Basis of Big Data

Discuss about the Security and Privacy Issues in Big Data Analytics.

Everything in today’s technological world is incomplete without data. Innovation in technology has altered the very definition of information and its aspects. Data is the raw and unprocessed form of information. Big Data is a term used to define the heaps of data used to analyze patterns and other statistical information for strategical purpose in variety of organizations. Data mining technology and its advancements has brought a revolution and is the genesis of Big Data (Kshetri 2014). It has its advantages and challenges. Big Data and its analysis has quite a number of challenges that are to be faced to keep it secure and private for the use of targeted entities. In this review, the concerns and its remedies are discussed in relation with the Big Data and its security in the network. Discussion of every aspect of security of Big Data analysis is provided in a very precise manner.

Big Data technology is different from the previously data mining and processing techniques. It allows larger amount of data to be processed at one time. This technique differs in three ways from the previous methods such as the conversion and transmission rate of data, types of data that is processed and the amount of data to be processed. This technology is advanced in way so that it can connect and accept data from different sources and in different formats. Analysis of data helps the organizations to make deductions, draw conclusions and predict market behavior. Cloud computing has also proven to be a vital platform for the storage, processing and collection of data (Moura and Serrão 2016). It is much easier to mine data available on the cloud because it is connected to the internet. The success of Big Data can be credited by the software Apache Hadoop. This software platforms helps Big Data to be collected and stored in clusters which makes it easier to process and analyze. It is scalable to use as it can access thousands of severs and its threads from one system (Moreno, Serrano and Fernández-Medina 2016).

Mapping study of the software is done in four stages that gives a comprehensive idea of how the data is scaled thorough different systems in a network of systems. The phases involved are:-

Questions of research: During the exploration of questioning methods it is found that issues related to objectives and other security dimensions are adhered to in some methodologies developed previously to counter these problems (Moreno, Serrano and Fernández-Medina 2016).

Systematic Study of Mapping

Methods of research: Different online libraries are searched and are used as methods. These libraries are Digital Library of IEEE, ACM, and SCOPUS.

Analysis of data: Data that is collected from different sources and filtered on certain parameters so that it can result in the targeted statistic development.

Data Interpretation: The data that is mined and processed is used for interpretation by different organizations and for different processes (Moreno, Serrano and Fernández-Medina 2016).

In spite of the advancements and features provided by the data mining technology, many technocrats are quite concerned about the security of this amount of large data, its privacy and its application. A heap of data contains many types of private data of the individuals that can prove detrimental to the people if leaked of made public. In acquiring these types of data, unauthorized access is done and undesirable procurement is executed. In this process the privacy of an individual in breached without his consent which calls for a serious issue of data security. To deal with such an issue a new technology in compliance with data mining has been developed which is privacy preserving data mining (PPDM) (Kshetri 2014). This technology aims at non-disclosure of data and to retain the application of data in the desired manner.There are two aspects of PPDM in data security. One deals with the privacy of sensitive individual’s data such as social security number, bank details user id and passwords. Another one deals with the analytical results of the large amount of data whose leakage can lead to serious security consequences (Xu et al. 2014).

According to Xu et al. (2014), privacy protection and its approaches  are as follows:

• Access Limitation: Data acquired by the collector is done in active and passive way. In the active way collection of data is done by filling a form of surveying which is done by the data collector itself. But in passive way data collector can gather it during the routine activities of the provider.
• Privacy of Trade and Benefit: In some cases the data provider compromises the security for the benefits of data transaction to a concerned collector. For future benefits and ease of business data is traded to a trustworthy collector.
• Availability of False Data: There are trust issues with some data collectors about their intentions to use the data. To protect data from getting misused data provides give false information to the collectors to protect the interests of individuals.

Increased risk of leakage of information: Data range and coverage in cyber space is unfathomable. Increased number of social networking sites and email services has exponentially increased the amount of data. As these details are personal, risk of leakage is huge. In data mining different types of data such as business transactions, classified government records. Two very important facets of information is derived from sources such as personal data and national data. Personal data can cause damage to individuals or group of persons in a very effective way. Data mining and Big Data provides innumerable facilities to the consumers and people using technology but can also cause grievous consequences if leaked or got in wrong hands (Thayananthan and Albeshri  2015). Personal information such as house address, telephone number, credit card details, emails accounts, passwords and one time passwords are some that can be exemplified. Leakage of these information can cause serious economic consequences to the affected person. Coming to the national security issue related with the data one can observe that automation of all the government agencies has been done rapidly. Due to this, government records and sensitive information is uploaded on the internet which automatically makes it vulnerable to leakage and disclosure. Defense records and other data of various departments are available for the taking to hackers and interested parties. This raises serious concern on the front of national security (Mengke et al. 2016).

Concern of Privacy in Data Mining

According to Mengke et al. (2016), security platform of big data can be divided into five layers that are listed below:

• Storage of data
• Processing of data
• Big Data Interface
• Big Data Security management
• Application security of big data

These five layers of Big Data Security also includes the application of the collected data.

As discussed above, Big Data contains variety of data from sources that are very much valuable to be oozed out in the hands of unwanted sources. Big data combined with cloud computing mechanisms are distributed in different layers. These layers are of application, hardware, network, business process, database and operation. Moreover, cloud computing provides its certain privileged assistance in the face of elastic storage, elastic computing, elastic networks, BPaaS, virtualization and AaaS. These processes and methods assists Big Data with the following facilities of faster decision making, Cost of DB License, Running of simulations, Scaling in and out, Reducing intensity of IT and running of environments (Gholami and Laure 2016).

As per Ahmed and Saeed (2014) Cloud computing structure that deals with the Big Data comprises of three stages:

Modeling: This model counters with the data-leakage and other attacks of the cyber space which targets the most vulnerable points of the storage.

Analysis: This is done to figure out the possibilities to avoid threats that have been made before.

Implementation: Application of the security models in the structure of cloud.

Cloud computing has certain security mechanisms that are mentioned as follows:

Security of Infrastructure:

1. a) Secured Computations in the Frameworks of Distributed Programming.
2. b) Best security methods for non-relational databases.

Data Privacy:

1. a) Preserving of Private Data in analysis and mining.
2. b) Data security center based on cryptography.
3. c) Granular Access Control.

Management of Data:

1. a) Secure Transaction logs and secure data storage.
2. b) Granular Audits.
3. c) Provenance of Data.

Reactive Security and Integrity.

1. A) Last-mile validation and filtering.
2. b) Monitoring of security in real-time.

Certain core technologies have been developed to provide security and privacy to Big Data:

• Polymorphic Detection of code of JavaScript to defend from APT Attacks: The author of this detection techniques establishes that this method detects malicious software and malware attacksusing suitable graphs as they have a certain relationship. The patterns thus created are recognized by a classifier known as Support Vector Machine. This classifier distinguishes between normal and malicious software (Yu, Mu and Ateniese 2015).
• An algorithm set based on Identification of fault in Wireless Mesh Network: An algorithm by the name of A-SRS is a very efficient method for positioning of malicious routers in communication with other routers in the mesh network. It also ensures connection with available network services and also tracks the least number of detecting routers due to presence of less numbers of routers present in the mesh.
• Procedure of user equipment registration in the networks that are Femto-cell enabled: Equipment registered with the process in this specific network is prone to denial-of-service attacks. For the compatibility of the network, the given mechanism uses the standard of 3GPP (Yu, Mu and Ateniese 2015).
• Trackable encryption with Public-key with Data Sharing for Cloud Storage in mobile:Mainly the public-key encryption is used in data storage and transfer from mobile cloud. Asymmetric-key protocol is applied in transferring of data.
• Hierarchical Structure of Attributes: This method is used to obtain access control systems of fine grained nature. This is based on symmetric fine-tuned encryption that is attribute-based.
• Public Auditing Scheme for Cloud Storage of Data: Other auditing protocols are not sound in achieving a desired level of privacy of the data. Even after deleting all the files of a data provider the malicious program continues to run on the cloud server. This protocol is weak even in the accurate security model.

• Heterogeneity and Incompleteness: When it come for a man to deal with a variety of data to be compiled by a human brain it is comfortably accepted. But in the case of computation by the algorithms from a system, homogeneous data is of much greater choice. As it is a requirement of analysis for the data to be scaled and structured before given as an input for compilation. It can be observed that in Big Data, the sources of information are very diverse and complex. For ease of access for the analytics and to design a specific algorithm data of similar nature is required (Subaira, Gayathri and Sindhujaa 2016).
• Scale: Size of the Big Data is of much concern as there must be a limitation of data size. Analysis of data can only be done up to some extent otherwise it will be very difficult for the designer of algorithms to put in the parameters. Big Data in itself denotes that the size of data will be huge but there has to be an estimation and a restriction of how big the data will be.
• Speed: Time taken by a mechanism to analyze the Big Data is also a matter of concern as any procedure is to be completed within a time frame. Time taken is also dependent on the size of the data as each and every bit of data needs to go through the same process. Complexity of data also determines the time of analysis (Yu, Mu and Ateniese, G. 2015).

Conclusion:

From the above review, it can be established that Big Data and its analytics has certain issues that needs to be taken care of for the security and privacy of consumers. A lot can be at stake if the data used for storage and analysis gets leaked or disclosed. The privacy issues and steps taken to remove the problems are stated in the review that will provide solution for some major issues. Since internet is an open arena for everyone, it is highly unlikely to assume that the data can be completely secured. There are many entities who can breach the security of data if intended to. The review also states the collaboration of cloud computing techniques with Big Data and how it affects the storage, collection and analysis of the Big Data. It has proved to be the base on which Big Data technology has advanced its features to some extent. Since the cloud computing supports Big Data, security of cloud is also of utmost importance. It has certain security mechanism of its own which are discussed above. Size of Big Data, time taken for analysis of the data and application of data in personal and public uses are also discussed in details. Thus it can be concluded that this cocktail of technology is turning tables in the recent IT segment.

Research works regarding this topic can be done more efficiently by targeting a Big Data Analytical company based on Cloud Computing. The real-time observations of problems that occur in the data collection, storage and analysis can be done with greater transparency. Taking a test subject of a Big Data division, each and every facet of security breach can be applied and tested for its resilience so that other problems and weak spots of the security mechanism can be clearly identified.  

References:

Ahmed, E.S.A. and Saeed, R.A., 2014. A Survey of Big Data Cloud Computing Security. International Journal of Computer Science and Software Engineering (IJCSSE), 3(1), pp.78-85.

Gholami, A. and Laure, E., 2016. Big data security and privacy issues in the cloud. International Journal of Network Security & Its Applications (IJNSA), Issue January.

Kshetri, N., 2014. Big data׳ s impact on privacy, security and consumer welfare. Telecommunications Policy, 38(11), pp.1134-1145.

Mengke, Y., Xiaoguang, Z., Jianqiu, Z. and Jianjian, X., 2016. Challenges and solutions of information security issues in the age of big data. China Communications, 13(3), pp.193-202.

Moreno, J., Serrano, M.A. and Fernández-Medina, E., 2016. Main Issues in Big Data Security. Future Internet, 8(3), p.44.

Moura, J. and Serrão, C., 2016. Security and privacy issues of big data.arXiv preprint arXiv:1601.06206.

Subaira,  A. S., Gayathri, R. and Sindhujaa, N., 2016, January. Security Issues and Challenges in Big Data Analysis International Journal of Advanced Research in Computer Science and Software Engineering, 6(2), pp. 530-536

Thayananthan, V. and Albeshri, A., 2015. Big Data Security Issues Based on Quantum Cryptography and Privacy with Authentication for Mobile Data Center. Procedia Computer Science, 50, pp.149-156.

Xu, L., Jiang, C., Wang, J., Yuan, J. and Ren, Y., 2014. Information security in big data: privacy and data mining. IEEE Access, 2, pp.1149-1176.

Yu, Y., Mu, Y. and Ateniese, G. 2015. Recent advances in security and privacy in big data. Journal of Universal Computer Science, 21(3), 365-368.