The Internet of Compromised Things 62
An anonymous reader writes: Jeff Atwood has a post about a security threat that's becoming more prevalent every day: spreading malware through a compromised router. "Router malware is the ultimate man-in-the-middle attack. For all meaningful traffic sent through a compromised router that isn't HTTPS encrypted, it is 100% game over." He links to a thorough technical analysis of how even HTTPS encrypted traffic can be subverted. Atwood provides a list of suggestions for keeping your router safe that probably won't be any surprise to people reading this site, and he further recommends only browsing on an unknown router if encryption is available. What I'm curious about are the long-term implications — is there a way forward to re-establish trust in our router infrastructure? What can the open source community do to speed this along?
Re: Don't routers already run BSD? (Score:1)
Nearly all home routers run Linux, genius.
Re:Don't routers already run BSD? (Score:5, Interesting)
Either way, the average person is only going to use the web interface or the software that the manufacturer provided that runs on the computer, if any. They won't be in position to fix anything that's broken if the manufacturer does not provide either an automatic means or a simple means to do so.
I think it'll eventually come down to a regulatory issue. Tech companies and those companies that use consumer-facing electronics (like car companies and their infotainment and body-control computers) have proven that they're not interested in maintaining their arguably defective products. Don't mince words, bugs are defects. Companies need to be taken to task over both this and over the increasingly rapid discontinuation of support (like factory-shipped apps on cell phones that stop working and can't be updated because new versions require OS updates that aren't provided) such that companies end up with mandatory windows of support until the last product ships, where all bugs and changes in communications protocols and services are maintained, such that devices that consumers have paid good money for actually last as long as their pricetags indicate that they should. For smartphones I think that window should be five years. For things line broadband routers, it should be at least five years, and I'd argue that it's not unreasonable to demand closer to a decade. For cars, where the average age of cars on the road is now something like twelve or thirteen years, it should be at least a decade for basic feature maintenance and probably another ten years for critical bugs that compromise the security of the vehicle's systems, like these easily unlocked cars we're hearing about.
Yeah, it sucks to have to maintain old code, but I'm very tired of having to pay for defective products whose features begin to stop working when the companies that wrote those features decide to change directions.
Re: (Score:1)
Yeah, it sucks to have to maintain old code, but I'm very tired of having to pay for defective products whose features begin to stop working when the companies that wrote those features decide to change directions.
So the solution is to mandate by law a minimum support period?
Re: (Score:2)
Re: (Score:3)
Manufacturers should have to support smart phones for five years? What have you been smoking, buddy? And can you pass me some? The hardware itself (screen, casing, battery, etc.) of most phones does not last 5 years - why should the software?
My phone is a Samsung Galaxy SII [gsmarena.com]. I bought it when it was newly launched. It is now four years old. My previous phone was a T-Mobile G1 [gsmarena.com], also sold as the HTC Dream, the first retail Android phone, which I also bought when it was newly launched. I still have it and it actually still works, but we replaced it in part because of application problems from being limited to Android 2.3.
Just because you replace your technology frequently doesn't mean that the rest of us do. Frankly I'd rather spend my mone
Re: (Score:2)
Just because you replace your technology frequently doesn't mean that the rest of us do. Frankly I'd rather spend my money on other pursuits rather than re-buying the same theoretically-durable goods all of the time.
Agreed. My previous cell phone was a Qualcomm QCP-1900 I bought in 1998 for $200. It was one of the first all-digital phones. It only made voice calls, which is all I needed. I had to get a new phone last month because nTelos (originally PrimeCo in my area) sold their spectrum to Sprint in my area and, for some reason, my phone would no longer work on the network.
Yes, I arguably got my monies worth from the phone, but it *still* worked just fine - grrr..
Just to note: I switched to Ting and got a Kyoc
Re: (Score:2)
Re: (Score:2)
Re: (Score:3, Insightful)
Or you could, you know, go to an ATM, withdraw some cash, then walk into $shop and buy one there, using the cash...
Re: (Score:3)
You can for ADSL routers. Cable service routers usually combine modem and router into one box, and DOCSIS authenticates this device with the other end of the network cryptographically - even if you wanted to replace it, you couldn't. If you check the fine print you'll usually find that the modem-router is the property of the cable company and serves as the demarcation point.
Re: (Score:1)
Although US cable companies will happily lease you a modem for an exorbitant monthly fee, you're not tied to using their modem or router. The only hardware I absolutely have to use from the cable company for my setup is a cable box, for TV. I also leave a device for switched digital video, and a cablecard. I don't have to use the last two. Maybe things are different where you are, or with a different provider, but being able to bring your own modem and router is pretty standard here in the US. Whether those
Re: (Score:2)
I'm in the UK. We have exactly one cable company in the country - there used to be more, but there was a string of mergers until only one stood. The one is Virgin Media, and they insist on the use of their own branded router, the 'Superhub.' You can set it to disable all the routing functionality and just act as a dumb modem though.
Re: (Score:2)
I could also just have my router do the dhcp and use some rules to decide what traffic uses the vpn.
My point is that even without a pass-through mode, you can use your own router to protect your devices.
Re: (Score:2)
FIbre to the door here. I don't even *need* a router to connect. Of course, I have multiple devices and desire wireless connectivity, so I do need a router or similar for those things. But my provider doesn't care what I plug in on my end--any off-the-shelf device will do, as far as they're concerned. Of course, the flip side of that is that, if they see that something that affects their network adversely is coming from my connection, they reserve the right to cut it off and make me ring them up to find out
Re: (Score:2)
Re: (Score:2)
You can for ADSL routers. Cable service routers usually combine modem and router into one box, and DOCSIS authenticates this device with the other end of the network cryptographically - even if you wanted to replace it, you couldn't. If you check the fine print you'll usually find that the modem-router is the property of the cable company and serves as the demarcation point.
On Cox Communications networks you can use your own Cablemodem, but it must be one from an approved list. Unfortunately that means no buying a used Cisco 2800-series router and throwing a DOCSIS module into the HWIC slot, even if theree was a DOCSIS module conforming to a new enough standard.
All one has to then do is contact them and have them "provision" the modem, which I assume means entering its OUI into your customer record that it appears in their allowed-devices list.
Re: (Score:2)
Why must the router and modem be a single bit of kit? Don't we have Ethernet to serve as a well-developed cross-connect?
What could possibly be gained from a suitably-new DOCSIS module in a Cisco 2800-series router (or, in my case, an impossibly-conformant VDSL module on aa Asus RT-N16) that cannot be accomplished with a provider-provided Ethernet-connected modem (whatever that is) and an Ethernet-connected router of my own choosing?
These Cisco 2800-series routers you speak of: They do have Ethernet, don't
Re: (Score:2)
Simpler and more likely solution for the government, just watch everyone.
If not that, Oh yes, you bought a router for cash, definitely watching anyone who does that.
Re: (Score:3)
http://myopenrouter.com/ [myopenrouter.com]
If you're interested in security, you'll buy a router which you can flash and program to your own liking.
I don't know how to counter a custom spying chip that might be embedded on your router's board, but defeating software is pretty damned easy.
Pfft. This all misses the point. (Score:4, Interesting)
If you actually care about security maybe you shouldn't let information from the internet control your computer. I mean sure they can edit your webpages via a router to insert whatever, but the inserting whatever part is could just also be malware on the internet proper.
The point of TNO is that you might as well assume your router is altering webpages and inserting malicious content (either that or there's some already on the internet). One should view the internet as a black box of security threats and then just go from there. Yeah, routers can do that stuff. Security is what you do with that understanding. You don't get a router you can trust. You don't trust the router or the rest of the internet.
trust consumer grade routers? ha! (Score:2)
if you want to trust a router, you better be sure there is a vested interest in it being secure. new consumer grade routers come out as often as car models and the firmware is common crap that has half-assed security in the name of features.
i wouldn't trust a router without it being open source and open hardware down to the ICs. so... is anyone making a RISC V based router?
Re: (Score:1)
Well, there's this [bananapi.com], but it's probably not open enough for you, and almost certainly tries to do too much at once, but it's a nice OpenWRT [openwrt.org] platform.
Trusted Network Fallacy (Score:4, Insightful)
The people who designed the internet had the right idea: Dumb network, intelligent edge. Perimeter security and trusted networks are dead ends. Communication is from endpoint to endpoint. The network shouldn't even matter. You might be running IP over avian carriers if that's what you need to do to get a connection. But if you need to trust the network between the endpoints, you're doing it wrong. Even if you could trust your own router, do you trust the ten or more routers behind it? Ubiquitous encryption and authentication with IPSec is possible with DNSSEC supplying the keys.
Re:Trusted Network Fallacy (Score:4, Interesting)
I think you still have to trust some aspects of the network. Sure, DNSSEC can provide some protection, but what if your ISP's DNS server is compromised to provide bad information? I suppose you could verify it against other servers. Can you trust that the routers your packets pass through are properly routing your traffic to the IP you want it to reach? If done right, compromising these things could be almost invisible to a lot of users. I think you have to trust certain aspects of the network, though you should use encryption to protect against MITM attacks. I think you can avoid many types of exploits, but you have to trust something in order for the internet to function. The idea of using HTTPS is a step in the right direction, except that CAs can't be trusted and the biggest one has a horrible record of security. Add to it that most users are ignorant of HTTPS and many applications don't reveal the protocol to the user and you have a problem. Can you trust that mobile apps and a lot of other software that doesn't explicitly reveal its protocols to the user makes use of encryption? Sure, you could sniff the packets, but who does that? I just don't think you can entirely remove trust from the equation, though we can do a lot better than we do now.
Re: (Score:2, Insightful)
what if your ISP's DNS server is compromised to provide bad information?
That's why you need to use DNSSEC, and by use I mean verify that you got authentic data, which DNSSEC lets you do.
Re: (Score:1)
You have to trust someone else if you don't already have a secure channel to all other communication endpoints. You're correct in observing that key management is a critical aspect in cryptographic systems, and that government actors are not automatically trustworthy. You are wrong though to assume that you have to trust all governments and commercial interest groups. DNSSEC does not work like SSL, where any CA can issue certificates for any domain. With DNSSEC you only depend on the delegation path to the
Re: (Score:1)
The people who designed the internet had the right idea: Dumb network, intelligent edge.
No they didn't. There was no default encryption. There was no built-in means to confirm that one was speaking to the host one thought one was speaking to. There was no method in the protocols to detect MIM attacks. Basic protocols require one to be honest about one's identity (like SMTP) and have only relatively recent bandaid attempts to mitigate the disaster that is their implementation. The original Internet was all telnet and UUCP and FTP and SMTP and POP3 and password flying in plain-text all over
The entire friggin' internet is compromised (Score:4, Informative)
Nevermind your own dinky router: any traffic you send on the internet is exploited by greedy "big data" companies and rogue 1984-style government agency. And encryption doesn't stop them from watching what you do...
Re: (Score:2, Interesting)
This is unfortunately the ugly reality: the internet as we knew it is dead. What many dreamed would be an empowering tool for the masses became the ultimate instrument of power and control for the Ruling Elite. We can't even leave it because all services are being brought online and online only. We have been enslaved and there's nothing we can do. In the end, I almost think the Ruling Elite deserves its great victory: they have been most astute and far-seeing in their acting. It's the culmination of a 20-ye
Re:The entire friggin' internet is compromised (Score:4, Insightful)
This is unfortunately the ugly reality: the internet as we knew it is dead. What many dreamed would be an empowering tool for the masses became the ultimate instrument of power and control for the Ruling Elite.
To be fair, it's actually a little bit of both.
Having access to all the compiled knowledge of mankind is empowering for any and every person with internet access, as is being essentially free to contact nearly every other Worldly citizen via the web. The ability to monitor an individual's access to that information is maddeningly power grubbing for the government's surveillance state.
Being realistic, if it was not advantageous to the ruling elite, would they let us keep it?
Re: (Score:2)
In what way has it ever been about empowering the masses?
Remember, freedom of the press belongs to those who own the presses. LIkewise, freedom on the internet belongs to those who own the internet - in this case, corporations who sponsor the backbones and who connect our homes with it.
Always been the case. As long as someone big and power owns a part of it, they control it.
Re: (Score:2)
No, but it can make watching you sufficiently expensive and impractical as to render it impossible on a non-targetted basis.
SSL interception is possible, but if any ISP or intelligence service does it on a large scale it will inevitably be noticed.
Re: (Score:3)
I wish this weren't modded up. Really, I do.
"any traffic" implies "all traffic" and it's simply wrong that "big data" is somehow exploiting, for example, the OpenVPN traffic between my laptop and my home mini server, nor are they making use of anything going on over SSH.
And encryption doesn't stop them from watching what you do...
And this is just silly. Of course it does! It is *not* a perfect tool, but it is a damned good one, the engineers did their job. As with any defensive/offensive technique, there are ways to mitigate it, and there are ways to bolster against
HTTPS is not the only encryption (Score:3, Insightful)
Re: (Score:3)
TLS is no more broken than SSL, and can be used by HTTPS sessions. If anything, SSL is the older and less reliable protocol, and that is what SSH is built over. So is sftp.
Regardless of whether you are using TLS or SSL, you are relying on the same public key infrastructure system to identify hosts. So I don't know where you get the idea that SSH is "more secure."
Re: HTTPS is not the only encryption (Score:1)
SSH does not use SSL and also has no CA (this is what parent talked about).
Re: (Score:2)
With no CA, how can you claim that it's "more secure"? You have no way of knowing the certificate was actually issued for the server you connect to if that's the case!
Here I'd always thought people were just using self-signed certs, but if there is no CA, it's not even that secure.
Re: (Score:1)
Re: (Score:2)
The first thing I notice about that article is that it help spreading the misconception that HTTP is the only use of Internet and HTTPS the only encryption scheme. I must say, I feel much safer knowing my SSH sessions are not HTTPS-encrypted, because the certification mechanism is completely broken.
The HTTPS certification infrastructure has problems, but to say that SSH is better because it doesn't have one at all is rather bizarre. If you'd like exactly the same sort of security from HTTPS that you get from SSH you can verify HTTPS certificate IDs manually, and you can install a browser extension that warns you when they change.
Re: (Score:1)
Sorry, but it does not work. People who manage SSH servers know what a private key is, they treat it as a precious file and keep it when, for example, restoring from a hardware failure. Only when the key is compromised do they change it. If they are really serious about it will even distribute the fingerprint along with other necessary information when opening new accounts. You can verify it carefully, and then it is once and for all in the known_hosts file.
People who manage HTTPS sites, on the other hand,
Re: (Score:2)
Re: (Score:1)
Absolutely not. My argument is that the TLS authentication architecture is broken beyond repair.
The SSH authentication system does not scale, but it is sound, and it could be made to scale without changing the base principle. The TLS authentication can not be repaired without changing it from the core.
Re: (Score:2)
Absolutely not. My argument is that the TLS authentication architecture is broken beyond repair.
The SSH authentication system does not scale, but it is sound, and it could be made to scale without changing the base principle. The TLS authentication can not be repaired without changing it from the core.
There is no SSH authentication "system". It's purely manual. How could it scale for use by billions of people who know nothing about security and are more than happy to click "OK" just to get that annoying dialog out of the way? Particularly without forcing far-reaching changes in the rest of the web infrastructure.
Also, I disagree that TLS' authentication architecture is broken beyond repair, for two reasons. First, in actual practice for the vast majority of uses, it's not broken at all. Given the scope
Re: (Score:1)
Re: (Score:2)
You didn't propose anything at all. You just complained about TLS and said SSH is great, not even substantiating either of those.
The reason for my lengthy posts is to engage in a technical conversation, but you seem incapable of doing anything but whining. If you'd like to actually respond to my points, perhaps propose how an SSH-style scalable system would work, or explain why you think the proposed fixes for the CA system (including the others I didn't mention... surely you're well aware of them), I'll
Re: (Score:2)
I'm seeing a few misconceptions in what you've said here as well as in your subsequent posts in these threads, so I hope you'll pardon the upcoming car analogies in response to this sentence:
I feel much safer knowing my SSH sessions are not HTTPS-encrypted, because the certification mechanism is completely broken.
To me, that reads like:
I feel safer knowing my Newegg packages are not Amazon-shipped, because the gasoline industry is completely broken.
To break that down a bit...
1) HTTPS is not an "encryption scheme" any more than Amazon is a courier service. Suggesting you can "HTTPS-encrypt" something makes about as much sense as saying that you can "Amazon-ship" a random package. Which is to say, just as Amazon relies on FedEx, UPS, and other
FCC to ban Open Source Routers Soon... (Score:1)
The FCC has an open rulemaking proceeding that would expand these requirements beyond the 5 GHz U-NII devices covered by the OET document to all Part 15 devices. See paragraphs 45 and 46 on page 18 of the Notice of Proposed Rulemaking (FCC 15-92):
We propose to modify the SDR-related requirements in Part 2 of our rules
based in part on the current Commission practices regarding software
configuration control. To minimize the potential for unauthorized
Re: (Score:1)
We should fill that public comments section up with traditional Slashdot posts. These should even include GNAA, Goatse, and many others calling them 'fags' or whatnot. It would be priceless.