The Zigbee Alliance this week offered — nine months after announcing it was forthcoming — an update on Project Connected Home over IP, an effort by Apple, Amazon, Google, Samsung, and others to create some kind of interoperability standard for the smart home. There are A LOT of cooks in this kitchen, with 145 separate companies and more than 1,300 individuals signed working on CHIP. It is a titanic effort.
The good news is that the standard is still on track, with a certification coming later this year and actual certified products anticipated in 2021. The Alliance also provided a list of devices that CHIP will focus on; it includes lighting, HVAC, locks, security, shades, and more, but neither smart speakers nor white goods, such as large appliances. And to round things out, it offered a GitHub repository and a detailed look at some of the levels where CHIP will seek standardization. But in many ways, the update was disappointing.
It was disappointing because CHIP represents such a huge shift in thinking about the smart home, one that could have incredibly important repercussions for the brands building connected devices. They want to know how things will work and what to expect, and they want to know as soon as possible because they still need to get products out the door. Even consumers aren’t sure if they should buy new gear or wait for CHIP.
In the meantime, developers are wondering how they should design their products and outfit their electronics to handle the inevitable updates. Indeed, based on my panel discussion during Silicon Labs’ Works With conference on Wednesday, I think most companies are participating in the CHIP groups so as to ensure their older devices aren’t rendered obsolete and that devices in development will continue to work with minimal changes.
But there are no guarantees. For example, none of the participants on the panel — which had representatives from Assa Abloy, IKEA, Philips Hue, and ADT — knew how CHIP might handle device provisioning or backward compatibility. So I thought it might be worth compiling a list of things we don’t know about CHIP. This list includes questions I had sent to the Zigbee Alliance all the way back in April and others I had sent ahead of the panel discussion with the hopes of getting an answer at the event.
The Alliance has answered a lot of the questions, especially those around timing and what the project plans to cover. But there are a few unknowns still out there.
So here are most of the questions I’ve posed to the group, and where possible, the answers I’ve been able to glean from the Github repository and various panels that have included CHIP members.
- When will developers see developer kits or resources so they can start playing with the standard and thinking about their next products in the context of CHIP being a reality? What will that process look like for developers and device manufacturers? When do you think we’ll have CHIP certified products for sale?
There is some real momentum here. During the Works With panel, Sujata Neidig, who is the head of marketing at NXP and represented Project CHIP, said we should see a draft by the end of this year and products in the market by this time next year. As for developer resources, Google showed off an example app using CHIP over Thread (it’s at 1:33 in the video), which was pulled from the Github repository. A spokeswoman for the Zigbee Alliance says that developers should look to their silicon partners for information about resources and dev boards.
- There are several pain points for consumers when trying to adopt the smart home, and to make smart home gear “secure, reliable and seamless to use,” what elements will the project address? Will the project address device onboarding and provisioning? Security? Naming best practices for voice interfaces? Rules around how devices communicate? Device decommissioning?
The more detailed stack diagram shared in the Github repository answers some of those questions. It’s clear that CHIP will have security elements, frameworks for the aforementioned device types, and what I think are pre-configured ways for devices to interact with each other (see the interaction model actions noted in the diagram). There will clearly be some rules on how devices communicate, but what those will be remains unclear. The panelists said they expect CHIP to outline a common way to get devices onto a network, but didn’t say how that might happen. During a demonstration by Google of a CHIP device getting added to a home network, the device was discovered and labeled (incorrectly in the demo) by a Google Hub Max and then the user entered a PIN code to add the device to the network.
- Will Project CHIP address how data is shared within device ecosystems? This can be personal data from the user but also state data from devices working within a home.
We don’t yet have an answer to that question.
- When it comes to provisioning, the project members have taken very different paths. Apple has embraced a QR code to get a device onto the local network while Amazon asks that users share their Wi-Fi data or hub data so when you purchase a device from Amazon it will automatically join your network. Are we looking at a model that’s more locally controlled or applied via the cloud?
That is another question that we don’t know the answer to yet.
- How will devices communicate with each other? Will devices have the ability to talk locally to other devices as many items already do? Or will Project CHIP require cloud-to-cloud communications?
Based on the stack diagram and the fact that the CHIP news this week included bridges as a device type that the group will focus on, local communication could still happen and is likely, but it’s unclear if CHIP plans to address cloud-to-cloud interactions.
- What is the role of a smart speaker/digital assistant from the CHIP perspective? Is this just another device type or is it more like a hub that will enforce rules and logic associated with the standard?
To be sure, smart speakers aren’t even a device type that the project lists as a focus area. Regardless, its perspective is important here, but it hasn’t offered that yet.
- How will developers bring their older devices into the CHIP ecosystem? Will they be able to update their existing gear, or will they need new hardware or specialized hardware? Obviously, some of this depends on the processing and memory requirements of running CHIP, but the core of this question is whether it will be possible if someone has robust hardware already in the field, or if we’ll need some forthcoming physical security (or other) element that necessitates new gear.
Even on the panel, participants in the project didn’t have answers to those questions, although they obviously hoped their existing gear would still work with a software update as opposed to having to physically replace their devices.
- What about the app experience? Will consumers be able to download a single app and control all CHIP devices at some level?
We don’t have guidance around this yet. Notably, developers in the chat associated with the Works With event were also curious about this.
- If someone buys a CHIP certified device, will they have to make a choice to use one ecosystem (such as Alexa, Google, or Siri/HomeKit) or will that device work in a mixed hub/assistant environment?
- Is Project Connected Home over IP the final version or should we expect a rebranding at some point?
Samantha Fein Osborne, the director of global marketing at Samsung SmartThings who also heads up the Project CHIP marketing committee, told me “CHIP is an acronym and its a working name,” which leads me to think that we’ll see something different and more permanent for the standard and formal certification.
- If Project CHIP does succeed and gain market share, it stands to create confusion if it is managed by the Zigbee Alliance. Should we expect to see some sort of new Alliance or a rebrand of the Zigbee Alliance going forward?
Unclear, but as someone who is so bad at branding (I have two web sites and two brands associated with a media business run by one person), even I think a separation should happen at some point.
So, we actually know a lot more today than we did in December — or even back in April, when I last wrote about Project CHIP in depth. But there’s still plenty more to wonder about.