Ensuring Conﬁdentiality of Information When Processing Operational

Production Plans in Cloud Services

Radda A. Iureva

, Sergey V. Taranov

1,2

, Alexander V. Penskoi

and Artem S. Kremlev

ITMO University, Saint Petersburg, Russian Federation

SPbETU ”LETI”, Saint Petersburg, Russian Federation

Keywords:

Conﬁdentiality, Cloud Services, Operational Plans, Data Anonymization, Homomorphic Encryption.

Abstract:

This paper proposes two methods for ensuring the conﬁdentiality of information transmitted to cloud services

when processing the operational production schedule. The ﬁrst method consists of the consistent classiﬁcation

of critical information and the depersonalization of symbolic parameters, which may be personal or com-

mercial secret, concerning the type of anonymized data. The second method, as an additional gain, involves

homomorphic encryption of numerical parameters.For each of the proposed methods, the disadvantages and

advantages of its use and implementation are described.

1 INTRODUCTION

According to a study by the InfoWatch Analytical

Center, in 2018, there were 70 conﬁdential data leaks

through cloud servers and other insecure information

storages with Internet access. 28.2% of incidents

in the area of sensitive data leakage occurred in the

cloud, with more than 40% of them from storage fa-

cilities owned by high-tech companies (Figure 1). For

example, in August 2019, the personal data of almost

90 thousand customers of the Mastercard payment

system in Germany was discovered on the Web. Such

kind of information can be used for phishing mailings.

Figure 1: Distribution of fraudulent data incidents by chan-

nel.

The use of cloud computing extends the scope of

information security tasks (Park and Tran, 2015) to

protect resources, trade secrets, personal data, and

other user data (Figure 2).

Figure 2: Distribution of leaks from cloud servers by data

type.

When using cloud computing to generate an oper-

ational production schedule, the know-how of the en-

terprise, building plans, logistics ﬂow charts, as well

as personal data of employees are used. These data

are sensitive, as its loss may result in material and rep-

utational damage. As part of this study, anonymiza-

tion and data encryption procedures are considered to

protect the conﬁdentiality of this information. The se-

curity of using a cloud service is described as follows:

CS = (C

imp

), (I

imp

), (A

imp

), (1)

Where index impact = {low, moderate, high}., C

means Conﬁdentiality, I means integrity and A is for

availability.

The Security of Cloud Computing based on fully

Homomorphic Encryption is a rather new concept of

information security, so there are not so many re-

Iureva, R., Taranov, S., Penskoi, A. and Kremlev, A.

Ensuring Conﬁdentiality of Information When Processing Operational Production Plans in Cloud Services.

DOI: 10.5220/0009873201670173

In Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2020), pages 167-173

ISBN: 978-989-758-442-8

167

searchers on ot. In (Singh et al., 2019) is made a

review on this ﬁeld of science and described com-

mon advantages of using homomorphic encryption,

as it is enable to provide the results of calculations

on encrypted data without knowing the raw entries on

which the calculation was carried out respecting the

commercial secrets. In (Kavitha and Harsoor, 2018),

(Murthy and Kavitha, 2019) is made a review and an

analysis of The Security of Cloud Computing based

on fully Homomorphic Encryption to exhibit various

applications of in the real world. It is stressed also that

the system must work efﬁciently without compromis-

ing the required cloud security services. The paper

(Priya and Sumitha, 2018) provides another analysis

with comparison of performance on encryption algo-

rithms. The authors made a conclusion that Paillier

and Elgamal algorithms are considered to be the most

efﬁcient when it comes in terms of security

The purpose of this study is to increase privacy

security.

2 FORMATION OF

REQUIREMENTS FOR

METHODS OF ENSURING

CONFIDENTIALITY WHEN

PROCESSING PRODUCTION

SCHEDULE IN CLOUD

SERVICES

Consider the types of interaction, possible threats,

distribution of powers, and data ﬂows between enti-

ties (information owner and cloud service) to formu-

late requirements for methods of ensuring conﬁden-

tiality. Figure 3 shows the permissions for the cloud

user and information owner.

Figure 3: Powers of the owner of the information and user

of the service.

It is considered to be non-disclosure to third par-

ties in the process of cloud computing when planning

the assembly process of linear drives. To conduct a

vulnerability analysis of the system under study, a use

case diagram of the service user is constructed (Figure

4).

Figure 4: Use Case diagram for service user.

When using a cloud service and ensuring the con-

ﬁdentiality of data processed in the cloud, it must be

borne in mind that the cloud client must be able to

synchronize ﬁles in blocks and be able to quickly ac-

cess any data on the cloud to update or decrypt it

without the need to transfer large amounts of “left”

data. The Figure 5 shows the data ﬂow when they are

uploaded to the cloud for computing the operational

production plan. There are two types of cloud service

usage:

1. The information owner independently encrypts

the data (Figure 5), the key is stored in the infor-

mation owner’s database, sensitive data is trans-

mitted in encrypted form to the cloud. Advan-

tages: Conﬁdential information is not transmit-

ted to the cloud in an open or even anonymous

form. Disadvantages: When synchronizing and

updating information, it is necessary to re-encrypt

huge data sets on the side of the information

owner. Moreover, when encrypting with non-

homomorphic cryptographic algorithms, opera-

tions on numerical parameters will be completely

impossible.

2. The owner of the information fully trusts the

owner of the cloud service and transfers the data

to the representatives of the service for calculat-

ing the operational plan in the clear. Advantages:

Cloud services process open information most ef-

ICINCO 2020 - 17th International Conference on Informatics in Control, Automation and Robotics

168

ﬁciently without the need for pre-processing. In-

formation can be quickly updated by the owner of

the information Disadvantages: There is no pro-

tection against malicious exposure from the cloud

service.

Figure 5: Data ﬂow in the process of download data on

Cloud.

The main problem of encryption in data centers is

that suppliers store keys to user data so that the cloud

service administrator can see all the data related to

personal information and commercial secrets at any

time. Thus, customers of the cloud service have to

overpower themselves and show trust because their

data is not fully protected. The methods proposed in

this paper are going to solve this problem.

When creating distributed applications, it is advis-

able to visualize the network infrastructure of the soft-

ware system. The use of cloud computing power de-

termines the need to address issues related to the secu-

rity and stability of access to information of corporate

clients. Thus, the general conﬁguration and topology

of the software system, containing the image of the

placement of components on individual nodes of the

network, are presented in the Figure 6. It also of-

fers information transfer routes between devices that

are involved in calculating the operational production

plan of a distributed enterprise.

Based on the above requirements for the organi-

zation of ﬂows and interaction of entities, it is rec-

ommended to follow the next sequence of operations,

presented in Figure 7, to ensure the conﬁdentiality of

information transmitted by the information owner to

the cloud for the implementation of the production

plan generation process.

The main structural element of the interaction

scheme between the information owner and the cloud

service shown in Figure 7 is the projective transfor-

mation, which for the most part, rests with the task

of ensuring the conﬁdentiality of information. The

methods underlying this protective conversion are de-

scribed in the next section of the paper.

Figure 6: Deployment Diagram of Operating Scheduling.

Figure 7: The sequence diagram of operations when calcu-

lating the operational production schedule using encryption.

Ensuring Conﬁdentiality of Information When Processing Operational Production Plans in Cloud Services

169

3 METHODS FOR ENSURING

CONFIDENTIALITY OF

INFORMATION FROM

TECHNOLOGICAL MAPS

DURING PROCESSING IN

EXTERNAL CLOUD SERVICES

Before describing the security conversion proposed in

the paper, it is worthwhile to clarify the relationship

between the described privacy methods and the ex-

isting secure connection protocols (SSL/TLS, IPsec,

HTTPS). The purposes of ensuring conﬁdentiality

proposed in the paper relating to the protocol inter-

action of the information owner’s node and the cloud

service at the application level, therefore, the task of

providing a secure connection between the cloud and

the owner is not considered and it is believed that it is

solved at lower levels (network and transport).

To describe the proposed protection method, we

consider the simplest example of a technological map,

an accompanying enterprise database, and the pro-

cess of downloading information from an enterprise

database to an external cloud to optimize critical pa-

rameters of an operational production plan. The table

1 presents the simplest example of a ﬂow chart for

manufacturing a metal part.

Table 1: Example of operational plan.

Number Sequence

1 Select a workpiece for processing

2 Mark the workpiece in length

3 Cut the workpiece along the marking lines

4 ···

··· ···

Based on technological maps, enterprises form

technological processes that involve the necessary

personnel, equipment, and additional tools to create

the materials indicated in the technological maps. The

most straightforward database accompanying the ﬂow

chart may can consist of attributes with following

clarifying information:

1. number of specialists in the enterprise;

2. required number of workpieces;

3. labor costs;

4. number of workbenches.

In the corporate database, information from tech-

nological maps is concertized, the necessary capac-

ities for creating the product are compared with the

current resources of the enterprise. An urgent task

for an enterprise with continually changing external

(interaction with suppliers, equipment and materials,

prices, logistics, production calendar) and internal

(availability of certain personnel and specialists) con-

ditions is the selection of the optimal ratio of critical

parameters (labor costs, time of employment of per-

sonnel and equipment consumed resources, etc.) in

order to improve the quality or quantity of products.

When assessing volumetric ﬂow charts and taking

into account a wide range of critical parameters, the

task of selecting the optimal ratio becomes computa-

tionally complex and, accordingly, intractable within

a particular enterprise. In this regard, a frequent prac-

tice at present is the transfer of the procedure for se-

lecting the optimal ratio of critical parameters to ex-

ternal public cloud services (Margun et al., 2017).

Despite the fact that the protection of conﬁdential-

ity from external intruders and third parties is solved

at the TLS/SSL protocol level, the actual threat of

conﬁdentiality violation is the threat of access by un-

scrupulous administrators and cloud owners to cus-

tomer data stored and processed in the cloud. The

obvious and primary problem when transferring in-

formation on limited-access outside the enterprise is

the preservation of its conﬁdentiality.

Consider the approach to ﬁnding the optimal key

parameters by solving the system of inequalities. This

approach is used, in particular, in the preparation of

information for generating a production plan. For ex-

ample, in a routing, there are two consecutive steps A

and B, in each of which two different specialists are

involved, the labor of which is required to complete

the corresponding steps as d

and d

. Specialists can-

not replace each other, and parallelization of A and B

operations cannot be performed. Under such condi-

tions, the task of optimizing labor costs for the enter-

prise will be the solution to the following inequality

systems:











= A

+ d

;

= B

+ d

;

<= B

(2)

The last inequality is necessary to accurately de-

termine the sequence of operations. This paper as-

sumes that the cloud assumes the function of ﬁnding

the optimal parameters by solving systems of inequal-

ities similar to 2.

For the methods of ensuring conﬁdentiality pro-

posed in this paper, preliminary data processing is re-

quired, which ranks the information to the one that

will participate in the calculations (structure or nu-

merical data), and the one that does not directly par-

ticipate in the estimates (tiles or character data). Char-

acter data includes not only explicitly deﬁned conﬁ-

dential information but also, for example, variables

for optimizing production plan parameters.

ICINCO 2020 - 17th International Conference on Informatics in Control, Automation and Robotics

170

The ﬁrst method proposed in this paper to ensure

the conﬁdentiality of information downloaded from

technological maps is to depersonalize key parame-

ters when transferring data to an external cloud ser-

vice. The following approach is proposed when pro-

cessing information before moving it to the cloud,

which consists of two successive stages: depersonal-

izing parameters (the depersonalization method must

be utterly reversible for the information owner and

preserving the data structure) and marking the data

types necessary for performing operations by the

cloud service. The most simple way of depersonaliz-

ing data is the method of introducing identiﬁers. This

approach provides for the replacement of critical in-

formation from a security point of view with a random

set of generated identiﬁers.

In the enterprise database, replacement tables are

additionally generated, the purpose of which is to de-

personalize the data transmitted to the cloud. For ex-

ample, to solve the system of inequalities 2, the fol-

lowing additional attributes can be entered in the en-

terprise database - Tables 2 and 3.

Table 2: Example of the substitutional table to character

parameters.

Number Sequence Notion

Forwarded

data

Select a workpiece

for processing

5DCBD2

Mark the workpiece

in length

6EB3BE

Cut the workpiece

along the marking lines

1A0623

Finish making

workpiece

2645B2

Table 3: Example of the substitutional table to numeric data.

Position Notion labor costs unit

type

identiﬁcation

Locksmith d

20 man hours 557C86

Turner d

10 man hours 5F0228

In the event that replacement tables are formed on

the client-side, as shown in Figure 2, only the last at-

tribute from the substitution table 2 and the second

and previous attribute from the substitution table 3 are

transferred to the external cloud service. As a result,

the cloud service receives the following inequality to

solve:











6EB3BE = 5DCBD2 + 20;

2645B2 = 1A0623 + 10;

6EB3BE <= 1A0623.

(3)

That is, the information transmitted to the cloud

server contains exclusively anonymized data (anon

function) and marking for data types, as shown in Ta-

ble 4.

Table 4: Example of data format forwarded to cloud service.

Formulation of

the problem

Description of

data types

anon(Sequence of

manufactoring steps)

Invariable

anon(Specialists

and labor)

Invariable

Only numerical data

with the number

of needed workpieces

integer

allowed range

Number of specialists

with speciﬁc

qualiﬁcation

Integer

exact value

It is worth noting that to increase the resistance

to correlation attacks, the method of anonymization

based on identiﬁers can be complicated and modi-

ﬁed, in particular, using other methods of anonymiza-

tion, for example, the technique of mixing data. In

some cases, the depersonalization method can be re-

placed by the full encryption of individual variable

blocks (encryption must be performed in electronic

codebook mode so that the same variables in the cloud

have the same ciphertext values). Encryption can be

indispensable if, in the corporate network of the enter-

prise for sending tasks and processing the results re-

ceived from the cloud, individual network nodes, for

example, located in different branches of the enter-

prise, are responsible.

Even though most of the data is en-

crypted/anonymized, cloud services have all the

necessary information to solve the system of inequal-

ities. The last attribute of the substitution Table 3

allows you to determine whether absolute numerical

values (20 and 10 in our speciﬁc example) have the

same units of measurement and whether operations

are allowed on them. Since when depersonaliz-

ing/encrypting the same values with restrictions

in the proposed method, the same ciphertexts are

obtained, the cloud service can perform ranking

and ordering of variables based on inequalities

additionally introduced into the system.

The obvious advantages of the ﬁrst method is the

simplicity of its implementation, we need to modify

the data structure used in the enterprise by adding re-

placement tables and identiﬁer arrays to mark vari-

ables and units of measure. Another advantage is that

in a cloud service, based on the ranges of values de-

ﬁned by the client, parallelization can be performed

on computing stations. When transmitting informa-

tion to the cloud, the owner of the information con-

Ensuring Conﬁdentiality of Information When Processing Operational Production Plans in Cloud Services

171

verts the blocks of data independently of each other,

so each block can be changed on the owner’s side

and re-sent to the cloud. Disadvantage of the pro-

posed method is that we transmit in an open form

(for the cloud owner) numerical values to an exter-

nal cloud service. Numeric values can be used by cy-

bercriminals (dishonest cloud server administrators)

to recover partial or complete information about the

internal business processes of an enterprise. Also, the

owner of the cloud will know the sequence of opera-

tions, the allowed ranges of values for the processed

parameters, which, in turn, can provide indirect infor-

mation about conﬁdential client data.

Conﬁdentiality of numerical data can be ensured,

provided that partially or fully homomorphic cryp-

tographic algorithms are used for their transmission

(Gentry, 2009a; Gentry, 2009b; Rohloff and Cousins,

2014). Thus, the second method of ensuring conﬁ-

dentiality is the use of fully homomorphic encryption

for numeric data transmitted to external services. The

following modiﬁcation of the ﬁrst proposed method

is proposed. Each numerical value x from the sec-

ond attribute of the Replacement Table 3 is assigned

a Enc(x) cost encrypted using a fully homomorphic

cryptosystem, which is subsequently transferred to

the cloud. As an encryption algorithm, this method

proposes to use the fully homomorphic cryptosystem

L. Jian and D. Song (Li et al., 2012), which is a

cryptosystem adapted to ensure conﬁdentiality in the

clouds. Let us brieﬂy describe the main stages of en-

cryption of the cryptosystem under consideration:

1. The key k is generated as a random odd number

of length l -bit

2. The process of encrypting one bit of the message

∈ 0, 1

is carried out according to the following formula

= x

+ k + r ∗ k ∗ q, (4)

where r is a random value of D-bit length, q is a

constant representing a large integer, n is the serial

number of the message bit and ciphertext.

3. To decrypt information on the client side, it is

enough to perform the operation:

c mod k (5)

The paper (Li et al., 2012) proves the homomor-

phism of the presented cryptosystem. An additional

signiﬁcant advantage of the presented method is the

ability to verify the integrity of data sent to the cloud

without decrypting it. In particular, when using the

proposed method, the information owner can send en-

crypted messages to the cloud at some intervals to ver-

ify the integrity of numerical data:

client

= x

client

+ k + r ∗ k ∗ q (6)

The server, in turn, is able, using only the general syn-

chronization parameter q, to verify that the selected

clear text bits do match:

Integrity = (c

server

− c

client

) mod q (7)

If the integrity parameter integrity is equal to zero,

then the copy stored on the server matches the sent

parameter.

Note that using a homomorphic system, it is pro-

posed to encrypt exclusively numerical data and not

information completely transmitted to the server, in-

cluding the statement of the problem, operating pa-

rameters, and identiﬁers. The encryption process it-

self, when it is used out of control, will be time-

consuming and labour-consuming for the information

owner. Therefore, it is proposed to further rank the

numerical parameters before sending it to the cloud

to determine which disclosure of which carries the

greatest risk to the enterprise. And concerning criti-

cal numerical parameters, apply the proposed method

based on a homomorphic encryption algorithm. Ad-

vantages As in the ﬁrst method, the cloud server will

be able to perform ordinary actions on numerical data

and even, as shown in the paper (Chialva and Doom,

2018), perform comparisons, ranking, dividing pa-

rameters into ranges under certain conditions. Using

selective encryption, you can achieve an effective bal-

ance between the load on the client and the level of

information conﬁdentiality. When using the second

method, the probability of violating the conﬁdential-

ity of information is very low. The attacker will only

know the sequence of operations, which is almost im-

possible to correlate with the technological processes

of the enterprise.

It is also worth highlighting drawbacks, charac-

teristic of the second method, on the client-side, it

is necessary to further modify the allowed type and

range for numeric and character parameters, because,

after encryption, the received value may not satisfy

the requirements deﬁned for the initial settings. Mod-

iﬁcations may concern, among other things, the state-

ment of the problem as a whole if, as a result of ap-

plying the second method, the optimization parameter

of the production schedule was encrypted.

4 CONCLUSIONS

The paper raises the urgent problem of ensuring the

conﬁdentiality of information when it is processed

in cloud services. In this paper, an analysis of in-

formation ﬂows between interacting entities was car-

ried out, a model of threats to information privacy

violations in solving the problem was constructed.

ICINCO 2020 - 17th International Conference on Informatics in Control, Automation and Robotics

172

The paper suggests using a comprehensive method for

anonymizing and marking conﬁdential information

with the preservation of the data type for subsequent

processing as a protective transformation for data

transferred to cloud services. As a second method,

it is proposed to strengthen the original algorithm for

the protective transformation using schemes of homo-

morphic encryption of numerical data. For each of the

proposed methods, the disadvantages and advantages

of its use and implementation are described.

ACKNOWLEDGEMENTS

This work was ﬁnancially supported by Government

of Russian Federation (Grant 08-08) and by the Min-

istry of Science and Higher Education of Russian

Federation, passport of goszadanie no. 2019-0898.

Section three funding was provided by Russian Sci-

ence Foundation (RU) (Grant No. 19-19-00566).

REFERENCES

Santosh kumar Singh, Pankaj Manjhi, Dr. V.R Vadi (2019).

Homomorphic. Encryption in Cloud Computing Se-

curity, Conference

C. R. Kavitha, B. Harsoor A survey on Homomorphic en-

cryption in cloud security, January 2018

Shrujana Murthy, Kavitha C.R. Preserving Data Pri-

vacy in Cloud using Homomorphic Encryp-

tion, Conference: 2019 3rd International con-

ference on Electronics, Communication and

Aerospace Technology (ICECA), June 2019 DOI:

10.1109/ICECA.2019.8822127

Changigarh S Manju Priya, J. Sumitha. Comparative Anal-

ysis of Homomorphic Encryption in Cloud Comput-

ing. Conference: 4th International Conference On Sci-

ence,Technology And Management, December 2018

Data in the clouds: leaks through unprotected servers in-

creased by 43%, https://www.infowatch.ru/company/

presscenter/news/15559, 2019 (accessed 22.10.19)

Future Factory: How Technology Is Transforming

Manufacturing:Resource Planning and Sourc-

ing/Operations Technology Monitoring and Ma-

chine Data, http://cresprit.com/en/2019/04/05/

future-factory-how-technology-is-transforming-

manufacturing2resource-planning-sourcing-

operations-technology-monitoring-machine-data/,

2019 (accessed 22.10.19)

Hong-Seok Park, N.-H. Tran, Development of a cloud based

smart manufacturing system, Journal of Advanced

Mechanical Design Systems and Manufacturing 9(3),

July 2015

Gartner Forecasts Worldwide Public Cloud Revenue to

Grow 17.5 Percent in 2019, https://www.gartner.com/

en/newsroom/press-releases/2019-04-02-gartner-

forecasts-worldwide-public-cloud-revenue-to-g, 2019

(accessed 22.10.19)

J. Li, D. Song, S. Chen and X. Lu (2012). A simple

fully homomorphic encryption scheme available in

cloud computing, 2012 IEEE 2nd International Con-

ference on Cloud Computing and Intelligence Sys-

tems, Hangzhou, pp. 214-217

C. Gentry (2009). A fully homomorphic encryp-

tion scheme, PhD thesis, Stanford University,

crypto.stanford.edu/craig.

C. Gentry (2019). Fully homomorphic encryption using

ideal lattices, Proceedings of the 41st Annual ACM

Symposium on Theory of Computing, pp. 169–178

K. Rohloff and D. B. Cousins (2014). A scalable imple-

mentation of fully homomorphic encryption built on

NTRU, Financial Cryptography and Data Security

Workshops, V. 8438 , pp. 221–234

D. Chialva, A. Dooms (2018). Conditionals in Homomor-

phic Encryption and Machine Learning Applications,

The International Association for Cryptologic Re-

search.

A. A. Margun, A. A. Bobtsov, I. B. Furtat (2017). Algorithm

to control linear plants with measurable quantized out-

put, Automation and Remote Control, V. 78, N. 5, pp.

826-835

Ensuring Conﬁdentiality of Information When Processing Operational Production Plans in Cloud Services

173