The number of connected devices, as well as the amount of data they generate, grows exponentially each year and will continue doing so. IoT Analytics estimates that by 2020, 25 billion connected devices will produce around 44 zettabytes of data.
This enormous set of insightful information will represent a disrupting power capable of creating trillion-dollar business opportunities and use cases. Within 10-15 years, IoT data will become a major business asset, with more and more companies building their business models solely upon the connected data insights.
The real value of IoT data hides in its quantity: the more statistics are generated, the more tangible benefits businesses are able to get. However, to derive this real value from the data, one should get to the core of it and then apply the result to a certain business problem. The overall shift towards data-driven business models gave impetus to the emergence of IoT platforms as hubs for the IoT data storage, analytics, visualization, and application.
IoT Platform vs. IoT Middleware
IoT solutions, particularly, platforms have appeared as a backbone of the connected ecosystem, bringing together Big Data and the tools for its handling. Nowadays, there are more than 300 platforms on the market, and they differ greatly in terms of technological depth and purpose.
Experts argue whether all solutions available today can be called full-fledged platforms and here is why.
The logic architecture of a full-scale IT platform contains the following “building blocks”:
- Device connectivity
- Data processing, analysis, and management
- Presentation and business connectivity (integration of analytics and application of insights to a specific problem)
- Most solutions available on the market work at the connectivity level, serving as gateways for cross-device communication. This is how KAA, Microsoft Azure, Eclipse, Hortonworks and other major market players operate. Though device connectivity is a cornerstone of an IoT platform, taken separately, it represents a mediator between a device and a real platform. That is what we call IoT middleware.
So, what do we call a real platform then?
At least a three-layer infrastructure built around the IoT gateway:
- A source layer to derive the insightful data from. It may contain data from the connected devices, social media, weather conditions, prices for a certain commodity, etc., whatever data is relevant for a particular company or industry;
- An ETL layer to process and store the collected information;
- A presentation layer to represent the data in the form so that they could easily be applied to a specific problem.
- Thus, if you purchase one of the existing solutions, you are very likely to acquire a piece of middleware, not a full-scale IoT platform.
Full-Scale IoT Platforms in a Nutshell
To build a full-fledged IoT platform, one needs to consider not only its multilayer structure but also the diverse needs of customers. The thing is that a truly universal IoT platform should serve a retail store with a couple of beacons and an oil tanker with 1,000+ sensors on board equally well. Besides, not only the amount of the incoming data matters but also how fast it can be transformed into an applicable insight.
Thus, for an IoT platform to properly fill the gaps between device sensors and data networks, it should satisfy the following requirements:
- Interoperability. Open APIs and SDKs enable businesses to bring different hardware and software standards to the common denominator, thereby encouraging faster evolvement of the universal IoT infrastructure. As a result, transactions within the IoT environment become more transparent, secure and consumer-oriented.
- Smooth cross-device communication and 100% message delivery. Internet of Things connects various types of devices, from smartphones to fitness bracelets to radio nannies. In this regard, seamless connection and exactly-once delivery become the core of proper system operation, especially when a message transmitted from one device to another requires an immediate follow-up action.
- Support of zero downtime and failure tolerance.Imagine a gambling company that accepts bets in real time. At peak periods, for example, during important matches, the number of messages from various devices may reach as much as 2 million per second. If a server suddenly fails and millions of bets turn out to be undelivered, it may entail huge financial losses for both parties, which is, of course, unacceptable.
- Support of any number of end points and users.It is essential for an IoT platform to be able to gather data from the unlimited number of devices and sensors seamlessly, especially when it deals with high-load systems. As the number of connected devices is only growing, in the long run, this feature will become a cornerstone of any IoT platform.
- 100% security and privacy. Personal data privacy is a “stumbling block” on the way to unrestrained connectivity in the true sense of the word. On the other hand, unfortunately, this is mostly connectivity that causes sensitive data leakages and unauthorized usage today. That’s why healthcare, finance and other industries working with sensitive data require any software to be fully compliant with their privacy standards. Otherwise, there is no chance for a solution to enter the market.