A couple years ago I thought a pond would be a nice addition to an empty space on the side of my old house. I didn’t know anything about constructing ponds at the time but I thought that actually doing it would be a good way to learn. After reading a few web articles on the subject I got out the shovel and started digging. The pond came together very quickly but it turned out I was kind of naive and bit too impatient in pond v1.0.

The problem was that the plastic pond liner just doesn’t lie out perfectly in a complex-shaped pond. I had some places that weren’t perfectly covered so I took all my leftover scraps that were big enough and layered and duct-taped them over those spots. I knew it wasn’t going to be perfect but I figured that enough layers of imperfect covering would eventually catch just about all the leaks. I also put down a 3-inch thick layer of cement over much of the pond.

After completing the pond I was anxious to fill it with water and enjoy the peaceful trickling of the waterfall. It was great the first night but the next morning I went outside to find the pond half empty. I filled it up again only to find it half empty again the next day. I spent half my summer patching, lining, sealing, and cementing but all I ever managed to do is slow down the rate of leaking—to the point where it would only lose a fourth of its water each day.

I finally just gave up, took a sledge hammer to it and started over from scratch—this time doing it right with careful planning. That summer I learned a lot about building ponds, got some good exercise and a pretty good tan but I wasted a lot of time and money.

What’s funny about it is that as a security guy I was defending from leaks by adding layers of protection. But water doesn’t care about stuff like that. You can take a leaky bucket full of water and put it inside another leaky bucket and repeat that with as many layers of leaky buckets as you want and all the water will still get through.

Today I read a patent application recently submitted by Microsoft that caught my attention. This invention claimed by Microsoft is titled “Disposable red partitions.” I do like the practice of using colors for security purposes because many colors have intrinsic meaning (of course, sometimes colors can just make things confusing). In this case red seems to be in place because you also see terms in the document like “panic button” and “torched” that have the same feel to them. This had to be good.

The invention mentioned in the patent application is a method for easily disposing of suspected malicious or infected system files without affecting user data. It sounds like some kind of “Format C:” that still leaves all your documents intact. According to the application, this file disposal can be triggered by “…pushing a ‘panic button,’ by a timer event, by a anti-virus checker event, or any other event that sets a standard for disposal.”

On the surface it sounds like a great idea, who wouldn’t love just having a PC with its own panic button?

But like so many security ideas, I wonder how much they thought this through. I’m getting kind of tired of all these technologies that mostly help you but they always seem to have that one small leak. I’m starting to wonder if all these 97% solutions are really just a bunch of layers of leaky buckets.

So here are some of the questions I have about “Red Partitions:”

1. Can you really know the difference between a system file and user data to the extent that you are comfortable deleting those files? What about a programmer whose user data is executables? What about system files that look like documents? What about scripts an admin writes, are they user data or system files? What about configuration files, are those user files or system files? What about configuration files that cause malicious activity?

2. Will eliminating system files clear out viruses? What about office documents or scripts that contain malicious code? Even worse, what about user files that are normally benign but are malformed to cause an overflow in the program that handles the files?

3. Would it really be that hard for a virus to fool the system into thinking it is user data? What about a virus that makes user data look like system files so it all gets “torched” along with the system files?

4. And when the system restores important files, where does it get them from? What if a virus attacks the source files instead and the system restore process actually installs the virus?

Of course I don’t know enough details about this system to say if these are valid questions or not. I’m certainly not trying to single out this one technology or even Microsoft. Perhaps they have already thought of all these things. Perhaps it’s just an invention they will never actually do anything with. Perhaps this truly is a system that makes malware irrelevant.

Or perhaps it’s just time to break out that sledge hammer.

No tag for this post.

Related posts

• No related posts.