In my last blog I looked at how both:
 

  • Physics is going to get in the way for some decision making and you need to reflect that in your approach

and

  • Local embedded analytics is key, with decisions and data abstracted to the data lake to allow medium and long term decision making to be performed

These constraints, in my view, are driving, a new set of new capabilities needed from the underlying technologies:

  • Network edge based analytic decisions – not all decisions can be made by the overall system through centralized analytics – the system must have a level of edge based decision, storage and analytics built in, at as low a power as possible. The resulting decision must always backhauled to be aggregated and, where possible, the data itself batched as well to enrich the data lake. Local data fabrics and analytics must allow trusted decisions via rules and machine learning.

And

  • Data context aware intelligent networks – whether it is the industrial device, my flight across the Atlantic or my Nike device connected to my medical records, it is key that the data is context aware and intelligent. The data should follow me to my destination to provide the best response and, at point of usage – provide the lowest latency. Where analytics need to be performed locally, in needs to be abstracted with  the supporting data and be backhauled to the data lake.

What does this mean?
 
In my view, intelligent infrastructure has a long way to go – the big players like Cisco, EMC, Netapp, etc. have a real opportunity to innovate and drive new capability into these areas. There are differentiation plays for them with a huge market opportunity in the next 10 years for the players that get it right.
 
Exciting times.