رسائل ماجيستير
English
2014
Data security in cloud storage services
Mai Mansour Dahshan, Sherif El-Kassass
كلية العلوم والهندسة-الجامعة الأمريكية بالقاهرة · مصر
Cloud Computing is considered to be the next-generation architecture for ICT where it moves the application software and databases to the centralized large data centers. It aims to offer elastic IT services where clients can benefit from significant cost savings of the pay-per-use model and can easily scale up or down, and do not have to make large investments in new hardware. However, the management of the data and services in this cloud model is under the control of the provider. Consequently, the cloud clients have less control over their outsourced data and they have to trust cloud service provider to protect their data and infrastructure from both external and internal attacks. This is especially true with cloud storage services. Nowadays, users rely on cloud storage as it offers cheap and unlimited data storage that is available for use by multiple devices (e.g. smart phones, tablets, notebooks, etc.). Besides famous cloud storage providers, such as Amazon, Google, and Microsoft, more and more third-party cloud storage service providers are emerging. These services are dedicated to offering more accessible and user friendly storage services to cloud customers. Examples of these services include Dropbox, Box.net, Sparkleshare, UbuntuOne or JungleDisk. These cloud storage services deliver a very simple interface on top of the cloud storage provided by storage service providers. File and folder synchronization between different machines, sharing files and folders with other users, file versioning as well as automated backups are the key functionalities of these emerging cloud storage services. Cloud storage services have changed the way users manage and interact with data outsourced to public providers. With these services, multiple subscribers can collaboratively work and share data without concerns about their data consistency, availability and reliability. Although these cloud storage services offer attractive features, many customers have not adopted these services. Since data stored in these services is under the control of service providers resulting in confidentiality and security concerns and risks. Therefore, using cloud storage services for storing valuable data depends mainly on whether the service provider can offer sufficient security and assurance to meet client requirements. From the way most cloud storage services are constructed, we can notice that these storage services do not provide users with sufficient levels of security leading to an inherent risk on users' data from external and internal attacks. These attacks take the form of: data exposure (lack of data confidentiality); data tampering (lack of data integrity); and denial of data (lack of data availability) by third parties on the cloud or by the cloud provider himself. Therefore, the cloud storage services should ensure the data confidentiality in the following state: data in motion (while transmitting over networks), data at rest (when stored at provider's disks). To address the above concerns, confidentiality and access controllability of outsourced data with strong cryptographic guarantee should be maintained. To ensure data confidentiality in public cloud storage services, data should be encrypted data before it is outsourced to these services. Although, users can rely on client side cloud storage services or software encryption tools for encrypting user's data; however, many of these services fail to achieve data confidentiality. Box, for example, does not encrypt user files via SSL and within Box servers. Client side cloud storage services can intentionally/unintentionally disclose user decryption keys to its provider. In addition, some cloud storage services support convergent encryption for encrypting users' data exposing it to “confirmation of a file attack.” On the other hand, software encryption tools use full-disk encryption (FDE) which is not feasible for cloud-based file sharing services, because it encrypts the data as virtual hard disks. Although encryption can ensure data confidentiality; however, it fails to achieve fine-grained access control over outsourced data. Since, public cloud storage services are managed by un-trusted cloud service provider, secure and efficient fine-grained access control cannot be realized through these services as these policies are managed by storage services that have full control over the sharing process. Therefore, there is not any guarantee that they will provide good means for efficient and secure sharing and they can also deduce confidential information about the outsourced data and users' personal information. In this work, we would like to improve the currently employed security measures for securing data in cloud store services. To achieve better data confidentiality for data stored in the cloud without relying on cloud service providers (CSPs) or putting any burden on users, in this thesis, we designed a secure cloud storage system framework that simultaneously achieves data confidentiality, fine-grained access control on encrypted data and scalable user revocation. This framework is built on a third part trusted (TTP) service that can be employed either locally on users' machine or premises, or remotely on top of cloud storage services. This service shall encrypts users data before uploading it to the cloud and decrypts it after downloading from the cloud; therefore, it remove the burden of storing, managing and maintaining encryption/decryption keys from data owner's. In addition, this service only retains user's secret key(s) not data. Moreover, to ensure high security for these keys, it stores them on hardware device. Furthermore, this service combines multi-authority ciphertext policy attribute-based encryption (CP-ABE) and attribute-based Signature (ABS) for achieving many-read-many-write fine-grained data access control on storage services. Moreover, it efficiently revokes users’ privileges without relying on the data owner for re-encrypting massive amounts of data and re-distributing the new keys to the authorized users. It removes the heavy computation of re-encryption from users and delegates this task to the cloud service provider (CSP) proxy servers. These proxy servers achieve flexible and efficient re-encryption without revealing underlying data to the cloud. In our designed architecture, we addressed the problem of ensuring data confidentiality against cloud and against accesses beyond authorized rights. To resolve these issues, we designed a trusted third party (TTP) service that is in charge of storing data in an encrypted format in the cloud. To improve the efficiency of the designed architecture, the service allows the users to choose the level of severity of the data and according to this level different encryption algorithms are employed. To achieve many-read-many-write fine grained access control, we merge two algorithms (multi-authority ciphertext policy attribute-based encryption (MA- CP-ABE) and attribute-based Signature (ABS)). Moreover, we support two levels of revocation: user and attribute revocation so that we can comply with the collaborative environment. Last but not least, we validate the effectiveness of our design by carrying out a detailed security analysis. This analysis shall prove the correctness of our design in terms of data confidentiality each stage of user interaction with the cloud.
