The Modares Journal of Electrical Engineering
Volume:16 Issue: 3, 2016

The unprecedented growth of internet contents, specially social media, makes a challenge to the load of cellular networks. Latency is one of the most important metrics at end-users. To this end, we propose a resource allocation (RA) algorithm to design both caching and delivery policies with the aim of minimizing total latency of mobile users (MUs), where in the caching phase, the content placement is investigated and in the delivery phase, we allocate radio resources (i.e., transmit powers and subcarriers) in a multi-cell orthogonal frequency division multiple access (OFDMA)-based network communicating with a data center via backhaul links.
In order to achieve an efficient caching policy, we propose an optimization problem to minimize the latency of MUs subject to maximum delivery deadline, maximum allowable transmit power of each base station (BS) and data center, and exclusive subcarrier assignment constraints. Hence, we devise an iterative algorithm to solve the main optimization problem and prove that the proposed approach converges to a near-optimal solution, when the number of iterations increases. Moreover, simulation results illustrate that devising the transmission-aware caching policy can significantly improve the performance, compared to the conventional proactive caching policies which are only based on the popularity of contents and the storage capacity of BSs.

Secure outsourcing is essential to growth of cloud usage. There are some protocols allowing any functionality to be outsourced. However, specific constructions are necessary in order to do so in an efficient way. In this paper, we consider the problem of secure outsourced pattern matching. Our solution is based on Bit-Parallel Shift-ADD algorithm. The properties of this insecure algorithm allow our construction to search in an outsourced text, without revealing any non-trivial information to the computing server. We achieve a round optimal protocol that allows us to search for patterns with wildcards and handles the Hamming distance computation. Since the protocol has no leakage to the server, it cannot be optimal considering communication complexity; however, we suggest efficient techniques to achieve communication optimality through outsourcing of decryption as well. The security of our protocol is proved in the semi-honest setting. Then, in order to retain the efficiency of the protocol, we omit the correctness property in the malicious setting and prove that the scheme remains private in the presence of malicious adversaries.

3D is a secret-key block cipher, designed to secure and fast encryption of large amounts of data. This block cipher
uses multi-dimensional states to generalize the design of Rijndael. Thus, while maintaining the benefits of the AES design, 3D operates on 512-bit blocks of data and can also be used as a cryptographic primitive in the cryptographic systems with the large internal states. Since its proposal in 2008, the cryptanalysis of 3D has been considered in several papers. While the previous impossible differential attacks on 3D cipher can analyze up to 10 rounds of the cipher, this paper, using a new 6-round impossible differential, presents an impossible differential attack on 11 rounds of 3D. The proposed distinguisher begins in the input of AddRoundKey operation of round 3, and ends in the output of ShiftRows of round 8. Results show that the proposed attack on 11-round of 3D cipher requires about 2501 chosen plaintexts and a time complexity of about 2495 11-round encryptions.

Oblivious transfer (OT) is a basic building block in many cryptographic protocols. A common approach in designing secure multiparty computation protocols is to assume that messages of the protocol are being transmitted over an authenticated channel, where entities have been authenticated to each other before the actual flows of the protocol. However, the mentioned aspect leads to some restrictions in design and development of secure multiparty computations. In this paper, we exploit some well-known authenticated Diffie-Hellman-based key exchange protocols to build three authenticated 1-out-of-2 oblivious transfer protocols. As a result, our schemes incorporate the authentication within the oblivious transfer protocol itself, instead of performing authentication via a separate sub-protocol. We show that the proposed protocols are secure in the semi-honest model. We also compare our new schemes with the previous methods (performing authentication via a separate sub-protocol) which illustrates that our schemes decrease computational and communication complexity for both sender and receiver.

Counting the minimum number of differential active S-boxes is a common way to evaluate the security of block ciphers against differential and linear cryptanalysis. In this paper, we use mixed-integer linear programming (MILP) to calculate minimum number of active S-boxes of the some Feistel structures. We focus on Type-II of Feistel structures with four and six partitions and explain how to analyze them by MILP when they have more than one MDS2 matrices (like Clefia) in their structure. Moreover, we propose a new four partitions Feistel structure with three multiple MDS matrices which have more active S-boxes rather than Clefia structure. We also generalize Clefia structure in to six partitions Feistel structure by three multiple MDS matrices for 192 bits block size.

During the past few years, the number of malware designed for Android devices has increased dramatically. To confront with Android malware, some anomaly detection techniques have been proposed that are able to detect zero-day malware, but they often produce many false alarms that make them impractical for real-world use. In this paper, we address this problem by presenting DroidNMD, an ensemble-based anomaly detection technique that focuses on the network behavior of Android applications in order to detect Android malware. DroidNMD constructs an ensemble classifier consisting of multiple heterogeneous one-class classifiers and uses an ordered weighted averaging (OWA) operator to aggregate the outputs of the one-class classifiers. Our work is motivated by the observation that combining multiple one-class classifiers often produces higher overall classification accuracy than any individual one-class classifier. We demonstrate the effectiveness of DroidNMD using a real dataset of Android benign applications and malware samples. The results of our experiments show that DroidNMD can detect Android malware with a high detection rate and a relatively low false alarm rate.