security for personal health record using attribute based ... · security for personal health...

National Conference on “Advanced Technologies in Computing and Networking"-ATCON-2015Special Issue of International Journal of Electronics, Communication & Soft Computing Science and Engineering, ISSN: 2277-9477

29

Security for Personal Health Record using Attribute BasedEncryption in Cloud Computing

Tejashri J. Madavi Dr. S. S. Sherekar Dr. V. M. Thakare

Abstract — Attribute based encryption determines encryptionability based on a user’s attributes. In a multi-authority ABEScheme, multiple attribute-authorities monitor different sets ofattributes and issue corresponding decryption keys to user andencryptor can require that a user obtain keys for appropriateattributes from each authority before decrypting a message.Personal Health Record (PHR) is maintained in the centralizedserver to maintain the patient’s information. The patient recordsshould be maintained with high privacy and security. The securityschemes are used to protect the personal data from public access.This paper proposes a novel patient-centric framework and asuite of data access mechanisms to control PHRs stored in semi-trusted servers. To achieve fine-grained and scalable data accesscontrol for PHRs, an Attribute Based Encryption (ABE)techniques to encrypt each patient’s PHR file. Multiple dataowners can access the same data values. The proposed schemecould be extended to Multi Authority Attribute Based Encryption(MA-ABE) for multiple authority based access controlmechanism.

Keywords — Personal Health Records, Cloud Computing, DataPrivacy, Fine-grained access control, Multi-authority AttributeBased Encryption

I. INTRODUCTIONCloud computing is a computing environment where

in computing resources which may be hardware or software aredelivered as a service via a network. Cloud computing entitlesresource sharing to achieve best utility over a network. ThePersonal Health Record (PHR) sharing among a wide range ofpersonnel has been identified as an evident application in thefield of cloud computing [6]. A personal health record is ahealth record where health data and information related to thecare of a patient is maintained by the patient himself. Thepurpose of PHR is to provide accurate medical details aboutthe patient, which can be accessed online also. PHR can covera wide variety of information including prescription report,family history, allergy details, and laboratory test results and soon.

Personal health record has emerged as a patient-centric model of health information exchange. It enables thepatient to create and control her medical data which may beplaced in a single place such as data center, from where accesscan be made by different individuals [7]. Due to the high costof building and maintaining specialized datacenters, manyPHR services are outsourced to or provided by third-partyservice providers, for example, Microsoft Health Vault, GoogleHealth.

Personal Health Record (PHR) is maintained in thecentralized server to maintain the patient’s information. Thepatient records should be maintained with high privacy andsecurity. The security schemes are used to protect the personaldata from public access. Patient data can be accessed by manydifferent people [9]. Each authority is assigned with accesspermission for a particular set of attributes. The access controland privacy management is a complex task in the patient healthrecord management process. Cloud computing is a colloquialexpression used to describe a variety of different types ofcomputing concepts that involve a large number of computersthat are connected through a real-time communication network.It is a synonym for distributed computing over a network andmeans the ability to run a program on many connectedcomputers at the same time. Data owners update the personaldata into third party cloud data centres.

In this paper, propose a novel patient-centricframework and a suite of data access mechanisms to controlPHRs stored in semi-trusted servers. To achieve fine-grainedand scalable data access control for PHRs, an Attribute BasedEncryption (ABE) technique to encrypt each patient’s PHR file[10]. Multiple data owners can access the same data values.The proposed scheme could be extended to Multi AuthorityAttribute Based Encryption (MA-ABE) for multiple authoritybased access control mechanism.

II. BACKGROUND

This paper is mostly related to work incryptographically enforced access control for outsourced dataand attribute based encryption. To improve upon the scalabilityof the above solutions, one-to-many encryption methods suchas ABE can be used [1]. A fundamental property of ABE ispreventing against user collusion. In addition, the encryptor isnot required to know the ACL.

A) Trusted authorityA number of works used ABE to realize fine-grained

access control for outsourced data [3] [4]. Recently, Narayan etal. proposed an attributebased infrastructure for EHRsystems, where each patient’s EHR files are encrypted using abroadcast variant of CP-ABE [5] that allows direct revocation.

There are several common drawbacks of the aboveworks. First, they usually assume the use of a single Trusted


30

Authority (TA) in the system. This not only may create a loadbottleneck, but also suffers from the key escrow problem. Inaddition, it is not practical to delegate all attribute managementtasks to one TA, including certifying all users’ attributes orroles and generating secret keys.

B) Attribute Based EncryptionIt is a well-known challenging problem to revoke

users/attributes efficiently and on-demand in ABE.Traditionally this is often done by the authority broadcastingperiodic key updates to unrevoked users frequently [6], [7],which does not achieve complete backward/forward securityand is less efficient. In this paper, bridge the above gaps byproposing a unified security framework for patient-centricsharing of PHR in a multi-domain, multi-authority PHR systemwith many users. The framework captures application levelrequirements of both public and personal use of a patient’sPHR and distributes users’ trust to multiple authorities thatbetter reflects reality.

III. PREVIOUS WORK DONE

Zhiwei et al. [1] proposed a watermarking method inthe architecture of cloud computing, to mitigate the risks ofinsider disclosures. Private health information once confined tolocal medical institutions is migrating onto the Internet as anElectronic Health Record (EHR) that is accessed by cloudcomputing. The author proposed an insider threat model for thehealth data storage in the cloud, finding the security gaps. Thecloud-based watermarking method including architecture andimplementation, for enhancing cloud security. Rob et al. [2]describe a study in the domain of health informatics whichincludes some novel requirements for patient confidentiality inthe context of medical health research. The author present aprototype which takes health records from a commercial dataprovider, anonymises them in an innovative way and makesthem available within a secure cloud-based Virtual ResearchEnvironment (VRE). Data anonymity is tailored as required forindividual researchers’ needs and ethics committee approval.VREs are dynamically configured to model each researcher’spersonal research environment while maintaining dataintegrity, provenance generation and patient confidentiality.Sven et al. [3] proposed privacy in business processes forproviding personalized services is currently a matter of trust.Business processes require the disclosure of personal data tothird parties and users are not able to control their usage and sotheir further disclosure. The current work on usage controlmainly considers formalization of usage rules. The authorinvestigate on digital watermarking as a way of enforcingobligations for further disclosure of personal data withoutmandatory trust in service providers. Yuyu et al. [4] proposednew SaaS Confidentiality Risk Management Framework basedon literature research, expert talks and the working experience.It enhances the client side confidentiality management in apublic SaaS integrated IT environment and especially focuseson small to medium sized enterprises, which are oftenconfronted with rigid contracts enforced by cloud serviceproviders and have weak and lacking ability to negotiateService Level Agreements (SLAs). Tatiana et al. [5] proposedEHR as a subset of electronic medical record shared acrosshealth centers (HC) by medical workers. The cloud computing

approach does not just provide adequate data storage capacitiesand facilitate storing of health data in one centralized place.

IV. ATTRIBUTE-BASED ENCYPTION

In ABE system, users’ private keys and cipher text arelabelled with sets of descriptive attributes and access policiesrespectively, and a particular key can decrypt a particularcipher text only if associated attributes and policy are matched.A) Key-Policy Attribute-Based Encryption (KP-ABE):

It was introduced. In this cryptography system, ciphertext is labelled with sets of attributes. Private keys, on the otherhand, are associated with access structures A. A private keycan only decrypt a cipher text whose attributes set is authorizedset of the private key’s access structure. KP-ABE is acryptography system built upon bilinear map and Linear SecretSharing Schemes [9].

B) Multi-Authority attribute-Based encryption:In a multi-authority ABE system, it have many attributeauthorities, and many users. There are also a set of systemwide public parameters available to everyone (either created bya distributed protocol between the authorities). A user canchoose to go to an attribute authority, prove that it is entitled tosome of the attributes handled by that authority, and requestthe corresponding decryption keys. The authority will run theattribute key generation algorithm, and return the result to theuser. Any party can also choose to encrypt a message, in whichcase he uses the public parameters together with an attribute setof his choice to form the cipher text. Any user who hasdecryption keys corresponding to an appropriate attribute setcan use them for decryption [12].

V. PROPOSED METHODOLOGY

The main goal of the system is to provide secureaccess of PHR in a patient-centric manner and efficient keymanagement. First, the system is divided into multiple securitydomains like Personal domain (PSD) and Public domain(PUD). Each domain controls only a subset of its users. Foreach security domain, one or more authorities are assigned togovern the access of data. For personal domain it is the ownerof the PHR itself who manages the record and performs keymanagement. This is less laborious since the number of usersin the personal domain is comparatively less and is personallyconnected to the owner. Public domain consists of a largenumber of professional users and therefore cannot be managedeasily by the owner herself. Hence it puts forward the new setof public Attribute Authorities (AA) to govern disjoint subsetof attributes.

In this framework, there are multiple SDs, multipleowners, multiple AAs, and multiple users. In addition twoABE systems are involved: for each PSD the YWRL’srevocable KP-ABE scheme is adopted; for each PUD, thisproposed revocable MA-ABE scheme. Each data owner is atrusted authority of their own PSD, who uses a KP-ABEsystem to manage the secret keys and access rights of users intheir PSD. Secondly, so as to achieve security of healthrecords, a new encryption pattern namely Attribute basedencryption (ABE) is adopted. Data is classified according totheir attributes. In certain cases, users may also be classifiedaccordingly into roles. PHR owner encrypts their record under


31

a selected set of attributes and those users that satisfy thoseattributes can obtain decryption key in order to access the data.However, in the new solution pattern, an advanced version ofABE called multi-authority ABE (MA-ABE) is used. In thisencryption scheme, many attribute authorities operatesimultaneously, each handing out secret keys for a different setof attributes. A Multi-Authority ABE system is comprised of kattribute authorities and one central authority.

Each attribute authority is also assigned a value, dk.The system uses the following algorithms:

A) Set up: A random algorithm that is run by the centralauthority or some other trusted authority. It takes as input thesecurity parameter and outputs a public key, secret key pair foreach of the attribute authorities, and also outputs a systempublic key and master secret key which will be used by thecentral authority.

B) Attribute Key Generation: A random algorithm run by anattribute authority. It takes as input the authority’s secret key,the authority’s value dk, a user’s GID, and a set of attributes inthe authority’s domain and output secret key for the user.

C) Central Key Generation: A randomized algorithm that isrun by the central authority. It takes as input the master secretkey and a user’s GID and outputs secret key for the user.

D) Encryption: A randomized algorithm runs by a sender. Ittakes as input a set of attributes for each authority, a message,and the system public key and outputs the cipher text.

E) Decryption: A deterministic algorithm runs by a user. Ittakes input a cipher-text, which was encrypted under attributeset and decryption keys for that attribute set. This algorithmoutputs a message m.

Using ABE and MA-ABE which enhances the systemscalability, there are some limitations in the practicality ofusing them in building PHR systems.

Fig.1. Architecture of patient record sharing

VI. RESULT & ANALYSIS

The system is design to manage Personal HealthRecords (PHR) with different user access environment. Thedata values are maintained under a third party cloud providersystem. The data privacy and security is assured by the system.The privacy attributes are selected by the patients. The data canbe accessed by different parties. The key values are maintainedand distributed to the authorities. The system is enhanced tosupport Distributed ABE model. The user identity based accessmechanism is also provided in the system. The system isdivided into six major modules. They are data owner, cloudprovider, key management, security process, authority analysisand client.

A) Data Owner: The data owner module is designed tomaintain the patient details. The attribute selection model isused to select sensitive attributes. Patient Health Records(PHR) is maintained with different attribute collections. Dataowner assigns access permissions to various authorities.

B) Cloud Provider: The cloud provider module is used to storethe PHR values. The PHR values are stored in databases. Dataowner uploads the encrypted PHR to the cloud providers. Useraccess information's are also maintained under the cloudprovider.

C) Key Management: The key management module is designedto manage key values for different authorities. Key values areuploaded by the data owners. Key management processincludes key insert and key revocation tasks. Dynamic policybased key management scheme is used in the system.

D) Security Process: The security process handles theAttribute Based Encryption operations. Different encryptiontasks are carried out for each authority. Attribute groups areused to allow role based access. Data decryption is performedunder the user environment.

E) Authority Analysis: Authority analysis module is designedto verify the users with their roles. Authority permissions areinitiated by the data owners. Authority based key values areissued by the key management server. The key and associatedattributes are provided by the central authority.

F) Client: The client module is used to access the patients.Personal and professional access models are used in thesystem. Access category is used to provide different attributes.The client access log maintains the user request informationfor auditing process.

CONCLUSIONIn this Paper, presented the detail design and detail of

proposed a novel framework of secure sharing of personalmedical records in cloud computing. Considering partiallytrustworthy cloud servers, that to fully realize the patient-centric concept, patients shall have complete control of theirown privacy through encrypting their medical record files toallow fine-grained access.

A framework of secure sharing of personal healthrecords has been proposed in this paper. Public and Personalaccess models are designed with security and privacy enabled


32

mechanism. The framework addresses the unique challengesbrought by multiple PHR owners and users, in that thecomplexity of key management is greatly reduced. Theattribute-based encryption model is enhanced to supportoperations with MA-ABE. The system is improved to supportdynamic policy management model. Thus, Personal HealthRecords are maintained with security and privacy.

REFERENCES1] Zhiwei Yu, Clark Thomborson, Chaokun Wang, Jianmin Wang and RuiLi.”A Cloud-Based Watermarking Method for Health Data Security”, IEEEpaper 2012 ISBN: 978-1-4673-2362-8/12.2] Rob Smith, Jie Xu, Saman Hima & Dr. Owen Johnson.”GATEway to theCloud”, IEEE Computer Society 2013 IEEE Seventh International Symposiumon Service-Oriented System Engineering, DOI 10.1109/SOSE.2013,46.3] Sven Wohlgemuth, Isao Echizen and Noboru Sonehara.”On Privacy-Compliant Dislosure of Personal Data to Third Parties using DigitalWatermarking”, Journal of International Hiding and Multimedia SignalProcessing VOL 2, NO. 3, pp 2073-4212, July 2011.4] Yuyu Chou, Olga Levina, Jan Oetting.”Enforcing Confidentiality in a SaaSCloud Environment”, IEEE paper 2011 19th Telecommunications forumTELFOR, ISBN: 978-1-4577-1500-6/11.5] Tatiana Ermakova and Benjamin Fabian.”Secret Sharing for Health Data inMulti-Provider Clouds”, IEEE Computer Society 2013 IEEE InternationalConference on Business Informatics, DOI 10.1109/CBI.2013.22.6] Ming Li, Shucheng Yu, and Wenjing Lou, “Scalable and Secure Sharing ofPersonal Health Records in Cloud Computing using Attribute basedEncryption”, IEEE Transactions On Parallel And Distributed Systems 2012.7] S.Yu, C.Wang, K. Ren, and W. Lou, “Achieving secure, scalable, and fine-grained data access control in cloud computing,” in IEEE INFOCOM’10,2010.8] J. Benaloh, M. Chase, E. Horvitz, and K. Lauter, “Patient controlledencryption: ensuring privacy of electronic medical records,” in CCSW’09,2009, pp. 103–114.9] Y. Zheng, “Privacy-Preserving Personal Health Record System UsingAttribute-Based Encryption,” master’sthesis, Worcester Polytechnic Inst., 2011.10] S. Narayan, M. Gagne ́, and R. Safavi-Naini, “Privacy Preserving EHRSystem Using Attribute-Based Infrastructure,” Proc. ACM Cloud ComputingSecurity Workshop (CCSW ’10), pp. 47-52, 2010.11] J. Hur and D.K. Noh, “Attribute-Based Access Control with EfficientRevocation in Data Outsourcing Systems,” IEEE Trans. Parallel andDistributed Systems, vol. 22, no. 7, pp. 1214-1221, July 2011.12] S. Ruj, A. Nayak, and I. Stojmenovic, “Dacc: Distributed access control inclouds,” in 10th IEEE TrustCom, 2011.

AUTHOR’S PROFILE

Tejashri J. MadaviT. J. Madavi has completed B.E. Degree in ComputerScience and Engineering from Sant Gadge Baba AmravatiUniversity, Amravati, Maharashtra. She is persuing MastersDegree in Computer Science and Information Technologyfrom P.G. Department of Computer Science andEngineering, S.G.B.A.U. Amravati.

Dr. Swati Sherekar received the degree ofM.Sc. and Ph.D in computer science fromSGB Amravati University, Amravati.Presently working as Associate professor inthe P. G. Department of Computer Scienceand Engg. and having 19 years of teachingexperience. Her area of research is Networksecurity, data security, Image Processing andcompleted her Ph.D. in multimediaauthentication. Completed one MRP. Numberof papers are on her credits at National &International level journals and conferences.

Dr. V. M. ThakareDr. Vilas M. Thakare is Professor and Head in PostGraduate department of Computer Science and engg,Faculty of Engineering & Technology, SGB Amravatiuniversity, Amravati. He is also working as a co-ordinator on UGC sponsored scheme of e-learning andm-learning specially designed for teaching and research.He is Ph.D. in Computer Science/Engg and completedM.E. in year 1989 and graduated in 1984-85.He has exhibited meritorious performance in hisstudentship. He has more than 27 years of experience inteaching and research. Throughout his teaching career hehas taught more than 50 subjects at various UG and PGlevel courses. He has done his PhD in area of robotics, AIand computer architecture. 5 candidates have completedPhD under his supervision and more than 8 are perusingthe PhD at national and international level. His area ofresearch is Computer Architectures, AI and IT. He hascompleted one UGC research project on "Development ofES for control of 4 legged robot device model.". OneUGC research project is ongoing under innovativescheme. At PG level also he has guided more than 300projects/discretion. He has published more than 150papers in International & National level Journals and alsoInternational Conferences and National levelConferences. He has also successfully completed theSoftware Development & Computerization of Finance,Library, Exam, Admission Process, Revaluation Processof Amravati University. Also completed the Consultancywork for election data processing . He has also worked asmember of Academic Council, selection Committeemember of various Other University and parentuniversity, Member of faculty of Engineering & Science,BOS (Comp. Sci.), Member of IT Committee, Member ofNetworking Committee, Member of UGC, AICTE,NAAC, BUTR, ASU, DRC, RRC, SEC, CAS, NSD etccommittees. .

security for personal health record using attribute based ... · security for personal health...

Documents