Weekend Ramble - A Rough Week For The Cloud And Internet Of Things

About a month ago Revolv announced it was ending support for their Revolv home automation hub, and this week a number of sites have picked up on this, starting a discussion of just what this means for cloud computing and the Internet of Things (IoT). What has a lot of people concerned is that, unlike many devices, the Revolv is COMPLETELY dependent on the Revolv service to do ANYTHING. Without that service, users will on May 15th have a brick, and a non-functional home automation system. And we’re in a world where things we have purchased can stop working with little or no warning.

Revolv was bought by Nest back in 2014, apparently more for their expertise than their product line. The Revolv hub ceased being offered for sale at that time, although the service the devices depend on (more about that shortly) was still active. Which leads to the first interesting point, no Revolv hubs still in use are under warrantee. I’m assuming the Revolv service was free, so one question is, does a company have any obligation to keep a service running that their products are dependent on, absent any contractual agreements? In true IANAL spirit I think it’s unlikely. Certainly a number of hosted, or “cloud”, services have gone poof over the years (I was a fan and user of Google Reader for quite a while). Usually when that happens there are alternatives one can move to, at some level of inconvenience, but this is one of the first cases I’ve heard of where a service retirement completely renders a physical product completely useless. Even my 3Com Audrey still can display a web page (kinda), and I hear WebTV boxes can be repurposed to a degree also. But it looks like the Revolv is destined to become a hockey puck. Even worse, from the comments I’ve read it’s not clear if owners realized how dependent the Revolv was on the Revolv service.

While there may be no obligation to do anything to keep a product discontinued 2 years ago running, there are still questions and concerns raised by this incident that should be addressed. Why, for example, is a device like this so dependent on a cloud service? Sure, some sort of hosted service would be needed to allow for remote control and monitoring of the system, but if I just want devices in my home to talk to each other and be controllable in the house, why the outside dependency? It doesn’t sound like Revolv was making any income on the cloud service, so milking the customer isn’t it. Is it just bad design decisions?

As the IoT, and other cloud services, continue to grow, what happened with Revolv should be an object lesson. How do we avoid this in the future? Use of standard, open protocols would be a good start, as would better system architecture. The IoT is supposed to make devices smarter and more flexible, not allow them to be broken at a vendor’s whim. Perhaps expansion of warrantee terms to cover reasonable expectations of service, and what works when that service is gone, as well as traditional coverage for defects would help as well. I don’t think it’s reasonable to expect vendors to run services for devices long after they reach end of sales, but is it really ok to just brick the device? Surely we can do better.

Written on April 9, 2016