Slashdot Mirror
Mirai and Bashlight Join Forces Against DNS Provider Dyn (arstechnica.com)
A second wave of attacks has hit dynamic domain name service provider Dyn, affecting a larger number of providers. As researchers and government officials race to figure out what is causing the outages, new details are emerging. Dan Drew, chief security officer at Level 3 Communications, says the attack is at least in part being mounted from a "botnet" of Internet-of-Things (IoT) devices. "We're seeing attacks coming from a number of different locations," Drew said. "An Internet of Things botnet called Mirai that we identified is also involved in the attack." Ars Technica reports: The botnet, made up of devices like home WiFi routers and internet protocol video cameras, is sending massive numbers of requests to Dyn's DNS service. Those requests look legitimate, so it's difficult for Dyn's systems to screen them out from normal domain name lookup requests. Earlier this month, the code for the Mirai botnet was released publicly. It may have been used in the massive DDoS attack against security reporter Brian Krebs. Mirai and another IoT botnet called Bashlight exploit a common vulnerability in BusyBox, a pared-down version of the Linux operating system used in embedded devices. Mirai and Bashlight have recently been responsible for attacks of massive scale, including the attacks on Krebs, which at one point reached a traffic volume of 620 gigabits per second. Matthew Prince, co-founder and CEO of the content delivery and DDoS protection service provider CloudFlare, said that the attack being used against Dyn is an increasingly common one. The attacks append random strings of text to the front of domain names, making them appear like new, legitimate requests for the addresses of systems with a domain. Caching the results to speed up responses is impossible. Prince told Ars: "They're tough attacks to stop because they often get channeled through recursive providers. They're not cacheable because of the random prefix. We started seeing random prefix attacks like these three years ago, and they remain a very common attack. If IoT devices are being used, that would explain the size and scale [and how the attack] would affect: someone the size of Dyn."
The name of the bot is Mirai.
Here is the source: https://github.com/jgamblin/Mirai-Source-Code
A lot more then that.. there's also a few hundred thousand Shopify stores that were affected as well to include my company's store. Generally directly interfearing with commerce is a big no no as well.
There may very well be something I'm missing here, but I have a suggestion for how to deal with the random prefix attack.
Keep a running count of the number of requests for non-existent subdomains. Once they exceed a certain number in a short period of time, cease to respond to requests for subdomains that aren't already cached as valid.
Example: foo.com, www.foo.com, and mail.foo.com are cached. A flood of requests for (random chars).foo.com starts up. Once this exceeds 100 requests in a minute, all requests for foo.com subdomains are ignored except for foo.com, www.foo.com, and mail.foo.com.
This would still cut off access to infrequently-accessed subdomains, but subdomains with enough traffic to be in the cache would remain reachable.
Proud member of the Weirdo-American community.
"Dan Drew, chief security officer at Level 3 Communications, says the attack is at least in part being mounted from a "botnet" of Internet-of-Things (IoT) devices."
Gee, who could have seen this coming? Oh, that's right, lots of people, including me.
The IoT (Internet of Terrors) is upon us. Buckle up, baby. It's going to be a bumpy ride and it's going to get worse before it gets better...if it ever does, that is.
Personally I'm not holding out much hope- the damage is done. Millions and millions of craptastic insecure IoT gadgets are out there right now, happily botting away.
Even if starting tomorrow every single new gadget sold was 100% secure, it's too late- the world's infrastructure is already infected with mountains of this consumer-grade garbage that will be around for a long, long time.
Just cruising through this digital world at 33 1/3 rpm...
Could there be no way for consumers of IoT to secure their own devices
If they cannot be arsed to change the default passwords, thinking they'd bother with running such an app is fantasy. And that's how these botnets spread
Many of these articles seem to implicate a "bug in busybox" or "bug in telnet", but they do not describe any activity consistent with exploiting CVE-2011-2716. At most the articles might suggest elevation of privileges after getting in via a default password, perhaps via CVE-2013-1813, but probably just due to busybox not originally having been intended as a multiuser runmode so such holes are more likely to be present there.
The "bug" seems to be just journalists not understanding that a default password is not the same thing as a software bug, nor is the language or platform/OS on which malware is targeted at fault for running a program written for it.
Anyway, since vendors seem to only find it economically viable to make these should-be-local devices totally reliant on overcomplicated cloud services, or even just like to leave hardcoded test accounts on them, and many of the devices contain closed SoC/peripherals so there's no equivalent of OpenWRT for them, even enthusiasts cannot really secure them easily enough to maintain any enthusiasm for the product. They'll end up cutting the feet of some shoeless child in a 3rd world landfill as soon as their manufacturer goes bankrupt or abandons the product line.
Incidentally if you have an old busybox where and you can alter the udhpc default script and prevent the use of DHCP-acquired hostname/domainname etc, that might be worth the effort if you cannot just reasonably upgrade it.
Someone had to do it.
I have two routers. Ones a Zytel provided by the phone company and then I also have one of the russian make one (TP / Archer).
How would I know if they are part of the botnet?
Some drink at the fountain of knowledge. Others just gargle.
Your ISP's name servers don't have the records for each name. Instead, it goes like this:
Your computer asks your ISP for the IP of mail.yahoo.com .com names?" The root server says "ask dns.root.com, aka 1.2.3.4." The ISP asks dns.root.com "which DNS servers know about yahoo.com?" Foo.root.com replies "four.dyndns.org knows about yahoo.com names."
Your ISP asks the root servers "which DNS servers know about
The ISP asks four.dyndns.org "what's the IP for mail.yahoo.com?"
four.dyndns.org has the record for mail.yahoo.com and sends it back to your ISP.
The ISP sends it to you.
The ISP caches the answer for a few minutes, in case your neighbor wants to access mail.yahoo.com too.
If this seems convoluted and slow, it is. In fact, doing DNS lookups for all the ads, javascript, and crap on a web page is a major proportion of the total load time. It's not that loading of the ad banner itself is slow, it's doing the (very indirect) DNS lookups for the domain that counts ad impressions, another domain with Javascript that loads the ad, another domain where the actual ad image is, etc. Plus the site logo is on a CDN domain, the html on the home domain, some other part of the page on images.foo.com, etc. Your browser can easily look up 20 or 30 different names to load just one page.
Dyn comes in on the other side of the equation. You use your ISP DNS server (or Google's 8.8.8.8, etc.) to look up addresses. But the people running the servers have to publish those addresses somewhere in the first place, and to do so, some of them use a service like Dyn.
To use a simplified phone analogy, Dyn publishes a phone book and your DNS server is 411. If you call 411 and the operator can't find the right phone book, they can't give you the number you want.
Thanks to the War on Drugs, it's easier to buy meth than it is to buy cold medicine!