云通信中密文访问控制的可撤销技术

This project aims to study the revocation techniques of ciphertext access control in the cloud communication. Cloud storage is one of the novel techniques deriving from cloud computing, which serves the function as providing the online data storage for the users by the cooperation of various kinds of storage devices. Different from the traditional mode where users themselves take charge in the control of data storage and computing, the cloud storage depends on the cloud service provider instead. Therefore, how to protect the data confidentiality and the efficient access control from sophisticated attacks has increasingly been an important topic. Attribute-based encryption provides the fine-grained access control and secret key revocation facilitates the flexibility in user management. Therefore, the combination of these two techniques can introduce a series of novel access control methods and higher security levels. Unfortunately, it brings about another unaddressed problem of secret key leakage and compromise, requiring us to prohibit the illegal users from continuously accessing the sensitive data in the cloud storage. Consequently, revocable attribute-based encryption which can be classified into two categories named authorized revocation and unauthorized revocation is a valuable and extremely challenging topic. However, the existing solutions possess their own shortcomings in whatever respect of low security levels and large overhead and rare practical techniques are proposed to perfectly address this problem. Revocable attribute-based encryption was firstly proposed all over the world by our research group with the paper published in ASIACCS 2009 focusing on the situation of authorized revocation. Then, our group and Waters' group respectively studied the unauthorized revocation under two different models in 2011 and 2012. Unfortunately, there still exist the following critical problems unaddressed: (1) How to simulate the decryption oracle to achieve adaptive CCA security in the authorized revocation? (2) How to achieve adaptive CCA security in the standard model in the unauthorized revocation? (3) How to design the attribute-based encryption schemes simultaneously satisfying the properties of revocability and traceability? Traceability refers to the ability to trace the identities of users who maliciously delegate their secret keys or whose secret keys are compromised by the adversary. (4) How to solve the problem of attribute revocation in the multi-authority setting? This project plans to study the abovementioned aspects and tries to make a breakthrough on these challenging problems. It can significantly stimulate the application of the ciphertext access control in the cloud storage and guarantee the security, confidentiality and efficient access control of the users' data in the cloud communication system.

本项目主要研究云通信中用户访问存储密文的权限撤销技术。属性基加密可以为存储密文提供细粒度的访问控制策略，而私钥或属性撤销则有助于灵活的管理用户权限，这两者结合可以为云通信的密文访问控制提供丰富的访问策略和更高的安全级别。可撤销的属性基加密第一篇文章起源于2009年我们发表在ASIACCS 2009上的论文，我们研究的是有授权的情形。2011年和2012年我们团队和Waters团队分别在两个不同模型下研究了无授权的情形。但是，下面几个关键问题仍然没有解决：（1）有授权的可撤销技术中，如何模拟解密语言机将安全性提高到CCA安全？（2）在无授权的可撤销技术中，如何达到标准模型下CCA安全？（3）如何设计可追踪且可撤销的属性基加密方案？可追踪是指系统能够追踪用户的私钥泄露或被泄露。（4）如何解决多机构情形下属性撤销问题？本项目将研究这些问题，并力争取得突破，推进云通信中密文访问控制的实用化。

本项目系统地研究了云通信中用户访问存储密文的权限撤销技术，取得了一系列重要突破，发表49篇学术论文，其中SCI检索23篇，EI检索49篇，有13篇属于CCF权威期刊/权威会议A类或JCR一区论文，包括TDSC, TIFS, TPDS等期刊和INFOCOM等会议长文。项目负责人和主要研究人员还应邀在国际会议、全国会议以及在高校、科研院所或省级学会做邀请报告30余次，获得美国和中国发明专利授权3项。.本项目主要创新性工作包括：.（1）解决了构建授权、未授权等多种撤销功能的密文数据访问控制机制问题，提出了授权路径自治加密机制，数据包合作传输的证据隐私在随机预言机模型下达到CCA2安全，安全的细粒度多文件共享方案在标准模型下达到CCA2安全。.（2）解决了可追踪方案的可撤销问题、多机构方案的属性撤销问题，从而构建了同时具有白盒可追踪、可撤销的多权威机构属性基加密方案。完全解决了如下两个问题：设计可追踪且可撤销的属性基加密方案和多机构情形下属性撤销问题。.（3）提出了世界首创的零开销追踪机制和云数据隐私保护的同态数据封装方法（不依赖（全）同态加密且只使用一次公钥加密的完全数据隐私保护技术），并给出了对基于云的安全计算和隐私保护的应用。.此外，本项目还扩展了许多研究，对云通信中用户访问存储密文的权限撤销技术的衍生做出了贡献。.受资助期间，团队获得了教育部自然科学二等奖和全国党政密码科学技术二等奖，分别有一人获得上海市领军人才和ACM上海优秀博士学位论文，还培养毕业5名博士、10名硕士以及3名博士后出站，相信这是一支推进云通信中密文访问控制实用化的生力军。