One of my favorite sections draws from the EFF Threat Modeling page. “Threat modeling” may sound like something a management consultant would explain to you with 19 PowerPoint slides for only $45,000. But it really just consists of considering these five questions:
What do I want to protect?
Who do I want to protect it from?
How bad are the consequences if I fail?
How likely is it that I will need to protect it?
How much trouble am I willing to go through to try to prevent potential consequences?
Ultimately the goal of information security is not to protect the information assets absolutely. Protecting anything absolutely is not even theoretically possible. What we’re trying to do here is, make the information assets more trouble to attack successfully than they’re worth. If stealing a new sprocket design from the engineers at Spacely Sprockets is worth $4 million, then we have to make it cost an expected $4.5 million or more to get. That way, even success is failure for the attacker.
But if preserving that design is worth $4 million to us, we’d be idiots to spend $4.5 million defending it. We could post it on Facebook and save ourselves $500,000.
Threat modeling is really just taking a breath, refusing to panic, and applying all-too-UNcommon sense.
It’s not hard to figure out why ad-blocking is left out of a NY Times online article. But I will say that until the publications who pay for it exert some pressure on the ad networks to clean up their act, I will continue to block ads 100%.
If they refuse to let me visit, I will gladly go elsewhere.
I predict that the publications will never do this, because the cost of ad-borne malware is a complete externality to them. They never feel the tiniest pinch. They leave that to us.
Welcome to the wonderful world of the Internet of Things. You’ve probably seen this term in the news a bit lately. Perhaps you read about it in connection with a massive botnet called Mirai, or it’s even more potent descendant, IoT_reaper.
The term Internet of Things (IoT), refers to items – other than computers, tablets or mobile phones – that are connected to the Internet and communicate back to their manufacturers or distributors. A prime example of this is, printers and copiers that provide supplies consumption and problem diagnostic data back to the manufacturer. This allows service calls and supply replenishment to arrive with minimal delays in production. A great benefit, to be sure.
The problem arises when large numbers of consumer devices start using this same capability, but without much in the way of careful design or attention to the possible security compromises. A buyer of a $1,500,000 production printer may safely assume that some attention has been given to this issue by the manufacturer. They also know that $1.5M worth of business gives them quite a bit of leverage to press the manufacturer to fix it if something is wrong. But a buyer of a $20 “smart” light bulb has neither of these safety factors. For $20, you get what you get.
As more low-cost consumer devices all start turning up with internet capability, we start to see some very odd ideas expressed in this technology. Late in 2015, we learned about a vulnerability in Samsung refrigerators that exposed customers’ GMail logins (including passwords) to cyber-criminals. Many people had questions about this. “How could this happen?” “Have they fixed the problem?” My question was, “WTF were REFRIGERATORS doing with GMail logins?” This illustrates the first principle of IoT security
1st Principle of IoT security: Don’t give your devices information they don’t need. Think about what could be the impact, when information you give to something like a refrigerator is leaked to cyber-criminals. If a device works and does what you want despite the fact it’s still asking for some information, drop the matter. Its feelings won’t be hurt; it has no feelings.
As I have said a number of times in this space, the essence of security is not absolute, but relative safety. Make trade-offs intelligently between risks and benefits.
When I get a new device, one of first things I do is assess what I will gain by connecting it to my network and to the internet, vs. what might be at risk if the device’s security is not up to snuff. Most of the time, my conclusion is, “don’t connect it at all” or “connect it to the home network but keep it off the internet.” If your router has a parental controls feature, where you can restrict your kid from getting online, you can also use that to restrict your fridge from getting online. Most devices’ main reason for being connected to the Internet is to feed data back to its manufacturer that can — at the most benign end of the spectrum — be used for marketing purposes. Consider that when assessing the risk side of this question.
2nd Principle of IoT security: Don’t allow devices to connect directly to the Internet or the rest of your home network unless necessary. Figure out what you’re really giving up if you don’t connect the device. And if the answer is, “not much”? Don’t plug in the wired connection, don’t give it the WiFi password, just say no.
Brian Krebs is an information security researcher (hacker!), with a blog that is very popular in our field. He does a lot of independent investigation of cyber-criminals, and as a result he often draws their ire. He has had heroin shipped to his door, and they have spoofed phone calls to police that result in the SWAT team being dispatched for the non-existent “hostage situation.”
Last fall, Krebs’ blog website was attacked by the largest denial-of-service that had ever been seen to that point: a botnet directed over 660 gigabits/second of bogus traffic at his server. For comparison, the fastest connection available from Time-Warner in Rochester is 50 megabits/second, so this was larger by a factor of 13,200. All of that focused on a single web site will disable the servers just because of the volume.
Upon investigation, the source traffic was found to have been infuriatingly simple. The attackers had just scoured the internet for connected IoT devices and checked them to see if they still used the manufacturer’s default username and password to allow remote access. They were able to find millions that did, mostly CCTV cameras and cheap routers. Those were harnessed by the criminals to start sending Krebs a synchronized tidal wave of garbage network traffic. It’s tempting to say they were “hacked” but they weren’t, really. Their owners had offered them to the public with the documented default logins, effectively free to use for all comers.
3rd Principle of IoT security: Change the default username and password. If the install process forced users of all new devices to choose any non-default username and password, that alone might have been sufficient to stop the attack on Krebs.
So to recap: our three principle of IoT Security are:
Don’t give your devices information they don’t need.
Don’t allow devices to connect directly to the Internet or the rest of your home network unless necessary.
Change the default username and password.
Yes, there are problems in IoT security, and we’re going to need the manufacturers to address poor designs and worse implementations. But by applying these three principles, we can reduce the impact on our own lives, so that we still get some benefit from these modern things.