11
Mohan much prefers an edge computing architecture that uses many low-cost power
efficient wireless sensors that send data back to a single central processing unit located
somewhere in the building. "This is like a thin client system," he says. "You have one edge
gateway or compute box in a corner of the building with an internet connection. It talks
wirelessly to all the sensors." Mohan points out that today's edge computing devices
can run anything you would run in the cloud, including data processing and learning
algorithms for smart control functions.
In addition to using lower-cost sensors that are easier to install and require less
maintenance, a big advantage of this architecture is how it supports fast, easy system
upgrades. Mohan says, "If your compute box upgrades from 4 gigabytes (GB) of RAM
to 8 GB of RAM, or it goes from a dual-core to a quad-core processor, you can literally
send the upgraded box to the customer to swap out. Then, you can run all the software
upgrades on the back end, which is much easier than having someone go through the
building and change all your sensors."
There is an additional advantage to using the thin client edge computing architecture:
security. If you have a system with multiple sensors connecting to the cloud, every sensor
becomes a hole in your network and part of the attack surface. With an array of sensors
connected to one edge computing device, only that one connected edge computing
device is part of the attack surface, which makes the system much easier to defend.
For a smart building control system to succeed, it must be intuitive, easy to manage,
easy to upgrade, and easy to operate. It must provide TCO advantages, and it must be
secure. Solution design plays a key role in achieving these goals.
"You can run
all the software
upgrades on
the back end,
which is much
easier than
having someone
go through the
building and
change all your
sensors."
