Hacker News Reader

Comment by supergeek133 | Hacker News Reader

I work for a residential IoT company.

The issue with remote access is you need a service in the middle to provide that remote access, and provide it securely. Otherwise you're just talking something in the home (BLE, etc).

If you don't have some type of service the average consumer will not be able to use it, people aren't going to setup NAS devices, or other servers in their house to collect video data. Even a wireless DVR system can be a pain (another box, likely another TV to see it, etc).

I know that this community is more on the side of the skepticism of collecting data, but keep in mind the total addressable market is mostly people who don't want to deal with these problems.

Now if you're talking about "if the cloud service is down my device doesn't work" then I completely agree, these devices need to have enough local functionality to operate/buffer without their service working for extended periods of time, or being removed.

View on HN

Ring already has a base station. Why doesn't that base station have a solid state drive to store all the video/images, and allow the user to view them on their smart phone by connecting to the base station through the wifi?

The argument "you need to upload all your data to our servers or you'll likely have to get another to TV to view it" is just beyond lame.

> Why doesn't that base station have a solid state drive to store all the video/images

Everything is cheaper per pound in a data center. And they can force updates if the client is dumber.

And the customers won't complain when the SSD eventually dies.

> and allow the user to view them on their smart phone by connecting to the base station through the wifi?

Because people will want to do it across WAN. LAN-only is one piece of code, WAN-only is another. Going up to WAN for everything is redundant but costs less dev time.

Yes, it's frustrating for us who care about privacy. The motive is never to build a good product that respects people and does what they want. The motive is always to make money. Every corner must be cut.

> Everything is cheaper per pound in a data center....Going up to WAN for everything is redundant but costs less dev time.

Yes, agreed. That's just a totally different argument than the one supergeek133 made, so my point stands.

Yes my argument was business and ease for the customer, not what this is.

I suspect the answer is not a technical one, but a business one. I'm guessing that adding local-only usage features just aren't profitable enough (not enough increase in market share by implementing them) for Ring to justify the effort.

Running a platform costs money, many IoT products turn cost negative quickly if you don't introduce some type of value added service/ongoing revenue.

That being said I've always been a huge proponent of allowing some type of local control. Think of people in the automation market (Control 4, CEDIA type stuff) where it would simply be faster.

Even Google/Amazon allow local lambda function type activities that would be much better experience if we allowed local control.

But that also changes your security landscape.

Everything else that the other people listed (cost model, pricing, etc.) PLUS, if it is a box, you can follow the cables or look where the TV/router is, find the box, take it with you, and the evidence is gone.

Like this, even if the perpetrator takes the camera/bell with him/her, the evidence stays on the cloud.

Also, for evidence to be admissible in a court of law, one has to prove the chain of custody. Something in your laptop is more prone to editing than a video sitting on a cloud where logs vouch for its integrity.

Edit: don't let me invoke the gods of security frameworks. My laptop at home is not ISO27001 compliant, has never been assessed for security etc. The infrastructure of Ring (I believe) is regularly assessed/audited.

My question was rhetorical. As I said elsewhere, I was criticizing supergeek133's particular justification, not actually looking for the real reason.

The ""you need to upload all your data to our servers or you'll likely have to get another to TV to view it"" argument was involving the typical 8-camera wired DVR systems you commonly see in restaurants and in some homes.

It needs to be connected to a TV to view the video locally, or you need to have some technical knowledge to setup RTSP on your LAN.

Again, total addressable market.

Ok, but then how is that relevant to the topic? Folks in this thread are complaining that Ring wants to gobble up all the user data for services that could be offered without a cloud connection (including remote video watching). The reasons, as far as I can tell, are that (1) it's a bit cheaper through the cloud because of economies of scale and (2) Ring makes money off the data. It is not because it unavoidably requires technical sophistication/effort by the user.

Do you disagree?

The issue with remote access is you need a service in the middle to provide that remote access, and provide it securely. Otherwise you're just talking something in the home (BLE, etc).

Does Apple use iCloud for HomeKit, or is it all on the iPad/AppleTV that acts as the home hub?

I've had two HomeKit cameras at home for a while now. HomeKit cameras default to stream only (via the homebridge) while any of the people who have access to the camera are home. When everyone's out, it records/streams to iCloud, and this part is purportedly end to end encrypted, so that only devices of people who've been granted access to the camera(s) can watch it.

Homekit devices also require a mFI chip to participate in HomeKit (or develop software to connect to another device in the home with mFI chip).

> The issue with remote access is you need a service in the middle to provide that remote access, and provide it securely.

This is a pain, but solvable by punching holes in all the NAT pair combinations. At a previous startup we all communication peer to peer (phones, custom home devices, laptops), no data traversed our servers ("the cloud"). It worked beautifully.

> keep in mind the total addressable market is mostly people who don't want to deal with these problems.

> This is a pain, but solvable by punching holes in all the NAT pair combinations.

https://www.youtube.com/watch?v=i3RqDp7i308

I think I wasn't clear in the first post. The customer did not have to do anything (i.e. no need to configure any NAT passthroughs on their router etc), this was entirely transparent to the user.

I get that most people don't like to deal with that sort of set up. But we live in a world where the Raspberry Pi is extremely popular, yet also the most average consumer unfriendly computer in the entire world. Not everything needs to be a mass market product to be successful.

What's the sales of Raspberry Pi versus something like a connected camera (Nest cam, dropcam, etc).

"Popular" is relative.

Reply on news.ycombinator.com