Neo4j’s Emil Eifrem looks at why graph databases have enormous application to IoT

Graph is now a very familiar idea regarding its usage in capturing relationships and powering in social networks, with many of the social web giants such as Google, Facebook and LinkedIn having pioneered its application. 

However, there are also very promising applications of graph databases that are not so obvious. It’s the mapping networks of things – more specifically, sensors, devices and machine-to-machine relationships – that are the basic fuel of the complex IoT (Internet of Things) systems we want to build.

Smart homes will need to run a large number of sensors, networks, devices, cameras, power grids and smart water and thermostats to work, for example. But they will only really be effective if they are linked up together as a connected Internet (network) of many things (devices).

[easy-tweet tweet=”An IoT structure will have to underpin our future smart homes and cities.” hashtags=”IoT, AI”]

In other words, an Internet Of Things (IoT) structure will have to underpin our future smart homes and cities. When a new item of equipment or sensor comes online, it will want automatically to seek local controllers or other devices that it needs to listen to or share data with, while the powering up or down of just one individual sensor may create or end dozens of connections – maybe hundreds. Eventually, thousands.

The abstract that out, and you’re working with a complex data structure of many nodes. Understanding connections are the key to understanding dependencies and uncovering cascading impacts – and graph technology may be the only feasible way of capturing all that density and inter-connectedness.

Graphs and the smart home

Isn’t it possible to manage this volume of connections with my existing database technology, however? While it’s true that some smart home IoT problems could be handled by a relational database, they’re not an especially satisfactory fit. That’s because they represent data as tables, not networks, and IoT queries strain a data structure not set up to map connections.

That’s why observers like the analysts think that this level of device IoT functionality can best be implemented in a graph database, as they process complex, multi-dimensional networks of connections at speed, especially if these are native or thoroughgoing graphs. Alex Woodie, Managing Editor of Datanami, says, “One of those key enabling technologies [of IoT] is graph databases.” Tony Baer at Ovum: “Graph technology will allow the Internet of Things to be represented transparently… without the need to force fit into arbitrary relational models.” And, Matt Aslett, Research Director at 451 Research comments: “Graph databases offer the potential to store and analyze data from the IoT”.

Another independent commentator agrees: analyst Quocirca says that “The prime benefit of a graph database is that it operates on a pattern-matching basis, [disregarding] everything that is not related to what it needs and focusing only on what is relevant. Graph databases are therefore ideal for dealing with the real-time nature of IoT.”

Nordic and Baltic telco leader Telia, for example, has embraced graph technology, with its new digital ecosystem and platform for broadband connections, Telia Zone. It can potentially serve 1.2 million users, with graph-powered smart home management a major feature.

The vision is to be part of the big eco-systems alongside Apple, Google, and Amazon, giving customers a slew of fun and useful broadband-delivered consumer digital services.

Telia Zone is the basis for a future smart home system, starting with VOD and home entertainment, and is expected to have 13 million devices as individual nodes, with 20-30,000 events per second.

It’s, therefore, a sizeable IoT network, with the Neo4j graph database there to help the Telia team create new connections on the fly and make new APIs out of any that may become desirable. Also, the firm will soon add in AI (artificial intelligence) and Machine Learning, and it says that graph technology is the best way to handle that.

Graph analytics is at the heart of how Telia Zone understands data, in effect – using the graph with the router as the node, apps as connections, and other locations or devices where a subscriber connects to Telia Zone giving a wider context. Predictive analytics can also look at when two known methods are in the same zone or when a subscriber is likely to arrive home.

With the backing of the analysts and the example of pioneers like Telia showing the way, it’s hard to disagree that it’s graph-based IoT management that promises to be the shortest way to get us to the smart domestic future we all want.