The current state of the industrial IoT (IIoT) is defined by decreasing hardware and component costs as well as unsatisfactory data analytics implementations. Cheap sensors and actuators allow us to have eyes and ears everywhere however most enterprises (SME) struggle to analyse the load of information and to benefit from the flood of data.
Precise and sophisticated big data analytics are the key to success in the future as the intelligence retrieved from data helps companies to increase efficiency, implement customer feedback in products and optimise production. Data became the world’s new currency but there are not many banks which can help industrial corporations to successfully manage and invest it.
Few companies have the distributed systems and data engineering expertise to embrace big data, NoSQL and streaming analytics technologies and are capable to develop and integrate custom analytics platforms on their own.
The startup Ubeeko aims to change that and developed a highly innovative platform targeted at customers who want to take advantage of big data analytics without heavy expenses: HFactory – a fully customisable solution which stands out due to the velocity of analytics, simple configuration, high flexibility and easy-to-handle user interface (UI).
HFactory is a full-stack platform which fully hides away the complexity of distributed systems. The smart solution is able to store and digest billions of data points as well as to stream and analyse data in real-time.
In terms of data management Ubeeko relies on the Apache HBase wide column datastore for ultimate robustness and versatility.
On the data analytics side HFactory integrates with Apache Spark for distributed processing and data analytics.
NoSQL data is digested within the abstraction layer of the platform and delivered through a high-level, business-oriented RESTful interface.
HFactory is an integrated, self-contained environment including an application stack on top of Hadoop which handles HBase and Hadoop for you. Therefore enterprise users can concentrate on the business logic and set up the platform without programming skills.
All of this translates into pain free interaction with big data and smooth, advanced analytics: the Spark API makes domain-specific machine learning algorithms as easy to deploy as a function call.
The freedom to decide whether HFactory is deployed and managed on own servers, hosted in the cloud or fully managed by Ubeeko as a service further helps to find a perfect fit for each company and budget.
For the industrial IoT Ubeeko developed a clever extension called “HFactory TSA” which adds the capability to monitor and analyse time series data. This solution is tailored to the management of IIoT-style data and works out of the box with input from all kinds of devices like sensors, IT systems, robots or SCADA systems. Thanks to the Apache Spark Streaming API the platform can easily cope with 10.000 of sensors transmitting data points at high frequencies.
In a nutshell the HFactory TSA covers the complete data lifecycle from real-time capturing to interactive visualisation.
As Ubeeko recently joined the EBV supported accelerator Le Connected Camp which is part of the IoT Valley in Toulouse, France, we have been able to talk to Ghislain Mazars, CEO and founder of Ubeeko, in order to get some more insights:
What is the core element of this incredibly flexible system you are developing?
What’s making HFactory really unique is the abstraction layer that we have built on top of reference Big Data & NoSQL technologies. Instead of low-level interfaces, we map NoSQL data to their JSON representation and provide a business-oriented API to simplify data manipulation. In other words, our core value resides in offering high level and simple web interfaces on top of a distributed Big Data infrastructure.
Which key challenges can your solution address within industrial IoT systems?
HFactory makes it easy to develop real-time monitoring applications and alert systems on various types of timeseries data, from aggregated data residing in standard OT systems to raw data generated by new IoT-style connected sensors. HFactory first abstracts away the complexities of Big Data technologies to provide a programmatic access to data to the application developer and then helps him build tailored applications using off-the-shelf, reusable graphical and analytic components.
What is the most exiting feature or application area you see for HFactory?
HFactory facilitates and accelerates the development of applications and alert systems on various types of timeseries data, for use cases such as manufacturing analytics, connected agriculture, building performance management or smart grid monitoring. Remote equipment monitoring and predictive maintenance as part of Industry 4.0 projects are definitely the area where we see the most interest and growth potential.
What are the major technical challenges you face during development and roll out?
As we were working with cutting-edge technologies, we have often faced compatibility issues when integrating new components to the core platform. With the focus now shifting to the development of vertical packages and applications, these technical challenges have significantly lessened. Now the biggest challenge is in transitioning our focus to front-end development and bringing on board new design and UI skills.
How is LeConnected Camp helping you to achieve your goals?
Le Connected Camp has already proved to be very helpful in connecting us with prospective customers and opening up new business opportunities. As an example, we are about to start a project with an industrial company that we met in the course of the acceleration program. Being part of the dynamic IoT Valley ecosystem is also very beneficial to attract new technical recruits and improve the visibility of our startup.