Using Intelligent Agents Paradigm in Big Data Security Risks Mitigation

Introduction

The potential of the big data approach is far from being fully exploited nowadays. This happens due to the slow adoption of cloud based architectures that are big data main support technologies. The interest in cloud adoption is high at the level of economical and research based environments (HARDY, 2016). Yet, there are a lot of problems that make this transition slow. One consists in the higher costs involved by rewriting the commonly used applications as native ones. To solve this transition the most common method is to virtualize the real machine together with its applications. This solves the cloud transition and decreases long time maintenance costs. Unfortunately this means an inefficient use of the cloud resources, so the porting of the used application or movement to newly appeared native similar cloud applications is required (Badola, 2015).

The cloud solution has not yet reached its maturity as technology. As a result there are a lot of problems concerning its security (Khana & Al-Yasirib, 2016). The same security problem appears in the majority of applications because the process of designing and implementing a secure application is at least twice expensive than making an application with a reasonable security level. So, many common applications are not adapted to the asymmetric informational war that is ongoing nowadays at global level (Lasconjarias & Larsen, 2015). This must involve fundamental changes at any levels of the information society beginning with the application design level (Mumtaz, Alshayeb, Mahmood, & Niazi, 2018) and ending with paradigm changes in global security approach.

The public cloud based systems or private ones will be the computing nodes in future global cyberspace (Sharma, 2016). If we take this into account, then there is a need of fundamental changes in the used methods for security risks mitigation. A good idea will be to use some of the military grade protocols. If this approach is chosen it may involve a significant deployment cost increase. It is true that military approaches are better prepared from a security point of view (Eggen, et al., 2013) and in many cases there is a physical isolation between public communication networks and their internal ones. Yet, as normal, there are still two situations when this isolation is partially broken. One situation concerns the needed intersection points among the public domains and the military ones. Inhere there are a lot of problems generated due to disrespecting the internal security procedures by the people or targeted attacks of the third parties (private or from governmental level). The other one refers to the cases when, sometimes, the existing public communication network is used to handle high sensitive information. In these cases even if supplementary security measures are taken, no one can guaranty a full security.

To make the situation even worse, a new era in designing cryptographic algorithm begins due to the fragility of current approaches when quantum computing is used (Mood, 2016). Besides that, the information flows grow exponentially and the dedicated existing structures specialized in information security assurance begin to be overcome. As a result, major changes at the level of used cyber defense paradigms are required no matter if the military public or private domains are analyzed.

To design a new cyber defense paradigm for big data specific infrastructure some aspects must be taken into account:

• •

  The redesign of the main big data architecture by increasing the security involvement.

• •

  The use of the Artificial Intelligence (AI) based assisted decision support systems.

• •

  The use of the AI dedicated entities that will emulate at low and middle level a security expert administrator.

• •

  To make a leap from a static based structure of the cyber defense networks to a dynamic one that will have a better reaction time and collaboration in handling the current attacks.