Earlier this week, a Nest outage lasted for 17 hours. Nest cameras didn’t capture any video footage during that time. This downtime was likely a minor inconvenience, in most cases, if it was noticed at all. But for others who experienced some type of incident during the 17-hour window and that video footage would have been valuable to have, it’s a complete fail.
Yep. And my elderly father fell. Only the two times I needed it history was deleted.
— Proud Knights Fan (@3600dollarsgone) February 25, 2020
I’m using the word “fail” for a specific reason. As smart home systems mature and gain more mainstream acceptance, the failure of a cloud-based device or service becomes less acceptable. One possible solution is to start engineering these devices with some type of local failover, even if it’s limited in function.
Google says that the Nest outage was caused by a server update that didn’t go as planned. Having managed servers in Fortune 100 companies, I get that. And I’m not specifically calling Nest out here. Amazon Echo devices have occasionally experienced similar outages as have Ring products, which are part of the Amazon family.
Just in the past 30 days Ring device owners have experienced some service disruptions, as noted by Ring’s outage history page:
And last week, some owners of the PetNet smart feeding system saw their pets go hungry due to a service disruption with one person saying “My cat starved for over a week,” in a Twitter response to PetNet support.
The point here is that people are fully reliant on these types of smart home products to work. Not most of the time, but all of the time. When a supporting cloud service (often paid for in subscription fees) does go down, it can have very negative implications.
So what’s the answer?
We need smart home companies to deliver on the promises of local controls for existing products, and we need new products designed to smartly failover in some local capacity.
Last year, Google and Amazon both announced more localized services and smarts at the edge. Yet we haven’t seen much progress on this front. If new localized controls and smarts have found their way to our smart homes, I haven’t seen either company make a big news splash about it.
When it comes to new products, most of the ones I’ve seen are still focused on the subscription revenue model, which generally means some sort of cloud service for integrations or video storage. I don’t bemoan companies making money from services, but a local failover of some kind would improve the customer experience and, therefore, could sell more products and services in the long run.
Take the example of cloud-connected cameras and video doorbells, which are both a hot category right now. Having them solely dependent on a web connection to some servers is a recipe for disaster. Sure, they need the cloud in many cases for person recognition, data storage, or other services, but they’re IP-based devices on a home network, too. Before the smart homes of today, we had IP-based cameras that we could view in real-time from a phone.
Why can’t today’s smart devices failover to some localized viewing mode and rudimentary notification system during an outage? And if you’re going to add that, why not a limited amount of on-device storage, or a storage expansion slot for times like that?
Yes, there’s cost involved to add such storage or slots for a memory card. But as Wyze has proved with its $20 WyzeCam, it can’t be that much money. I have my own 32 GB memory card in my WyzeCam for this very purpose.
Something’s got to give here because the smart home is increasingly being relied upon by millions to monitor, react to and inform us of changes in the roof over our head. Server outages are a question of when, not if, even for the best of companies that have large-scale redundancy. It’s time for smart device makers to consider building in local failover options for when the inevitable system outage occurs.