Ask Slashdot: Security Monitoring Company That Accepts VPN Video Feeds?

mache writes: My cousin is finishing up a major remodel of his home in Houston and has installed video cameras for added security. At my suggestion, he wired up all the cameras to be on a separate VLAN that only uses wired Ethernet and has no WiFi access. Since the Houston police will only respond to security alarms if the monitoring company is viewing the crime in progress, he must arrange for the video feed to available to a security monitoring company. I told him that the feed should use VPN or some other encrypted tunneling technique as it travels the Internet to the monitoring company and we proceeded to try and find a company that supported those protocols. No one I have talked to understands the importance of securing a video feed and everyone so far blithely suggests that we just open a port on his home router. Its frustrating to see such willful ignorance about Internet security. Does anyone know of a security monitoring company that we can work with that has a clue?

Comments

IP matching

[TWX]

There is a degree of understanding for why a security company might not want to use your VPN solution; if they have to monitor a lot of customers' cameras then they'd have to have a lot of different VPN clients running that might cause problems when the networks overlap private IP addresses.

Configure your firewall to allow their IP address range to port-translate to the NVR's IP and port(s). ACL-off your security VLAN from your user VLAN(s), and vice-versa, and allow only the correct ports through from your user network(s) to the NVR.

Re:IP matching

[silas_moeckel]

Not at all, VPN does not mean NAT traversal, IPSec is perfectly capable of security on public IP to another hell that was a design goal.

Re:IP matching

[laird]

A big problem with "VPNs" is that there are a ton of incompatible, proprietary VPN tools out there, generally horrible and incompatible with each other, and/or expensive, so it's far from surprising that the security company doesn't want to deal with any of it.

If the goal of the OP is to make sure that nobody is watching his home's video other than the security company, I'd suggest using https streaming (which is what his system likely already does) with a certificate configured at the security company, verified by the sending side (in the house) so that can prove the stream is going to the security company. For extra credit, put a cert on the sending side, verified by the receiver.

Of course, you still have to trust the security company. But there's no way around that.

Re:IP matching

[silas_moeckel]

Thus why I said SSL with pined certs and went for IPSec site to site also with certs since that is a standard and goes between vendors every day.

Re:IP matching

[mysidia]

If the goal of the OP is to make sure that nobody is watching his home's video other than the security company, I'd suggest using https streaming

The OP's cousin should probably just accept the risk that some unauthorized third party could in theory be watching the video data, Concentrate on making sure a third party can't Control or Disable the camera (What is really important!), understand that risk, and mitigate it by placing the cameras where they will meet security objectives without a huge risk to privacy objectives.

Legally speaking.... the OP's cousin will have already lost any legal expectation of privacy, the second they hired an outside company and shipped camera video out of their exclusive physical control, local law enforcement/FBI/etc can compel the disclosure of such video data in the hands of a 3rd party without needing a warrant or subpoena, and even bar the monitoring company from informing that footage has been provided; it's no different than requesting transaction records from a retailer.

The reality of the situation is; security monitoring companies are totally focused on physical security, the ones likely to offer affordable services are going to be local SMBs for the most part, and they are likely to have little knowledge of IT Security topics.

Unless you've found a highly exceptional security provider in the local area that promises privacy of your video footage, most companies will just not care , and might not fully understand these issues, the issue is so universally neglected, that you will need a contract requiring end-to-end encryption, AND the OP will have to pay the monitoring company enough money for them to care.

And then, unless you are paying a couple thousand a month, then it is probably a tall order to require a custom VPN solution "just to protect a home's ccTV feeds".

Can you lock the IP address?

[khasim]

If those companies want a port open on the router, can you lock the port to only the IP addresses that that company would be using?

That should be fairly standard on most of the firewall/routers available today.

Re:Can you lock the IP address?

[mwvdlee]

How well does that withstand IP spoofing?

Re:Can you lock the IP address?

[radiumsoup]

IP spoofing is only good in one direction, generally, seeing as how a sender using a falsified address would not be able to get ACK packets or other 2-way data back from the recipient (in this example, the IP camera system's video feed). In other words, spoofing is great for DOS attacks or injections of some sort, not so great for trying to view video. (Caveats for sending unauthenticated backdoor commands to alter firewall rules, etc., but that's an application layer problem, not an open port problem)

IPsec or simple ssh like tunneling

[yes-but-no]

VPN may be too heavy weight a solution. VPN is used when different sites [like branch offices of say a bank in a city] want to appear as though they are co-located in a single site. In this video surveillance use-case, it's just that you need to send the data one-way securely from point A to B. Just using an L7 secure TCP [like ssh tunneling] or using L3 IPsec like protocol should be sufficient. May be there are dedicated devices that do this.. or you may be able to run a script/software in the PC in the home which acts as a middle-man doing this tunneling and sending out of the data to the remote server. Of course the other end should be able to receive and do the necessary decryption.

Re:IPsec or simple ssh like tunneling

[marcansoft]

If the camera is HTTP, just reverse-proxy it with something like nginx into HTTPS, and let it handle basic HTTP authentication. HTTPS should be as secure as most VPNs in practice, and the authentication at the proxy level stops pre-authentication exploits against the camera. Now that Let's Encrypt is a thing you can even get a real cert easily. The security company doesn't have to know that you're doing this; you give them HTTPS URL and off they go.

Re:IPsec or simple ssh like tunneling

[kyubre]

Please pardon my ignorance and be gentle in rendering a /. style re-education, but isn't this precisely what https transports are supposed to do?

I set up a similar thing for some family members but its purpose is only to monitor and document any intrusions (they already have a commercial security system, that apparently doesn't work all that well). The cameras aggregate to a single machine within their lan and my (remote) server periodically scp's all the video/picture data from all the cameras to my location where it is further consolidated to a single video file per day, per camera.

I don't think its any more vulnerable than any other ssh connection using secure keys. It's been running flawlessly for 3 years, but if I had it to do over again, I'd probably stage the data on an https file server and skip the ssh stuff.

Re:IPsec or simple ssh like tunneling

[pi_rules]

don't think the OP gets a choice in the matter. Most IP cameras and IP camera system transmit the video feed over a UDP-based protocol; it's generally RTP/RTSP; and support for SRTP/SRTCP is sparse and far between (Unless you spent mucho more $$$ to purchase high-end equipment that specifically supports it!). The only way you're going to be wrapping RTCP in HTTPS is with a SSL VPN that supports encapsulating arbitrary TCP protocols and the UDP-based RTP streams as well.

Bluecherry (http://www.bluecherrydvr.com/) can wrap up RTSP into an HTTPS transport nicely... and I don't consider it all that expensive at $500 for a 32 camera license which is good as long as that version is supported. Hell, it can suck up MJPEG and coax cameras and rebroadcast them in HTTPS encapsulated RTSP too. It's not even a selling point of the product, just a smart decision that they made when designing it. Basically the Bluecherry server pulls up all the RTSP H.264 or MPEG, MJPEG, and coax camera feeds, transcodes them on the fly into H.264 for storage, and then rebroadcasts them over RTSP for their cross-platform client to display live. You can also suck that RTSP feed into VLC or whatever else you want.

And, perhaps nicest of all given the topic, Apache handles the HTTPS traffic, not their server, so anything possible in Apache is already at your fingertips.

The product has warts, but they designed the thing so logically that I can't help but love it, and their team bangs out stuff awful fast. The product is probably 300% better than it was in 2012 when I started using it and I didn't dislike it back then.

No, you completely misunderstand.

[Anonymous Coward]

No, he wants an encrypted tunnel to the security company. Not to a third-parrty proxy.

Re:No, you completely misunderstand.

[93 Escort Wagon]

Ah, got it. But then why not just find a security company that sends their camera video over https?

Re:No, you completely misunderstand.

[Anonymous Coward]

You stop now or start posting as A.C.

The video surveillance monitoring agency is not providing an authenticated service. They want the video to be hosted and available for their monitoring server to download. The very best an HTTPS connection can offer in the scenario you describe is security through obscurity.

silly topic

[Anonymous Coward]

You and your cousin need to get a life and stop worrying about highly optimizing the design of security systems that have almost no practical value. The reality is for most users, 99.9% of the security value of their system mostly comes from the visual appearance of the camera as a deterrent factor.

Really?

[ledow]

What's wrong with a port forward?

Get them to tell you THEIR static IP, and only apply port forwarding from their address to your internal VLAN.

Problem solved.

Have to do it all the time for telephony, CCTV, remote software support, etc. I let them have a port-forward but only if:

a) they give me their source IP (I get the asked the same when I set up VPN's etc. anyway, so everyone does this!)
b) they only get one set of port-fowards to the internal system
c) I reserve the right to cut that connection off for 99.9% of the time until they actually NEED to do something. They ring me up, I open up JUST THAT PORT to JUST THAT IP, then they have to tell me when they are finished.

It makes it much easier to manage, to log, and to control your devices.

Nobody sensible opens up any port to the world unless they have a public-facing service on that port and have secured it properly (e.g. email, web, vpn). But "port-forward" does not mean you let the world into it.

And if the attackers know and can spoof the IP of your remote support, then you're in bigger trouble anyway! That's not the kind of attacker that you're going to be able to easily defend against. But with a plain port-forward, all they'll get (if you've done it properly) is into the VLAN and the cameras, not your systems.

And, guess what. The only device that traverses several VLANs should really be your gateway anyway. There's no point VLANning off and then having everything sit on all the VLANs. So you might as well just have the gateway port-forward and then all the config is on one device.

(Not only that, VPN setup like you suggest is a pain in the arse for most people anyway. If you have a hundred customers, with a hundred VPN's, it quickly becomes stupendous to put them all on 24/7, because of IP subnets stomping over each other and all sorts of confusions. That's before you get into the million-and-one variations of VPN and VPN settings and managing certs and credentials).

Re:Really?

[Anonymous Coward]

I would find two issues with your proposed solution.

The first is that there is no way to know how they manage their IP addresses, which presents at least two problems. The first being, what if they do not manage the addresses themselves and some other company gets control of the IP address. This would mean the feed can be viewed by unauthorized users. The second issue is that they may use a range of IP addresses, adding new ones, dropping old ones, and in the case of an emergency, a new address used may not be have been updated locally, which is also a issue with the first scenario.

The second issue is that it does not address sending the video feed across the Internet in the clear. Do we even know if they can use an HTTPS connection? What if they only support insecure HTTPS configurations? What happens when a new vulnerability is discovered in HTTPS? Will the user promptly update their server configuration?

Get a provider in the commercial space

[guruevi]

You should have the Axis security suite or find one of their partners to install it for you, then some company might take you seriously. Once you get that contract, you can specify anything you want and pay accordingly. I've done IPSec lines for some of their customers, but you could be paying $10k/year easily to maintain a few camera recordings which are totally useless in actual protection or prosecution (unless the cops get extremely lucky with an extremely dumb criminal, they won't be looking for that one person or even recognize them when they get arrested on another charge).

But for home or small business, this is laughable, your camera's won't do anything, they will barely be able to see any silhouettes especially at night (unless you buy a $1000 camera, the 100' IR LED cameras all wash out the image due to reflection within the housing, and yes, I have tried a number of them). Your city doesn't require any camera for monitoring by police. You do need a permit and so does your alarm company. Perhaps your alarm company told you that but they are just trying to up sell you their camera system. https://www.houstonburglaralar...

You can do a DIY alarm system with a cheap alarm monitoring service for ~$500 (Honeywell Vista with a few sensors and remotes) and $5-15/month for the monitoring service (wired or wireless). You could hook up ZoneMinder into your Honeywell as well with an RPi or whatever, but make sure you understand the false alarm fees your city levies. Some city codes also require you to hook up at least one wired CO and smoke detector if you do get a system so you should calculate all that in, other codes require wired CO and smoke detectors on every level during renovations.

Re:So, in other words

[sumdumass]

VPN in modern slang times is generally used by people in one country trying to access restricted content in another (say copyright restrictions not allowing AU to view US shows on Hulu or something) or to obfuscate the original of the data being transmitted (dissident materiel or perhaps illegal to some extent material). The practical application of a VPN (secure tunneling access to a remote network like work access) seems to be forgotten by those using it for other reasons or the other reasons are more prevalent in certain circles that the reasoning doesn't flow as quickly.

I used to play a game online and people were constantly complaining about having to find a VPN. Turned out their ability to access the game came at a time there would be almost no one online in their country so they used a VPN to fake their location in an area where a lot of people would be online and had a better overall game experience. The game attempted to route you to servers in your time zone to prevent culture clashes and whatever which lead to a lot of boring sessions in odd hours evidently.

Re:This Is A Great Question!

[spire3661]

" camera or DVR should be accessible form teh internet and no video feed should traverse the internet without encryption"

Just NO. Not everything needs to be encrypted. In lots of cases it jsut adds bloat to whatever it is you are trying to do. I absolutely believe companies should make better devices, but in the end its up to US to use the vast tool chest we have available to us to make up for these shortcomings. Dont force obfuscation where its not needed so that you dont have to think.

Unsecure on alarm

[holophrastic]

I certainly understand the need to secure the video, fully encrypted, of my home. But I'd be willing to have it unencrypted, and fully open in fact, during a break-in. It's a big call for help for anyone looking, and it really ought not be that often. And anyone whe'd stage a robbery to see the footage as recon for next time, well, that sounds foolish.

So, while not perfect, why not switch to unencrypted during alarm scenarios?

You need ENCRYPTION not VPN

[Anonymous Coward]

Others have pointed some of these things out but let me spell it out in big letters.

OP started out by telling the security company "I want a VPN." He then came to /. to say to us "where can I find someone that will do a VPN."
The problem is that a VPN is the wrong tool. When you have a problem state the problem and let the /. world help you; don't state what you think the solution is and why nobody will do it. There's a good reason they won't -- it's the wrong answer.

VPNs are used to link separate private networks across a different (public, non-private, or other private) network. That's not what OP needs here.
What OP needs is end to end encryption to ensure the camera video is visible only to the security company -- not the Internet at large.

Some suggestions have been floated by other posters above me, and to summarize they are as follows. Note that the first by itself won't encrypt but any two of these together gurantee both AUTHENTICATION and ENCRYPTION, which is what OP wants.
- IP source address filter. If the connection doesn't come from the security monitoring company it doesn't allow the connection.
- HTTPS encryption with authentication
- IPsec tunneling

E

Build it ... Then ask someone to work around your

[BitZtream]

Custom worthless crap?

Bwhahaha ... No security company wants to deal with some jackass that thinks they know all about it but was too fucking stupid to think about how it might interoperate before he started and now he's shocked that people have no interest in dealing with him when he walks in the door telling they run their business wrong?

You guys are a joke. You got all wrapped in vlans and no wifi that you forgot that protecting your home was the point ... I'm not sure if that was actually the point or if you guys just wanted to waste a fuckton of money. Your security system was a waste, deal with it

I-View Now and Protection One offers exactly this

[Anonymous Coward]

A bit self-serving as the CTO of the company, but we provide this kind of service to commercial national account customers all of the time. Typically an IPSec VPN tunnel is established between the client site and I-View Now, and the DVR/NVR at the end of the tunnel is monitored for online status every 5 minutes (Which also helps keeps the tunnel alive). When an alarm is triggered, in under 5 seconds, the operator at the central station is viewing both a live feed from the camera associated with the zone that went into alarm, but also a 5-second pre-alarm clip of what actually tripped the alarm. This same video clip is delivered to the end users via a link sent in an SMS message so by the time they receive the call from the alarm company, they are seeing exactly what the operator is looking at as well. i-viewnow.com