Two well known outfits, SAP and HP, announced an IoT collaboration around SAP’s real-time database, Hana, and promise lots of edge computing firepower. We have been hearing about edges for so long that I think we’ve hit what I’ll call “peak edge” with this announcement. You would never want to underestimate the power of either company to make an impact in the IoT — their installed bases and channels are simpy too vast for that — but it seems like we’ve taken the conventional view of the edge —gateways absorbing tsunamis of data from dumb or pretty-dumb endpoints — about as far as it can go.
The next logical step for edge computing is to push this sort of work to the endpoint itself. When you think about it, the infatuation with the gateway as the “edge of the IoT” is pretty anachronistic when most of the non-IoT world is carrying or even wearing powerful endpoints every day. No one today would call your WiFi access point the edge of a broadband network — it’s your smartphone or your desktop terminal.
In the low power wireless IoT where I focus, re-defining edge computing is about organizing endpoints into what is effectively a NoSQL-like database. All endpoints are queryable. In real-time. Analytics at the endpoint. No more spitting useless data to a gateway only to see it discarded. Most low power IoT implementations don’t do this today, but they will.
The benefits of moving the edge to the endpoint vs. today’s “dumb terminal” status quo are myriad, but it has everything to do with choosing a networking stack that supports this. For more thoughts, check out how to Google IoT endpoints or divorce IoT endpoints from the cloud. Or click over to our website.
Create your free account to unlock your custom reading experience.