Security and Privacy Issues in Analytics

This study discusses the security and privacy issues faced by Netflix, including vulnerabilities in privacy and the risk of data breaches. It also explores the ethical issues related to analytics and the utilization of machine learning in the context of Netflix data. The study provides insights into the need for improved analytics and privacy measures.

Running head: SECURITY AND PRIVACY ISSUES IN ANALYTICS
SECURITY AND PRIVACY ISSUES IN ANALYTICS
Name of the Student
Name of the Organization
Author Note

1SECURITY AND PRIVACY ISSUES IN ANALYTICS
Executive Summary
The main aim of the study is to describe about all kinds of privacy issues which are being faced
by Netflix recently as seen by CTO. There is a need for the improvement of the analytics of the
company. It has been noticed that there is a significant research into the vulnerabilities of privacy
form the release of the training data of Netflix. The report comprises of a high technical level
showing the possibility of attack on the anonymity of the set of data. A discussion is made on the
privacy which is raised by the types of analytical datasets including the technical issues. The
report also includes the ethical issues as well as the analytics. The CTO has thought of bidding a
contract with the local government for building a recognition system of image for the
enforcement of law. A brief discussion is made on the ethical issues which may rise due to the
utilisation of machine learning in the context of Netflix data set.

2SECURITY AND PRIVACY ISSUES IN ANALYTICS
Privacy Issues
Netflix mostly commits the future of it in streaming movies to all the customers and it
almost relies on the cloud services exclusively for the infrastructure of it. This raises several
concerns of security which definitely needs a new thinking process. Netflix is seen to have been
developed software and pushes this into the ultimate production via the cloud and several
characteristics of data centers which are traditional are not tolerated due to this (Yang et al.
2017). There may be chance of the code being changed which are pushed to the systems which
are running. Therefore, there is a need for writing and replacing the older versions with all new
instances entirely. In all the data centres which are traditional, several teams have their own ways
of application deployment and updates them. Therefore, all the versions of application may
disappear as the groups mostly tweak them for using individually by the creation of slight
different versions which becomes almost impossible to sync. The cloud cannot support such
practices. Netflix does not have any control over the network which is underlying. The
deployments of cloud possess nodes ephemeral nodes which are mainly the instances which may
disappear at any time. If the security changes are viewed, it is seen that if the applications are
pushed and are remained unchanged till they are replaced, there must not be no problem of
integrity of file (Zhu et al. 2013). The monitoring of activity goes down as there are no reasons
for the administrators to log in and out to the patch. There is a need for the staffs of security to
add accounts of user, systems which are inventory, change of the configuration of firewall and
also take screenshots of the analysis drives. Despite all the cloud providers offering way for
addressing concerns of security, there still remains a number of problems. With a number of new
nodes which are being created containing full new nodes and a number of others are taken down
as they are getting replaced, the administrators are not being able to monitor all the IP addresses.

3SECURITY AND PRIVACY ISSUES IN ANALYTICS
Recently, it has been seen that a popular original show of Netflix “Orange is the New Black” is
leaked as an attack on the most streaming postproduction service company. This breach and the
extortion act is exposed as a chink which is critical in the cyber security (Zyskind and Nathan
2015). This will continue to be a cause of same detrimental cyber breaches if proper security
measures are not taken for defending against them. Third party vulnerabilities pose a huge threat
to the systems of cyber defence and may result in a huge loss of proprietary property. This
breach is an expected one as there has been an early warning about the weak security at the
vendors of third party for several years (Shmueli 2017). Despite several warnings, the leadership
at the Netflix and other producers of the content have not acted for ensuring third parties in the
ecosystem of production and there is no established policies for getting protection from these
breaches.
Netflix has published about 10million rankings of movie by about 500000 customers as
challenge part for the people to come with good systems of recommendation than what the
company is using recently (Casino et al. 2013). The data is then anonymised by the removal of
personal details and the replacement of the names with numbers which are random for protecting
the recommender’s privacy. It has been noticed that there is a new class of de-anonymization
attacks which are statistical against the micro data which are high dimensional like preferences
of individual, records of transaction and recommendations (Kshetri 2014). The methodology of
de-anonymization is to be applied to the prize dataset of Netflix containing ratings of anonymous
movies of about 500000 subscribers of Netflix which is considered to be the largest online rental
service of movie in the world (Jeckmans et al. 2013). It is demonstrated that an adversary
knowing only a little about the subscriber of an individual can identify this record of subscriber
very easily within the dataset. With the utilisation of database of internet movie as a source of a

4SECURITY AND PRIVACY ISSUES IN ANALYTICS
knowledge in background, Netflix records of the users which are known are to be identified by
uncovering their apparent preferences which are political and other information which are
potentially sensitive. Sets of data include all the legally protected preferences of individual as
well as transactions which people view as private. The risks of privacy of micro data publishing
are already known. Even identifiers like names and the security members are removed, the
adversary can utilise the knowledge background and cross relationship with other databases for
re-identifying data records of individual. Famous attacks mostly include de-anonymization of a
hospital discharge database of Massachusetts hospital by joining with database of public voter
and breaches of privacy caused by AOL search data. Micro data is mainly characterized by both
sparsity and high dimensionality. Each records comprises of many attributes which are viewed as
dimensions. Sparsity generally means for an average record, there will be no same type of
records in a space which is multi-dimensional as defined by the attributes. First contribution
must be a formal model for all the privacy breaches in micro data which is anonymised
(Sedayao, Bhardwaj and Gorade 2014). It can be defined by two ways, one which will be based
on the successful de-anonymization probability and the other one is the amount of recovered
information about the main target. This model thus involves a much wider class of the attacks of
de-anonymization then the simple correlation of cross-database. The second contribution is
nothing but a general class of the algorithms of de-anonymization which demonstrates the limits
of privacy which are mental in the micro data which is public. This algorithm of de-
anonymization is considered to be robust to the imprecision of the background knowledge of
adversary and to the perturbation which is to be applied to the data which are prior to be
released. It works only when a subset of the data which is original are to be published. The third
contribution is the final practical analysis of the Prize dataset of Netflix which contains movie

5SECURITY AND PRIVACY ISSUES IN ANALYTICS
ratings of 500000 which are anonymised. Netflix published this particular set of data for
supporting the data mining contest of Netflix Prize. An adversary who knows a little about some
subscriber can identify easily the record if present in the set of data. There is a need for the
background of knowledge of the adversary to be precise. The algorithm is robust which is
responsible for uniquely identifying a record in the data set which is published.
Ethical issues and Analytics
One of the main or the crucial ethical issues which is faced by Netflix is the problem of
throttling. The company is seen to have been reducing the video quality for the users who are
accessing network through Wi-Fi. This practice is mainly meant for helping users in avoiding
data caps of mobile (Townsend and Wallace 2016). This is definitely an ethical issue which does
not allow the users in enjoying the privileges of subscription for which they have paid.
Therefore, it has been observed that a program has been established for several universities to
enjoy the watching movies free of charge. Another ethical issue which is faced by Netflix is its
ability of obtaining adequate permissions as well as licenses for streaming shows as well as
movies. Due to the expansion of it into video streaming, Netflix has been seen to have been
battling lawsuits for the violations of copyright (Guimarães, Pinho and Milani 2016). Extra
claims of violations which are ethical due to licensing are properly mentioned throughout the
sources of internet and however limited information is available for discussing such kinds of
issues. There are several issues which are related the utilisation of machine learning in the
context of Netflix. There is a requirement for keeping the data of the customers and their
identities private. Privacy does not only mean secrecy as data which is private may need to be
audited depending upon the requirements which are legal, but that the data which is private are
obtained from the person with their own consent is not exposed for utilisation by some other

6SECURITY AND PRIVACY ISSUES IN ANALYTICS
individuals with any kinds of traces to their respective identities. The second issue being the
confidentiality of the shared information which is private. All the companies which belong to the
third party, share a lot of data which is very much sensitive and there is a huge need for having
restriction on how the information is further shared. Third issue being the transparent view of the
customers. It is very much needed for the users to have a view which will be transparent showing
how the data is being used and the capability of managing the flow of information (Johnson
2014). This information flow is private across the analytical system which is massive and
belongs to the third party. Another issue is the interference of big data upon the will of human.
The analytics of big data can both determine as well as moderate who the users are before they
make up their own minds. Netflix must begin to think about all the kinds of predictions as well
as interferences which are allowed and those which are not. All the database administrators of
Netflix are very much needed for handling bug data for providing a voice in the discussion based
on ethics about the way the data is to be utilised (Koene et al. 2015). There is a need for Netflix
in discussing openly about all the dilemmas in both informal as well as formal forums if any. The
transparency in Netflix is to be taken into consideration to a great extent. It is to be seen about
how much data is to be provided to the users. Most of the users realise that some amount of
information is being monitored as well as analysed (Soliman et al. 2013). This is a ubiquitous
experience online in the recent culture of internet. The consent is also to be taken into mind. It is
to be checked if some or all videos or films or other data can be used by the users without
consent. It is also to be checked if all the users are allowed to be anonymous online or not. There
are several general consensus which obligated ethically for obtaining some kind of user sign-off,
but there is no standard which is agreed-upon of what the sign-off must include (Savage 2017).
There are several other issues as well which are related to the data analytics of Netflix. Crossing

7SECURITY AND PRIVACY ISSUES IN ANALYTICS
line may create lawsuits. Netflix, while transparent about the leveraging data of users for
advertising can be named as a class action filed on behalf of the users. It is very much needed to
be honest by maintaining transparency of data. Transparency is how data which is utilised
enables the users in deciding if they wish to provide their personal data or not. It is very much
important for protecting the privacy of the users. It is to be checked if the data is not being
misused potentially during the sharing of data anonymously.

8SECURITY AND PRIVACY ISSUES IN ANALYTICS
References
Casino, F., Patsakis, C., Puig, D. and Solanas, A., 2013, September. On privacy preserving
collaborative filtering: Current trends, open problems, and new issues. In 2013 IEEE
10th International Conference on e-Business Engineering (pp. 244-249). IEEE.
Guimarães, J.A.C., Pinho, F.A. and Milani, S.O., 2016. Theoretical Dialogs About Ethical Issues
in Knowledge Organization: García Gutiérrez, Hudon, Beghtol, and Olson. KO
KNOWLEDGE ORGANIZATION, 43(5), pp.338-350.
Jeckmans, A.J., Beye, M., Erkin, Z., Hartel, P., Lagendijk, R.L. and Tang, Q., 2013. Privacy in
recommender systems. In Social media retrieval (pp. 263-281). Springer, London.
Johnson, J.A., 2014. The ethics of big data in higher education. International Review of
Information Ethics, 21(21), pp.3-10.
Koene, A., Perez, E., Carter, C.J., Statache, R., Adolphs, S., O’Malley, C., Rodden, T. and
McAuley, D., 2015, May. Ethics of personalized information filtering. In International
Conference on Internet Science (pp. 123-132). Springer, Cham.
Kshetri, N., 2014. Big data׳ s impact on privacy, security and consumer
welfare. Telecommunications Policy, 38(11), pp.1134-1145.
Savage, T.A., 2017. Ethical issues in school nursing. OJIN: Online Journal of Issues in
Nursing, 22(3).

9SECURITY AND PRIVACY ISSUES IN ANALYTICS
Sedayao, J., Bhardwaj, R. and Gorade, N., 2014, June. Making big data, privacy, and
anonymization work together in the enterprise: experiences and issues. In 2014 IEEE
International Congress on Big Data (pp. 601-607). IEEE.
Shmueli, G., 2017. Analyzing behavioral big data: methodological, practical, ethical, and moral
issues. Quality Engineering, 29(1), pp.57-74.
Soliman, O., Rezgui, A., Soliman, H. and Manea, N., 2013, August. Mobile cloud gaming:
Issues and challenges. In International Conference on Mobile Web and Information
Systems (pp. 121-128). Springer, Berlin, Heidelberg.
Townsend, L. and Wallace, C., 2016. Social media research: A guide to ethics. Aberdeen:
University of Aberdeen.
Yang, Y., Wu, L., Yin, G., Li, L. and Zhao, H., 2017. A survey on security and privacy issues in
Internet-of-Things. IEEE Internet of Things Journal, 4(5), pp.1250-1258.
Zhu, T., Li, G., Ren, Y., Zhou, W. and Xiong, P., 2013, August. Differential privacy for
neighborhood-based collaborative filtering. In Proceedings of the 2013 IEEE/ACM
International Conference on Advances in Social Networks Analysis and Mining (pp. 752-
759). ACM.
Zyskind, G. and Nathan, O., 2015, May. Decentralizing privacy: Using blockchain to protect
personal data. In 2015 IEEE Security and Privacy Workshops (pp. 180-184). IEEE.