The argument has long since been settled concerning traditional relational database management systems (RDMS) versus the new wave of open source database management systems (OSDBMS).
As opposed to being an either-or proposition, the consensus is that there is a place for both approaches in the grand scheme of things.
But the onset of the Internet of Things (IoT) is shifting the balance of power even further to the side of open source NoSQL databases, such as MongoDB, Redis, and Neo4J.
The IoT consists of a combination of sensors, components, equipment, devices, systems, microprocessors, and analytics, most which operate without human interaction. Now consider this:
From a little more than 5 zettabytes (ZB) of worldwide information and 50 million potentially connected devices and sensors, predicts that the digital universe will expand to 44 ZB and 200 billion potential IoT connection points by 2020. Some are even talking about trillion-sensor networks within a decade.
However realistic such predictions turn out to be, this trend will inevitably exert a massive impact upon the database marketplace. For example, Gartner research indicates that 50% of existing commercial RDBMS instances will be converted to open source by 2018.
By that time, more than 70% of new in-house applications will be developed on an OSDBMS. This dominance of NoSQL is only going to accelerate as IoT gains momentum.
Certainly, a portion of the traffic will fit snuggly onto the RDBMS. But most of it won’t. There is just too much data on too many disparate formats to make it viable to attempt to host IoT on traditional data structures while maintaining reliability.
The many flavors of NoSQL, on the other hand, can deal with such variety and scale with relative ease.
MongoDB, for instance, is designed to cope with very large data volumes, complex processing, and rapidly changing needs without being held back by rigid schema. That makes it perfect for many, but not all, IoT use cases.
For high performance and rapid data ingestion, in-memory of certain NoSQL tools is the best choice. Other options offer further versatility. Thus a combination of two or more OSDBMS systems along with the occasional mainline RDBMS is likely to be the winning formula as IoT matures.
Analytics + IoT = NoSQL
Scale and versatility are obvious reasons for the onslaught of open source. But once analytics are factored into the equation, the need for NoSQL becomes all the more apparent.
For the full vision of IoT to be realized, it is no longer sufficient to provide historical analysis to summarize the past month, quarter, or year.
Thus a new breed of versatile real-time, near-real, and data streaming analytics tools are being incorporated into open source databases. They possess the flexibility, agility, and velocity required by billions of connected things – which are transmitting largely unstructured data that demands rapid conclusions, adjustments, and changes.
Further, the typically centralized RDBMS arrangement of “all roads leading to Rome” won’t function under the volume of IoT-generated information.
The ability to distribute intelligence to the edge of the network in multiple integrated NoSQL instances is the way to eliminate bottlenecks.
It will also provide actionable insights from data drawn from an exponentially growing pool of diverse sensors, devices, and applications.
The OSDBMS is going to be king in a world dominated by IoT. The RDBMS will play a part, but it lacks the multi-directional immediacy to cope well with vast amounts of structured, unstructured, and semi-structured data. And with analytics becoming the cornerstone of the Internet of Things, NoSQL databases are set to take center stage.
They offer the only platform that can efficiently and cost-effectively bring data together into tools that can be used to analyze and make sense of all available information.
Read more about how MongoDB specifically addresses the unique challenges presented by IoT.