Internet Of Things: towards Smart Object Storage (Contextual NoSQL)

Just commented on a Tech Republic blog post, and thought it was worth developing the idea.

Many people see the Internet Of Things just like a pure buzz, a word with no other meanings or stakes than pure marketing. That something we can’t agree with, there are very specific challenges behind “IOT”. IOT is about cloud. Cloud is about storage – and computing -, and storage needs to become smarter, more contextual. Here’s why.

NoSQL was not invented because geeks were feed up with SQL. NoSQL is there because of its massive scalability and resilience, and breakthrough performance for cloud-scale applications and platforms. That’s not a way of life, that’s not a mood: it’s just a constraint we have to live with in the cloud world. Same for Object Storage, which is the step further.

 Let’s get back to Internet Of Things:

IOT is much more than just a marketing term. There’s internet, so internet based or cloud powered apps, and there’s things, so other kind of interactions.

Things are very different from humans, and require more sophisticated data processing. Things, Objects, however you call them, are dumb.

That’s the concept behind our Smart Object Storage YobiDrive Y3: taking into account at the cloud or internet level, however you want to call it, that Objects need a little help from their cloud friends. And they need instant help to push value added data to their subscribers. Connecting a temp sensor, I’m not interested in the room temp, but only mine, when the sensor is in my ear.

A mobile health sensor will need data processing, for example to clean up data before storing. A health diagnosis helper app will need prepared data, with archives, variances, trends.

With things, the mix of competences to build the full picture becomes way too complicated for a single company: you need electronics experts, data analysis and reformatting experts, and great app developers that deliver business value from this data.

That’s why we believe this can’t happen without having a new vision of cloud storage, allowing each of those competences to merge around a “Smart Object” that represent this dumb thing that you turn into a smart object by plugin in competences from everywhere.

After all, databases from the 80s (and before) had triggers, and demonstrated a real efficiency. Why should could storage be inert ? NoSQL has to become smarter.

Great apps can be built around sensors:

The case of the WiFi pulse oxymeter sensor such as built by Sensaris is a great one. Let’s add change actions upon data push such as data cleaning, automatic day per day archival, csv auto build… what do you get? A mHealth ready platform where you just have to plug a visualisation app, merging the best of electronics, data processing and UI competences.

Smart Object Storage also applies to basic cloud apps for humans…

Let’s configure Y3 and it’s API for 2 kind of objects, typically manipulated by humans: folders and files. Let’s add them metadata, roles, states and state change actions such as reader, writer, manager, publication state, invitation email sending, pki encryption configuration, … you get a full featured secure file sharing API on top of which you can build an APP such as YobiDrive FX with a development effort small enough to be an innovation in itself.

 

By adding a contextual behavior at the storage level, it becomes easier to plug applications from various sources that produce and consume that “smart” information, triggering object related intelligence.

Comments are closed.