188宝金博页面版

Performance trade-offs on a secure multi-party relationaldatabaseRogério PontesHASLab, INESC TEC &University of Minhorapontes@inesctec.ptMário Pintomariobpinto@gmail.comManuel BarbosaHASLab INESC TEC & DCCFCUPmbb@dcc.fc.up.ptRicardo VilaçaHASLab, INESC TEC &University of Minhormvilaça@di.uminho.ptMiguel MatosHASLab, INESC TEC &University of Minhomiguelmatos@di.uminho.ptRui OliveiraHASLab, INESC TEC &University of Minhorco@di.uminho.ptABSTRACTThe privacy of information is an increasing concern of soft-ware applications users. This concern was caused by attacksto cloud services over the last few years, that have leakedconf i dential information such as passwords, emails and evenprivate pictures. Once the information is leaked, the usersand software applications are powerless to contain the spreadof information and its misuse.With databases as a central component of applicationsthat store almost all of their data, they are one of the mostcommon targets of attacks. However, typical deploymentsof databases do not leverage security mechanisms to stopattacks and do not apply cryptographic schemes to pro-tect data. This issue has been tackled by multiple securedatabases that provide trade-of f s between security, querycapabilities and performance. Despite providing stronger se-curity guarantees, the proposed solutions still entrust theirdata to a single entity that can be corrupted or hacked.Secret sharing can solve this problem by dividing data inmultiple secrets and storing each secret at a dif f erent loca-tion. The division is done in such a way that if one locationis hacked, no information can be leaked. Depending on theprotocols used to divide data, functions can be computedover this data through secure protocols that do not discloseinformation or actually know which values are being calcu-lated.We propose a SQL database prototype capable of of f eringa trade-of f between security and query latency by using adif f erent secure protocol. An evaluation of the protocols isalso performed, showing that our most relaxed protocol hasan improvement of 5% on the query latency time over theoriginal protocol.Permission to make digital or hard copies of all or part of this work for personal orclassroom use is granted without fee provided that copies are not made or distributedfor prof i t or commercial advantage and that copies bear this notice and the full cita-tion on the f i rst page. Copyrights for components of this work owned by others thanACM must be honored. Abstracting with credit is permitted. To copy otherwise, or re-publish, to post on servers or to redistribute to lists, requires prior specif i c permissionand/or a fee. Request permissions from permissions@acm.org.SAC 2017, April 03-07, 2017, Marrakech, Moroccoc ? 2017 ACM. ISBN 978-1-4503-4486-9/17/04...$15.00DOI: http://dx.doi.org/10.1145/3019612.3019659CCS Concepts•Security and privacy → Data anonymization andsanitization; Management and querying of encrypteddata; •Information systems → Query operators;KeywordsSecure databases, Derby, Multi-Party computation, SecretSharing1. INTRODUCTIONCloud services are one of the best commodities of the mod-ern age, with an increasing interest by not only the generalpopulation but also from private and governmental sectors.From the vast myriad of providers, an interesting set of op-portunities are open to outsource storage and computationto services that have dif f erent degrees of scalability, availabil-ity, and reliability at an optimal benef i t-cost ratio. On theact of outsourcing personal information to an external entity,one must be aware that this information most likely is notkept private and secure from prying eyes. In fact recent me-dia scandals, regarding governmental surveillance programs,arose some trust issues regarding these cloud-based services.This concern on cloud security is one of the main con-tributing factors that inhibits most enterprises from movingtheir databases to a cloud service despite the many ben-ef i ts they would gain. In fact, bringing data privacy andsecure query execution to databases is an open challengeproposed to the database community [1]. The challenge isbeing tackled by multiple solutions that apply dif f erent en-cryption techniques to ensure the security of the database.Considering only solutions based on software, the existingwork can be divided into two main groups. The f i rst groupleverages multiple encryption techniques, where the secu-rity of these techniques is provided by computational hardproblems. This same group, can for a subset of the SQLlanguage provide secure queries. However to do so, all thedata is stored on a single location and the encryption tech-niques used have dif f erent security guarantees [9, 10]. Theother group of solutions is based on secret sharing schemesthat provide data privacy by dividing data in multiple se-crets stored at dif f erent locations and by using multi-partyprotocols to perform secure query computation [8, 11]. Themain advantage of this group of techniques is the use of a456