I recently received a "terms of service" update from Comcast, with the notification that they can now "monitor and record anything going through the network. Including, but not limited to: audio recording, video recording, ..."
for anyone interested I pulled up the Terms and found what I believe the OP is referencing:
`Monitoring and Recording. You agree that Comcast and its agents may monitor and record any telephone calls or other voice, data or image communications that are transmitted between: (1) Comcast and its agents and (2) you, your agents, any user of your Service(s) or Equipment, or any user of any phone numbers associated with your account.`
This is far broader. No where in the Subscription Agreement[0] is the scope or the uses defined or limited. From what's written here, any form of information sent to Comcast in any way will can be stored and monitored for any purposes.
IANAL, but this seems like clever lawyering to make consumers think they're only referring to customer service calls.
Well doesn't that just mean they're allowed to record the conversations between me and their tech support? How is that worrisome? I assumed the person you were replying to was asserting that Comcast had proclaimed its right to record all communications I had over the internet.
I don't know. I'm reading it differently. I'm reading it in a way that almost any other entity with the power to execute on a term like that probably would - liberally and with deference to their own interpretation of lawful. Not too mention the 100% murky scope of requests and behests of government.
I read that the wire is owned by Comcast. Its agents roam free and everything traversing that wire is monitored. 100%.
If you're going to read the language non-sensically, why even bother to point to the text? Why not just make up whatever you want?
The numbering and the "and" clearly mean that one of the two endpoints must be "Comcast and its agents." The language is there to allow them to record customer service calls because otherwise in some states that would be a violation of wiretap consent laws.
I think this language could allow for Comcast to monitor any traffic passing through their proxies?
I agree that it seems to be worded to imply that it's just CS monitoring but I don't believe that `transmitted between` would necessitate that the party be the intended endpoint.
Here's mine in New Zealand (Slingshot), which sounds a bit more fair
"We do not proactively monitor what content you download or access, however, we must act on lawful requests for information and/or interception as well as infringement notices which we receive under the Copyright (Infringing File Sharing) Amendment Act 2011. This action may include sending you an infringement notice"
You have to get this stuff on record in real-time so the adminstration after Pai's will have timestamped evidence to act on. Stop being so short-sighted and defeatist about this stuff, think strategically.
Well, yeah. But on the other hand, the complaint records aren't going to be deleted even under Pai.
The complaints likely won't go anywhere anytime soon. But when Pai's successor is working to repair the damage, there's a decent chance that one of the things they'll have to do is go through the agency's records to figure out what was ignored. Having a record of the complaint will at least give them options in the future.
So file the complaint, even if it's unlikely to matter in the short-term. :)
Federal record-keeping rules are pretty stringent. Generally speaking, you can't delete stuff. Especially important records. And public comments are a required part of federal rulemaking procedures--see 5 U.S. Code § 553(c) [0]--so they'd definitely fall into the "important" category.
In fact, the FCC's response to the bot activity was to point out that they aren't permitted to delete the comments, though a former FCC special counsel was quoted as suggesting that the FCC "might have an obligation under the Administrative Procedure Act to remove fake comments from its consideration."[1] But "removing fake comments from its consideration" isn't the same as actually deleting them, so I'd imagine that just means labeling them as "likely fake" and ignoring them in their deliberations. The same would apply to official FCC complaints. The FCC might ignore them, but they can't outright delete them without violating the law.
Comments were collected. No obligation existed to treat them as a vote. Very few contained new information to the FCC, so they weren't valuable. The entire "millions of comments" nonsense is wholesale irrelevant, because the process doesn't exist for the purpose of voting or conveying public opinion.
Comments are collected in case there are scenarios the regulatory agency didn't consider. They considered them, and decided a way you didn't like.
I recently signed up for a promo from DirectTV Now where the promo was cheaper than the free AppleTV that came with it.
Don't worry, I'm not promoting DirectTV Now because it sucks balls. I have ad blocking at the router level at home and DirectTV basically won't work because of it. Even when it does work the picture quality is awful, you can't easily skip commercials, changing channels is painfully slow. And there are ads everywhere.
But, my main point is that a few weeks after getting it, I get a mailer from Charter asking me why would I want to pay for channel bundles (I get only internet from Charter)... the only way that mailer makes sense is if they were watching my traffic and seeing that I'm a subscriber to a channel package from their competitor. (I've never gotten a similar mailing and it doesn't make sense absent spying, otherwise they're kind of arguing against their own main cable service.)
It really made me want to get VPN setup whole-house.
I just received a similar mailer from Time Warner, and we use Sling. However, I think they are just trying to capture a market segment that they entirely missed the boat on, and that it is coincidental. I guess I could ask my neighbors if they received the same mailer.
What you can do is use a third party DNS resolver like 1.1.1.1, ideally over HTTPS or TLS, and use HTTPS for all your web traffic. That dramatically reduces what they can inspect.
If that isn’t enough, your next option is a VPN or Tor.
Third party DNS is at best a very minor hiccup in the ease of an ISP monitoring where you're going, though in combination with things like HTTPS Everywhere it's a start - mostly if you're going to good-sized sites where most of your traffic actually goes to CDN providers.
Hm, thinking about it as I write, I could see how encrypted DNS plus everything being encrypted and served via CDN could actually cut down a lot on what carriers can see. Still far from perfect, but not quite as bad as I was originally thinking.
Edit: I wasn't familiar with Server Name Indication (destination hostname is unencrypted even though the rest of the URL and session are encrypted).
Without encryption, they can see who you talk to, what you are saying, the rate, the frequency, from where and when.
With encryption (https), they can see who you talk to, the rate, the frequency, from where and when. They can't see the actual URL (just the hostname) or data (encrypted).
With VPN, they see you are talking to a VPN, the rate, the frequency, from where and when. VPNs cut down on knowing who you are talking to (assuming they aren't logging or being monitored which they could easily do).
Other services could be added to obfuscate rate, frequency and when I would assume, but even then those services would only be additive obfuscation (unless you cache packets for a short term... just thinking as I type).
Someone check me if i'm off on this these points...
You're close. There's actually well understood technology for obfuscating the rate and frequency - the field for manipulating these variables is called "traffic analysis".
One simple technique is to always transmit X packets/sec where Y packets/sec are real and the other packets are dummy packets (Y < X). If the channel is encrypted, it's impossible to distinguish the dummy traffic from the real traffic, and if you're over a VPN, it's difficult to identify the destination.
A slightly more sophisticated approach is to vary X over time, to make it shaped like streaming video, for example, to obfuscate the fact that you're using traffic analysis countermeasures.
It's not one of encryption or tunneling or packet shaping that does the work, but rather all of these techniques together work to reduce an adversary's traffic analysis capability.
One of the problems I've run into with using a vpn service is that many services block IPs from common vpn services. While this may be due to abuse originating from those IPs this still seems like a lazy approach to abuse prevention.
One approach that could work around this in many cases is to run a VPS with a private VPN server on a cloud provider. This is beyond the technical ability of the average user though and costs more than most VPN services.
In practice this isn't much of a problem, and I get it with Netflix. I turn off my VPN if I am watching Netflix on my computer and I don't think my privacy suffers too much for it. It feels bad not to have 100% coverage but I honestly don't care who knows what I watch / how often I watch Netflix.
It is more of a problem if you want to run the VPN on your router upstream of your devices, which simplifies management and ensures all devices are protected from ISP snooping, including devices where I don't control the software like a Chromecast.
I had to disable router-level vpn for exactly this reason, which is frustrating.
Amusingly enough, when I did have it on, my Chromecast showed weather data for the vpn endpoint so it's using IP-based geolocation for weather. It could be smarter.
Depends on the traffic from devices. If you don't have crazy high traffic, flash your router with openwrt and install wireguard. If the flash is too small, you can compile your own image with wireguard selected which doesn't need as much flash space.
Next you can choose a cloud provider, which is metered (GCE, AWS, etc), or non-metered like OVH, or Digital Ocean(they don't charge you if you go over the 1TB for now).
Or you can choose a VPN service provider like Mullvad who have wireguard option (PIA should be getting it soon), if you trust them.
Be careful in choosing your provider since you might be annoyed with the latency over time, or just get used to it.
You can mitigate the speed issue by adding your own DNS local caches and then start blocking ads, trackers and malware sites and whomever you want (FB is added to the spyware list) using publicly available lists. All of this speeds browsing considerably.
Openwrt by default caches using dnsmasq, but the blocklist is a good idea. The problem is, even the blocklist is sometimes too big for the tiny flashes of some routers.
As another poster said, go with a mini-pc like Qotom. Uses little power, configure however you want (memory, SSD). Most web pages load instantly, and it handles a massive blocklist (Bind9). I'm slowly adding a list of always on packages like sync tools. Also, you can use an AP instead of router attached wifi. Move the Power-over Ethernet AP where you want. Ubiquiti Unifi is far better coverage than my previous consumer grade wifi. Do it all in your favorite Linux flavor.
The average user doesn't want to buy more hardware, and flashing an existing router can be done in an evening.
Also, why Bind9? I don't see what's wrong with dnsmasq, and changing hosts file for blocklist. Also, I often advise against network wide blocklists unless you're the only one using the network, since subtle things break.
I find having one piece of low-power hardware that is always on a handy tool. A homeserver+router, basically. I can decloud a lot of things. Having a beefier piece of hardware makes it a non-issue. I try to run things in Docker for modularity. Total hardware cost is competitive with a high-end router, but I think I get more.
Bind9 seems to be better for blocking. RPZ is made for it. I don't think dnsmasq supports RPZ though projects like Pi-Hole use dnsmasq. I'm not positive, but I think RPZ is more flexible. Bind9 seems to do anything you like. I may want to resolve DNS myself and not just forward.
I'm starting to look into configuring Bind9 to have different blocking per user using "views." Some want Facebook, some don't, so I can block accordingly. I'm not sure you can do that in dnsmasq. I did discover subtle things break, like you can't block Facebook and still access Instagram, thus the "views" approach. I don't want to change hosts file on every device, especially mobiles, and can even provide some protection for guests this way. I might do a captive page for a blocked domain and let people bypass in their view if they like, then I can have a "block-first" approach.
I do like network-wide blocking for the malware lists - if anyone acquires malware, it can't phone home (if it's on the list) and I can detect via logs. DNS as firewall seems to be a trend. I'm looking into blocking IPs via iptables as well using public lists. Maybe I'll even setup Snort or Bro. The possibilities are endless.
Just clarifying your comment for others: Your ISP would still see the IPs of the sites your packets visit unless you are using an encrypted VPN connection. The DNS-over-HTTPS is a great addition to a VPN.
(the VPN comment was several levels up so some might miss it)
Netflix and other video sites will be blocked from a lot of VPS providers. You may be better off going with a professional service that can keep their ips clean
True, but it can be mitigated by setting the routing table for Netflix IPs to avoid the tunnel. There's no standard way this can be done, but it's not impossible.
I did this for a while, and it was a decent headache to maintain. I had pfSense running at the Comcast edge and configured it to route all but specifically whitelisted traffic through a VPN I controlled. Speed was slower than max but acceptable due to choosing a VPS provider close to me, but the maintenance on the whitelist was cumbersome.
"In order to server our customers better and provide the best possible experience, VPN services will be blocked and will require a Business Tier service. We feel that unless you have a legitimate business reason to anonymize your network traffic we will provide this service in order to protect our subscribers and network integrity. Click Here to Speak to our Sales Representative"
That can potentially be a HIPAA violation unless medical providers/platforms get a Business Associates Agreement with the ISP. This could be a nightmare.
I don't even have an alternative in my area.