a firewall to protect 'poorly written software that is security-hole ridden'. Sometimes writting software to catch the exploits is easier and takes less maintenance completely eliminating each and every little bug.

Didn't this already exist?

I think they're just branding the "Disable ActiveX" checkbox.

"Invents?"

Wow, Microsoft has "invented" privoxy! [privoxy.org]

Why bother!?

This just gets on my nerves. They must of spent ages planning and coding this not to mention funding.. Why the hell didn't thy put these resources into IE7 instead? Screw this "We'll protect you from the exploits", make it to the exploits are oh.. I don't know.. FIXED

It already exists, and it's called the Proxomitron

Sounds like M$ has just "invented" a limited-functionality locked-in version of the marvellous Proxomitron [proxomitron.info]. An application I truly wouldn't be without. Scrubs HTML nasties right out of the box, and also allows you to see a web page the way you want to see it. It runs with any browser, not just Internet Exploiter. And it's the right price, too.

Showing the page anyway?

It goes without saying that I didn't read the article, but it sounds like they remove the bad stuff and then show the page anyway. Why? Why not just show a page that says, "These f***ing scumbags just tried to f*** up your computer. Quit going there, and punch them in the mouth if you meet them. In the mean time, find a less dangerous source of porn."

Funny thing, MS

They never, ever have believed in, and have only on a few occasions under very great pressure given in to, *removing* software (Clippy?); they just keep on adding instead. They must have missed that one important rule that everyone in a creative profession must once discover; 'writing is scratching', or 'prepare to throw one away' as it's called in software-land. If MS were a person, he'd be declared anally retentive; some many layers of compatibility, so many stick-on solutions that are supposed to work from below. Please guys - this is a seductive, but wrong approach - think again.

Sounds like they've re-invented the sandbox.

FTA: "We basically intercept the Web page, inject our logic and transform the page that is eventually rendered on the browser," Wang said. "We're inserting our layer of code at run-time to make the Web page safe for the end user.

"The essence of the sandbox model is that local code is trusted to have full access to vital system resources (such as the file system) while downloaded remote code (an applet) is not trusted and can access only the limited resources provided inside the sandbox" - Java Security Architecture [sun.com]

Great!

Now I can download cracks and keygens for MS products without fear!

Oh my....

There goes MS again. Let me guess: it will show a big ass shiny shield with a really cool animated graphic and ask "Are you sure you want to execute this malicious code?" and when the user clicks the Ok button it will ask once more just to be sure.

Personally I'm very affraid about MS sniffing my code. Experience shows that it will let tons of lines of malicious code pass, while locking down many good codes out there.

When those people will learn to stop trying to do magic tricks and be serious? A solution to browser flaws already exists and it's not magical at all, but technical: it's called "patch".

Bizarro!

WTF? This is the kind of approach that would be used on someone else's propriatary legacy software, or on some piece of hardware to keep it working without altering the thing itself. What are m$ saying? 'Our browser code is such a POS that we don't know how it works anymore'? 'We lost the source code ages ago and we cannot be bothered doing the job right'? 'We have so much market share that we really don't give a crap anymore, pass the crack pipe and the stock options'?

Potentially Unsafe Code Samples

UNSAFE HTML:                            REPLACE WITH:

<a href="*.apple.com*"*>*</a>           <a href="mailto:/webmaster?Subject=Your%20Site%20Suck s">I am a Communist.  Please mail me if you love Osama bin Laden --- I certainly do!</a>
<body*>*Linux*</body>                   <body>This page cannot be displayed due to faulty programming in the server's OS.</body>
<embed src="*.[^w][^m][^av]"*></embed>  <b><u><i><blink>This page contains content created using a pirated version of Windows Media Player.  Contact the police.</blink></i></u></b>

Strangling metaphors

That's like putting a Robin Reliant into a huge metal box to make it as safe as a Volvo. Or something. More coffee...

Re:Strangling metaphors

It's a Reliant Robin, dammit. The model is a Robin, made by Reliant.

Ahh much better now

So instead of this dangerous page which will try to install malware [mozilla.com] we'll get a cleaned-up and safe version [microsoft.com]

I'm sure glad MS is out to make the interweb a better place for everyone.

Magic

The research group tested BrowserShield against eight IE patches released in 2005 and found that BrowserShield--when used in tandem with standard anti-virus and HTTP filtering--would have provided the same protection as the software patches in every case, Wang wrote in a research paper.

I'm afraid without more information this sounds too much like magic. "Vulnerability-driven filtering should prevent all exploits (of a flaw) and should not disrupt any exploit free pages."

How is the technology filtering, what is it filtering, and how will it differentiate exploit free from exploit-ridden pages? If it can simply detect them why not just block them?

Microsoft Research has produced amazing technologies in the past and most of their current research is also very promising, in the area of GUI design, security, algorithms and so on. I just hope they are in tune with what Microsoft is already doing in Vista to avoid redundant layers of technology.

Also there's always the danger of Microsoft slapping a technology on IE for pure PR reasons ("haha Firefox has no filter!").

But I believe we have a case of poorly written article here. It's not uncommon that reporters simply have no idea what they are covering and coming up with wrong conclusions on what fundamentaly the shield is.

I'd say wait for the opinion to "mature" a bit on this technology.

Rice's Theorem anyone?

So they are trying to build a machine that can decide whether arbitrary code is malicious or not - I highly doubt that this is possible in respect to Rice's Theorem [wikipedia.org]. It basically says that every aspect of an complex system cannot be decided. A well known example is the halting problem [wikipedia.org]: You cannot decide whether a turing machine (or an algorithm running on it) will ever come to a stop, or is going to loop forever. And since binding processing time via infinite loops could be considered malicious behaviour, and most script languages are turing complete, an automaton will never be able to decide if a specific piece of code will harm your system. It is possible that certain aspects of a program (opening files in strange places, writing to files that should not be written to) raise suspicion, and certain chracteristics of code might also leed to detection similiar to the work virus scanners do - but I still prefer the good ol' evil bit [wikipedia.org].

Already in anti-virus software

This sort of thing is already in anti-virus software. I use Avast! (free edition) and it has a "Web Shield" module (sounds a look like "Browser Shield", doesn't it?) that transparently proxies web traffic. When it finds anything nasty it pops up dialog box asking you if you want to download/access it anyway or "abort connection".

While this is all well and fine, would it be too much for Microsoft to just patch their bugs?

This will be fun

for webpages made by Frontpage.

from Microsoft Research

Do note, this is from Microsoft Research and not a core developement team working on the browser. There will always be bugs in software, just like virus can exist on any OS (though some may have more than others). MSR has been renowned for coming up with interesting solutions for interesting problems. I mean Firefox, Opera, Safari, and any other browser out there has been hit with exploits before. I mean every update of Firefox I download has multiple security updates. I'm not saying a perfect browser can't exist, but the road to get there requires both time and effort, espeically while trying to add new features to keep up to date to be able to compete with other browsers.

Just like how AV software isn't the solution to viruses, it's done quite well in protecting many systems. I personally don't understand exactly how this browsershield works, but from what I can grasp, it seems to be an additional check before loading the page into the browser and removing any malicious code. How it detects the malicious code is not clear, but having seen interesting research come out of MSR, I have my faith in these guys to have come up with an interesting solution.

well it's the Microsoft way

1. create product with security leaks
2. receive complaints
3. do not solve security leaks but instead, build a wall around them
4. go to sleep and forget about 1.

Wrong-Headed!

*sigh* So they are STILL trying to put bandaids on their old, insecure, highly-patched (and therefore low quality) software rather than ditching insecure communications protocols and writing a simpler browser that is secure from the gound up.

Yep - Microsoft is all in favor of security - so long as it maintains backward compatibility and they don't have to throw anything away.

Bye bye karma

I know we all love to hate MS but this is a good idea.
First off, I have seen first hand some of MSResearch fairs and they is a lot of great stuff coming out of them. Anything that comes out of those labs is worth at least some thought before you dismiss it.

That aside, stripping nasties using a simple system before they reach a more complex system isn't really a bad idea. All of our mail servers have some sort of filter that does this (granted, more for dumb users). IIS 5 did this using a tool that was later built into IIS 6. Hell, firewalls aren't a much different idea. Most of us already run some sort of proxy software to block popups, scripts, or ads. All MS is proposing here is the equivalent of proximity or similar proxy software.

Do we just hate this idea b/c MS is doing it?

Tryed with anti-virus software. And failed.

Well, I thought anti-virus software vendors already failed at similar effort. Every new virus out there first disables all known anti-virus software.

It all boils down to question: how could you tell malicious content from good one??? You would have to resort to signatures. That wouldn't help against 0day exploits in no way, since on that day 0 most signatures are not yet updated.

From the article it sounds more like standard corporate firewall functionality: "block all what looks like HTTP redirect, since that can IE exploit", "block all .exe attachments since that might be Outlook exploit", "block .wmf since that might be IE/Outlook exploit", etc. Nothing new.

Malicious hackers typically embed scripts on Web sites and then use social engineering techniques to trick unsuspecting visitors into downloading Trojans, bots, spyware programs and other harmful forms of malware.

With BrowserShield, Wang argues, many such attacks could be blocked. BrowserShield can be used as a framework that rewrites HTML pages to deny any attempt at executing harmful code on browsers.

Buhahaha! Very funny!! They at Redmond take Windows security very very seriously - they have put best PR people on it!!!

Good luck at identifying that "harmful code," darling!

P.S. And for that "rewrites HTML pages" bit be sure to have M$' lawyers ready. Few content providers would like idea that their pages may be rewritten by the software monopolist.

P.P.S. Would M$ ever learn? How long they intend to have that "ActiveX" crap enabled in their browsers by default?? How many sacrifices they intended to make???

P.P.P.S. On related news from Germany, my employer (about 150 desktops) 1.5 year ago has banned M$IE. Firefox and Opera must be used to access inter/intranets.

Security from MS cannot work

No, hold on, not a MS-bashing comment, please read on.

It's not that MS is "inapt" or that they can't get their act together, it's simply that computers are computers, people are people and the mix of those is by its very nature unreliable and insecure. No matter how good you make it, there will always be tiny cracks in the security, be it for technical shortcomings or flaws in human nature that can be manipulated by social engineering.

Now, MS is the biggest manufacturer of operating systems. This shield will, invariably, also be present on every PC running their OS. So the first thing you have to defeat, as the attacker, is this shield. Can't get past it, don't bother continuing trying to defeat other security software that may or may not be present. This shield WILL be present!

So every attacker out there WILL have to come up with a cracking scheme. No matter what the cost, no matter how long it takes. It HAS to be cracked.

Thus security from MS cannot be relied on. Not because it is insecure in any way. But because every piece of malware HAS to come with some procedure to circumvent MS security. It will invariably have countermeasures in its arsenal.

It's been done

Filtering proxies, like the Proxomitron [usask.ca] or Privoxy [privoxy.org] will do some of this for you. The thing is that this doesn't really work that well for security. You can reduce some exposure [kyeu.info], but it there are things that will get past your checks.

And the MicroSoft implementation seems to be a limited sub-set. It won't even block ads.