Wireless-First Strategy,

what to think about?


(TL;DR) Have the correct data-strategy, think long-term and bi-directionality, understand that different technologies do different things. Think twice about using any long-range technology in buildings (like LoRa) – My advice is using Wireless Mesh.

I am advising clients all over the world on what their Connectivity-Strategy should look like in order to get the most bang for the buck. It could be Digital-twin companies needing data from buildings and existing systems. It’s General Contractors that want to be the next APPLE of the building automation industry. New entrants to the smart building industry wanting the know-how what the perfect edge gateway should look like. It’s Real Estate owners, wanting to see more of a return from buildings and to challenge the status quo.

It’s all about trying to offer the pros and cons with any strategy, choosing the technology based on what the customer wants to do right now, but also in the future.

  • How important is the total cost of ownership, flexibility, security, ease of install, recurring costs, how often to send the data, to where, how much data, battery-powered, location, firmware updates, usability, scalability and a lot of other factors?

And that is why I’m always interested in knowing more about the existing alternatives, weighing alternatives back and forth, and recommending what I think should be the best option. I need to know this. Partly because it’s my job, but most importantly that the future revolves around solutions that can pass the test of time as much as possible. The world depends on it.

The connectivity situation in buildings
Looking at the future projections, it seems that the volumes in the long-run (2025 the earliest) between licensed and unlicensed solutions are 50/50 but in totally different use cases, thus more complementary.

“Licensed LPWANs run on public cellular networks that use the licensed radio spectrum and support the GSM and 3GPP standards. … Unlicensed LPWANs, on the other hand, use the radio spectrum that is unlicensed and can be used by anyone without exclusivity.”

I wrote about it in my last article about “What Connectivity Horse to bet on” after an interesting message conversation with a leading expert in the field (The comment section is particularly interesting since it offers alternative views than my own).

However, for this piece, I am ONLY interested in what solution would be great for buildings. Indoors. It’s about the need right now, and also what future needs look like and how that resonates with the choice of technology.

Existing and future building needs are these

  • Local Control First
  • High throughput data for local smart grid inclusivity
  • Buildings as node-balancers in local energy communities
  • Bi-directional control in order to create automated buildings
  • Deploying sensors that can scale to thousands of sensors
  • Interoperable solutions with as little complexity as possible
  • Battery-powered sensors/actuators needed
  • No packet losses (AI/ML demands quality data; otherwise predictions will be corrupted)
  • Native to Existing building automation systems protocols (BACnet most likely)
  • Secure communication

All the above is the necessity if you want to deploy anything that is future proof. Monitoring purposes only is fine for now and for slow-moving data in small packets, but buildings can’t be run with 20-minute data in the future.

Slow-moving floppy-disk size technology could be used when detecting water leaks in buildings or change of value in temperature just for the insight of knowing something.

But the future of Smart Buildings and Smart Cities demand real-time data to a much greater extent than what a lot of these solutions can provide. Or it conflicts with the Primary law in building automation:

Local Control, First.

This removes the traditional approaches of LoRa/LoRaWans, the NB-IoTs, The CAT-Ms and everything else that needs to go to some cell-tower or some cloud in order to get the data back to the building for control. Because what happens if the connection goes down if any logic resides in the cloud and not at the edge? The building will break. And possibly the people in there. LoRa Edge sounds interesting, to be honest, but it will still take a while to get to widespread adoption.

“Edge computing is computing that’s done at or near the source of the data, instead of relying on the cloud at one of a dozen data centers to do all the work. It doesn’t mean the cloud will disappear. It means the cloud is coming to you.”

All buildings should be able to operate freely where the cloud is just for offloading and pushing down algorithms when necessary. Everything that happens in the building will be sent to the cloud and to be augmented in a portfolio-way. Very similar to that of TESLAs OTA (Over The Air) architecture in that the OS could be updated at all times. But the logic still resides inside the car (read building).

It’s all about reducing Total Cost of Ownership

And that’s why I get a bit annoyed when I see that these technologies that mostly focus on monitoring are finding its way into buildings. Not only because of security risks with LoRa, but mostly because it’s not a strategy that is built for the end with that kind of technology.

Think about it. It’s like knowing that you want hot water running fast next year in a streaming kind of way, but you only install a faucet allowing you 1 decilitre of cold water every 10 minutes. Would your needs change in the future? Do you need more now?

If you need more, you soon find yourself in a kitchen with 5 different faucets, leading to an additional IoT aspect of the infamous “Thousand cuts-problem.” Not only with integration to the other faucets and other applications, but also the cost of install for each faucet and the total cost of ownership when packets are lost, and sensors don’t report data. Because what happens is that companies need to send someone out to fix it, and things will get messy in no-time.

That’s one faucet (sensor). What about having thousands of these already deployed, and the only thing they can do is to give you cold water once every 10 minutes?

I’d rather install thousands of faucets that can deliver hot and cold water whenever I want it, and it’s up to me to decide how to use it for the next 10 years when I upgrade to the next faucet in the same product line.

Your own Digital World with Wireless Mesh

The only thing that I see can cut it in the realm of building automation, and what will pass the test of time is Next Generation Wireless Mesh. It’s like building up your own private digital world with powered routers the size of a mobile dongle, allowing you to deploy battery-powered sensors at will adding, or subtracting sensors that will auto-configure themselves.

These together build up a future proof tent, where sensors can be deployed in the thousands, auto-configurable, no packet loss, one API, one up-front cost, extremely low Total Cost of Ownership, super reliable and an architecture that allows real-time data over battery-powered sensors. Data will be generated at the edge, and this allows companies the power to choose.

If you just want monitoring to start with, go ahead. Run that cold water with 1dl every 10 minutes or every hour. But when there’s a need for more data, that infrastructure will break, and the 20-year battery-life will soon become 5 years, 2 years, 1 year, and you’ll scrap the entire thing because your data-strategy wasn’t matched with the architecture of choice, to begin with. And that’s the thing. It’s about understanding the existing and future use cases.

Using wireless-mesh in buildings will offer comfort knowing that, when customers want bi-directional when data-scientists and AI/ML-powered solutions cry for more real-time data, you have the solution for it already in place. One API, delivering all that data you want and that works seamlessly together with existing and future solutions.

There are a variety of wireless-mesh solutions out there, and one thing to be wary of is the routing tables, which can limit scalability a lot. Home automation stuff like Zigbee-mesh and other things like it won’t cut it in scale, nor for future data quality purposes with an error-prone architecture.

There are maybe a handful of solutions out there right now that are fit for purpose when it comes to smart buildings/smart factories.

Summary

It always depends, yes. It is vital to understand that the short-term and long-term business strategy and the data strategy are linked together.

But when it comes to data collection and control for building automation purposes of today, and tomorrow- Local Control First – is the way to go. Everything else is plain stupidity and possible power-moves from companies with hidden agendas and/or a lack of understanding of what physics allows.

Cloud companies, what do they want? Data. Licensed and unlicensed companies like Telcos and LoRa-companies, what do they want? Cloud and recurring revenue to help you operate the network. The security aspect is also an important factor where your data strategy could go hand in hand with Private-Cloud thinking and on the Edge processing and control. Why send the data to someone else to spy at?

In my opinion, edge and distributed intelligence are here to stay considering that the technology in, i.e. our smartphones are getting smarter, faster, and better and better. Put 10 of these together, and you’ll get your own private cloud with extreme calculating capabilities.

5G, Licensed and unlicensed, Network-slicing, will be phenomenal in offering applications at the top that can leverage data in buildings and provide stellar user interaction, also from a portfolio perspective. But the transport-layer and logic will reside at the edge whenever it comes to building automation purposes.

Future-proof, as it should be.

Get in touch with me if you agree/disagree with this, and if you have any great use cases of wireless of your own!

Originally posted on Linkedin as What connectivity option to choose?

Also related topics:

Smart City Future
The building automation industry is broken

Leave a Comment

    This will close in 20 seconds