Enhanced RAM Simulation in Succinct Space

We describe two recent results on space-efficient functional random access memory (RAM), which is RAM with non-standard functionalities. The first is about oblivious RAM, which enables a remote database to be accessed without revealing to the database owner which part of the database is being accessed. The other is about wear leveling, which enables the number of updates to be balanced among all the memory cells regardless of the content of the computation being performed on the memory.

Oblivious stable sorting protocol and oblivious binary search protocol for secure multi-party computation

Multi-party computation (MPC) sorting and searching protocols are frequently used in different databases with varied applications, as in cooperative intrusion detection systems, private computation of set intersection and oblivious RAM. Ivan Damgard et al. have proposed two techniques i.e., bit-decomposition protocol and bit-wise less than protocol for MPC. These two protocols are used as building blocks and have proposed two oblivious MPC protocols. The proposed protocols are based on data-dependent algorithms such as insertion sort and binary search. The proposed multi-party sorting protocol takes the shares of the elements as input and outputs the shares of the elements in sorted order. The proposed protocol exhibits O ( 1 ) constant round complexity and O ( n log n ) communication complexity. The proposed multi-party binary search protocol takes two inputs. One is the shares of the elements in sorted order and the other one is the shares of the element to be searched. If the position of the search element exists, the protocol returns the corresponding shares, otherwise it returns shares of zero. The proposed multi-party binary search protocol exhibits O ( 1 ) round complexity and O ( n log n ) communication complexity. The proposed multi-party sorting protocol works better than the existing quicksort protocol when the input is in almost sorted order. The proposed multi-party searching protocol gives almost the same results, when compared to the general binary search algorithm.

SGX-MR: Regulating Dataflows for Protecting Access Patterns of Data-Intensive SGX Applications

Intel SGX has been a popular trusted execution environment (TEE) for protecting the integrity and confidentiality of applications running on untrusted platforms such as cloud. However, the access patterns of SGX-based programs can still be observed by adversaries, which may leak important information for successful attacks. Researchers have been experimenting with Oblivious RAM (ORAM) to address the privacy of access patterns. ORAM is a powerful low-level primitive that provides application-agnostic protection for any I/O operations, however, at a high cost. We find that some application-specific access patterns, such as sequential block I/O, do not provide additional information to adversaries. Others, such as sorting, can be replaced with specific oblivious algorithms that are more efficient than ORAM. The challenge is that developers may need to look into all the details of application-specific access patterns to design suitable solutions, which is time-consuming and error-prone. In this paper, we present the lightweight SGX based MapReduce (SGX-MR) approach that regulates the dataflow of data-intensive SGX applications for easier application-level access-pattern analysis and protection. It uses the MapReduce framework to cover a large class of data-intensive applications, and the entire framework can be implemented with a small memory footprint. With this framework, we have examined the stages of data processing, identified the access patterns that need protection, and designed corresponding efficient protection methods. Our experiments show that SGX-MR based applications are much more efficient than the ORAM-based implementations.

A Constant Round Write-Only ORAM

The write-only oblivious RAM (ORAM) is proposed to efficiently protect the privacy of applications such as cloud storage synchronization and encrypted hidden volumes. For N blocks with size B = Ω(log2N), the most efficient write-only ORAM, DetWoORAM, achieves O(B) communication complexity with O(logN) rounds per logical write. We propose a two-level write-only ORAM and achieve O(B) communication complexity with O(1) rounds. Similar to the traditional bucket-based ORAM schemes, we set a rate for the write operation to further reduce the communication complexity. The top-level stores data blocks in a flat array and the write pattern is protected by writing blocks uniformly at random. The second level employs a binary tree to store the position map of data blocks. To avoid recursive storage, a static position map for blocks in the second level is used. Both the analysis and experiments show that, besides the achieved low communication complexity and rounds, the stash sizes in the top level and the second level are bounded to O(B) and ω(B), respectively.

Identifying Influential Spreaders in a Social Network (While Preserving Privacy)

In order to disseminate information in a social network, it is important to first identify the influential spreaders in the network. Using them as the seed spreaders, the aim is to ensure that the information is cascaded throughout the network. The traditional approach to identifying influential nodes is to determine the top-r ranked nodes in accordance with various ranking methods such as PageRank, k-Shell decomposition, ClusterRank and VoteRank. In the current work, we study the problem of ranking the nodes when the underlying graph is distributedly held by a set of individuals, who consider their share of the data as private information. In particular, we design efficient secure multiparty computation (MPC) protocols for k-Shell decomposition, PageRank and VoteRank. For improved efficiency, we employ the oblivious RAM construct in conjunction with efficient data-oblivious graph data structures. We are the first to propose a secure variant of the VoteRank algorithm. We prove that the proposed protocols are asymptotically more efficient and have lower runtime in practice than the previous best known MPC protocols for computing k-Shell decomposition and PageRank centrality scores.