Toward the advancement of security in Big Data

Big Data Expo held two weeks ago, on March 11th and 12th at the Paris Palais des Congrès. The theme of this brand new was “Accelerate the Future”.

No doubt, the record number of exhibitors and an energy and buzz never before seen will forge the memory of this event assembling AI enthusiasts. But beyond Data and its role in improving the performance of companies of all kinds and in all sectors, the future remains worrisome. Face to face with those who write algorithms and conjure up deep learning, visitors are sometimes frightened at the prospect of a total loss of control.

Currently, the challenge for all types of data is accuracy and protection, even before imagining what can be done with it. Many of these AI-oriented companies invite us to collect and store our data in the famous Data Lake. They’re trying to demonstrate that there’s a way to learn real lessons. And also to make real progress for people and their professions.

But can we trust them?


How can we better protect data in a world that’s becoming more and more open and digital?

To better understand how security is a prerequisite, what could be more inspiring than to listen to Squadron Vice-Admiral, Arnaud Coustillière, Managing Director of Digital for the Ministry of the Armed Forces, an information service which employs 20,000 people in France.

Working with extremely sensitive data in a culture where confidentiality is king requires both openness and a sense of caution. It’s all about the constant search for balance between protection and opportunities for use. “The challenge is daunting, but it’s the same for any organization,” explains the Vice-Admiral.


Hadoop and Gemalto

Next to take the stage were Hervé Bruneteaud and Raphaël de Cormis. Hervé Bruneteaud is the Director of the Datawarehouse Competence Center for Orange. And Raphaël de Cormis is the VP of Innovation Labs at Gelmalto.

Hervé Bruneteaud explained the advantages of using the Hadoop platform for hosting datalakes across various uses. While Raphaël discussed the need for more thorough identity checks within information systems. The goal remains to guarantee the quality of information shared remotely.

Let’s be clear: Hadoop’s platform is a software framework developed in Java and open source. Made popular by its free access, Hadoop has evolved from one version to another. But is still intended for the handling of raw data storage and relatively simple queries. Hadoop had been adopted by the majority of the largest American companies. Although it seems a bit less attractive today.


From its side, Gemalto works on the identification and accuracy of information. Which raises questions about the main criticism of Big Data: Does the race to amass huge volumes of data result in the agglomeration of everything and anything? Isn’t the applicability of the data more important than the quantity of data collected? After all, just a bit of bad data can render the analysis invalid. Algorithms would be fed by bad information, misleading us as a result. And what is more dangerous than an analysis based on bad information?


And if the critical element of Big Data is human, are humans the only ones who can guarantee the accuracy and security of the information?

Shivers in the room! Consequently, the warning was issued in a recent study (conducted by Cybersecurity Ventures). It indicates that in 2021, there will be more than 3.5 million cyber security profiles worldwide. Without a doubt, it’s in humans that companies will need to invest. As much as for extracting intelligence from Big Data as for ensuring its safety. The lack of employees trained in security issues is already being felt, and some companies are taking the initiative to offer outsourced solutions.


For example, Gareth Maclachlan, VP of Strategy and Product Development at FireEye, states, “We have developed Expertise on Demand to give organizations direct access to the best expertise in cyber security, whatever their needs and/or budget. By combining products and expertise in this flexible and expandable way, we supply our clients with a selection of resources to support their employees in building the best line of defense.


As such, there is no single vision for Big Data, and securing data remains a major challenge in light of the evolution of storage volumes. We can’t deny the fact that neither the Cloud nor open source currently offer sufficient guarantees, and they remain vulnerable.

Human “guardians” still have a future!


To read: Governance platform: a new tool for data security