Did it all start with the apple iPod? The perfectly joined up user experience with the cool device, allowing a one-stop shop to evolve into a global cloud platform: iTunes. Where does this leave independent software makers offering device solutions? Will analytic software makers get marginalised?
A lot of companies who are selling hardware are providing added value to their products by bundling or offering specifically device-engineered software. It seems like a perfect match.
Samsung phones and tablets are married to Google Android, but Samsung provides a lot of other devices where it created the software for TV's, video cameras, cameras etc soon it could be using just it's own OS (Tizen), joining all devices.
Car manufacturers, like Tesla, BMW, Toyota run software versions and analytics inside the car dashboard, which can be updated OTA (over the air) or via USB. General Electric's (GE) Predix has it's own IoT platform to maintain all their infrastructure, claiming to save billions of dollars in operating costs. There are lots more companies providing software working perfectly together with their hardware products pushing 3rd party providers out of this space, effectively closing their platforms not allowing competition.
From the other side we have Apple expanding more into the hardware space, cars, watches, headphones (Beats) and for a long time PCs and laptops. Joining Apple in this approach are traditional software makers, like Microsoft (mobile phones, tablets, games console), Oracle (Sun servers) etc.
Who will win?
In my opinion the balance is more in favour of hardware manufacturers expanding into the software business because it is not seen to be outside their core market. Without the software the device would be useless. Whereas traditional software providers are pushed by their investors/shareholders to not dilute their efforts going after hardware products as they traditionally have lower margins, but instead focus their efforts on new innovations and increasing their market share - aka the way of the #unicorn.
The same applies even for smaller companies and start-ups developing new devices and enabling them to connect to the internet. Everyone has an "app" which is built in-house and you can install on your phone and control the device with, may it be a speaker system, drones, TV, heating system etc
Where does this leave analytics?
This calls for more openness and more partnerships between hardware manufacturers and software makers, essentially OEM agreements. If the software industry doesn't, the hardware side will have no other option as to develop their own software (and take more of the profits). BIRST seems to be leading the wayaccording to Gartner in terms of offering embedded analytics (OEM):
Opensource software is now considered the preferred way of getting "OEM" software onto a device for hardware manufacturers, Pentaho was bought by Hitachi Data Services (storage solutions) and Jaspersoft by TIBCO both offering strong OEM analytics approaches. This will likely allow Hitachi to integrate it's analytics quite easily into lots of places. The list of acquisitions will not stop there while hardware is invading the analytics software space.
While this is all happening the big fight for the IoT platform space has still to happen, Google is developing and placing it's bets on Brillo and Weave, getting into it early. Time for the SAPs, IBMS and Information Builders of this world to start pushing the embedded analytics more and get out there forming partnerships!
Here an article I wrote earlier, related to analytics embedded into the cloud: https://www.linkedin.com/pulse/embedded-analytics-reborn-cloud-denis-sproten
What are the options?
There are two main approaches for IoT analytics solutions: centralised and decentralised.
- Source all the data from all devices into a Big data warehouse
- Analytics displaying on the device
Datawarehouse analytics already exist, but a so called micro-analytics engine being able to run on device doesn't yet. It may actually be more efficient for most devices to have a bluetooth connector so you can link it up to an iPad and send a query to the device. Or alternatively collect all the data from the device and then run the query faster. In which case we are back to hardware manufacturers releasing their app for the iPad or Android tablets. So it seems we are destined to go this way, where no-one has gone before.