Ravi Chandra's Home Page

During the course of my research, I was involved in the following projects:

Businesses that provide data storage facilities on the internet (IDP) have exploded recently. Such businesses provide the following benefits to end users: a) anytime, anywhere access to data; b) low cost; and c) good quality of service. Examples of data storage providers include Amazon S3 service, Windows SkyDrive, Nirvarnix, etc. Users face two challenges in utilizing the storage infrastructures of the IDPs: a) Heterogeneity: Different IDPs provide different interfaces to application developers to store and fetch data with them due to lack of accepted standards; and b) Security: Data outsourced to IDPs is vulnerable to attacks from the internet thieves and from malicious employees of IDPs. To combat the above problems we present DataGuard, which is a client side interoperable security middleware that adapts to the heterogeneity of interfaces of IDP and enforces security constraints on outsourced data. This significantly simplifies the effort for application development. To combat heterogeneity, DataGuard incorporates an abstract service model that can be easily customized to several IDPs. To address the security challenges, DataGuard supports a security model that protects the confidentiality and integrity of outsourced data. We propose a novel indexing technique that allows search on the encrypted data stored at the IDPs. We illustrate the feasibility/efficacy of DataGuard, by implementing the middleware and executing it on two of the popular IDPs, Amazon S3 service and Gmail.com.

For more information regarding the project please visit http://DataGuard.ics.uci.edu . Here you will also find the prototype of the system available for free download.

Outsourcing personal information to web based data services (WDS) is a common recent trend. Websites such as Gmail.com, Flickr.com and Google Docs are examples of such services. Current WDSs provide a variety of data services that allow data storage, backup, sharing, etc. These services have become so popular that they are now a part of our daily life. The biggest drawback of such services is the requirement to trust the service provider. Data outsourced is stored in plaintext and is vulnerable to outsider attacks from hackers/Internet thieves and insider attacks from malicious employees.
To combat the above issues, we propose DataProtector, a security middleware for web based data services. DataProtector acts as a transparent middleman between the browser and the web server and enforces security requirements at the level of HTTP requests. DataProtector utilizes a rule based framework for determining the flow of sensitive information via HTTP requests/responses and protects the confidentiality and integrity of such information using cryptographic techniques. DataProtector is a generic middleware that can protect data outsourced to a wide variety of WDSs. We implemented a proof-of-concept DataProtector prototype to measure its performance.

For more information regarding the project please visit http://DataProtector.ics.uci.edu .

DataVault is an architecture designed for Web users that allows them to securely access their data from any machine connected to the Internet and also lets them selectively share their data with trusted peers. The DataVault architecture is built on the outsourced database model (ODB), where clients/users outsource their database to a remote service provider who provides data management services such backup, recovery, transportability and data sharing. In DataVault, the service provider is untrusted. The confidentiality and integrity of the user's data is preserved using cryptographic techniques. The service provider manages encrypted data. DataVault utilizes a novel PKI infrastructure and encrypted storage model that allow data sharing to take place via an untrusted server.

For more information regarding the project please visit http://ics.uci.edu/~rjammala/DataVault . Here you will also find the prototype of the system available for free download.

Pvault software is a personal data manager that stores and retrieves data from a remote untrusted data server securely. The major advantage of Pvault is that it allows users to access their personal data from any trusted remote computer. We will describe the issues and solutions for maintaining data confidentiality and integrity when the data is stored at the remote sever, since the server itself is untrusted. Pvault also prevents Phishing and Pharming attacks and we will describe the solutions for the same.

For more information regarding the project please visit http://www.itr-rescue.org/pVault/ . Here you will also find the prototype of the system available for free download.