Information Security Applications

DNS leakage happens when queries for names within a private namespace spread out to the public DNS infrastructure (Internet), which has various privacy implications. An example of this leakage includes the documented [1] leakage of .onion names associated with Tor hidden services to the public DNS infrastructure. To mitigate this leakage, and improve Tor’s privacy, Appelbaum and Muffet [2] proposed the special use .onion domain name, and various best practice recommendations of blocking of .onion strings (hidden service addresses) at the stub (browser), recursive, and authoritative resolvers. Without any form of analysis of those recommendations in practice, it is very difficult to tell how much of privacy is provided by following them in various deployment settings. In this paper, we initiate for the study of those recommendations by analyzing them under various settings and conclude that while the unlikely universal deployment will naturally improve privacy by preventing leakage, partial deployment, which is the case for early adoption, will degrade the privacy of individuals not adopting those recommendations.

Internet DDoS attacks are prevalent but hard to defend against, partially due to the volatility of the attacking methods and patterns used by attackers. Understanding the latest of DDoS attacks can provide new insights for effective defense. But most of existing understandings are based on indirect traffic measures (e.g., backscatters) or traffic seen locally (e.g., in an ISP or from a botnet). In this study, we present an in-depth study based on 50,704 different Internet DDoS attacks directly observed in a seven-month period. These attacks were launched by 674 botnets from 23 different botnet families with a total of 9026 victim IPs belonging to 1074 organizations in 186 countries. In this study, we conduct some initial analysis mainly from the perspectives of these attacks’ targets and sources. Our analysis reveals several interesting findings about today’s Internet DDoS attacks. Some highlights include: (1) while 40% of the targets were attacked only once, 20% of the targets were attacked more than 100 times (2) most of the attacks are not massive in terms of number of participating nodes but they often last long, (3) most of these attacks are not widely distributed, but rather being highly regionalized. These findings add to the existing literature on the understanding of today’s Internet DDoS attacks, and offer new insights for designing effective defense schemes at different levels.

Software defined networking (SDN) is a novel programmable networking paradigm that decouples control and data planes. SDN relies heavily on the controller in control plane that tells the data plane how to handle new packets. Because the entire network may be disrupted if the controller is disabled, many attacks including SYN flooding aim to overload the controller by passing through the ingress switches. In this paper, we propose a security enhanced Open vSwitch (SD-OVS) to protect the controller from SYN flooding. The switch authenticates benign hosts by interchanging cookie packets and generates a short-lived security association (SA). The retransmitted SYN packet from these benign hosts is validated using SA and passed on to the controller. Our evaluation shows that SD-OVS protects the controller from SYN flooding at an acceptable time cost.

We evaluate security of WebSocket, one of HTML5 APIs, in the view of L7 DoS attack and design the countermeasure against Slowloris attack which is known as difficult to be detected by IDS and IPS. It is easy to disable services based on WebSocket by sending partial request packets slowly. The server no longer provide the service since Slowloris attack makes request buffer full. For the solution, we design a dual-buffer based countermeasure. The main features of countermeasure are separation of buffer according to status of connections and request acceptance without limitation. In this countermeasure, we propose structure of request buffer free from fullness by employing circular buffer. The connections after handshake process move out to another buffer not to be affected from the request attack. In our construction, when the request buffer is full, the oldest request would be overwritten with a new request. Finally, our proposal allows the benign requests to be successful during Slowloris attack. Our construction could be also applied to other applications including HTTP, FTP and etc.

Detecting encrypted traffic is increasingly important for deep packet inspection (DPI) to improve the performance of intrusion detection systems. We propose a machine learning approach with several randomness tests to achieve high accuracy detection of encrypted traffic while requiring low overhead incurred by the detection procedure. To demonstrate how effective the proposed approach is, the performance of four classification methods (Naïve Bayesian, Support Vector Machine, CART and AdaBoost) are explored. Our recommendation is to use CART which is not only capable of achieving an accuracy of 99.9% but also up to about 2.9 times more efficient than the second best candidate (Naïve Bayesian).

Buffer overflow (BO) attacks are one of the most dangerous threats in the area of network security. Methods for detection of BO attacks basically use two approaches: signature matching against packets’ payload versus analysis of packets’ headers with the behavioral analysis of the connection’s flow. The second approach is intended for detection of BO attacks regardless of packets’ content which can be ciphered. In this paper, we propose a technique based on Network Behavioral Anomaly Detection (NBAD) aimed at connectionless network traffic. A similar approach has already been used in related works, but focused on connection-oriented traffic. All principles of connection-oriented NBAD cannot be applied in connectionless anomaly detection. There is designed a set of features describing the behavior of connectionless BO attacks and the tool implemented for their offline extraction from network traffic dumps. Next, we describe experiments performed in the virtual network environment utilizing SIP and TFTP network services exploitation and further data mining experiments employing supervised machine learning (ML) and Naive Bayes classifier. The exploitation of services is performed using network traffic modifications with intention to simulate real network conditions. The experimental results show the proposed approach is capable of distinguishing BO attacks from regular network traffic with high precision and class recall.

Distributed attacks have reportedly caused the most serious losses in recent years. Here, distributed attacks means those attacks conducted collaboratively by multiple hosts. How to detect distributed attacks has become one of the most important topics in the cyber security community. Many detection methods have been proposed, each of which, however, has its own weak points. For example, detection performance of information theory based methods strongly depends on the information theoretic measures and signature-based methods suffer from the fact that they can deal with neither new kinds of attacks nor new variants of existing attacks. Recently, behavior-based method has been attracting great attentions from many researchers and developers and it is thought as the most promising one. In behavior-based approaches, normal behavior modes are learned/extracted from past traffic data of the monitored network and are used to recognize anomalies in the future detection. In this paper, we explain how to implement an online behavior-based engine for detecting distributed cyber-attacks. Detection cases of our engine are also introduced and some actual attacks/incidents have been captured by our detection engine.

Smart devices interconnected through Internet became one of everyday items. In particular, we are now able to access Internet anywhere and anytime with our smartphones. To support the ad-hoc access to Internet by using smartphones, the computer network structure has become more complex. Also, a certain network node is highly connected to support the diverse Internet services. In this paper, we note that when a node is infected by malicious programs, their propagation speeds from the node with a high level of centrality will be faster than those from the node with a low level of centrality, which identifies the most important nodes within a network. From experiments under diverse worm propagation parameters and the well-known network topologies, we evaluate the influence of Centrality-based random scanning strategy on early worm propagation rate. Therefore, we show that centrality-based random scanning strategy, where an initial infected node selects the victim based on the level of centrality, can make random scanning worms propagate rapidly compared to Anonymity-based random scanning strategy, where an initial infected node selects the victim uniformly.

Cyber attacks cause fatal blow by destructing critical infrastructure of nations with relatively small cost. North Korea uses cyber attacks as a major asymmetric strategy along with nuclear weapon. In this situation, South Korea needs to find better way to cope with North Korea’s aggressive cyber threats. In the present paper, we shall see North Korean cyber capability including its backgrounds, organization, and personnel. By exploring some key signifying structures consisting North Korean cyber power, it is intended to broaden understanding of the nature of North Korean cyber operations. The purpose of this paper is to come up with appropriate and effective ways to respond North Korean cyber attacks.

TLS protocol provides a secure communication environment by guaranteeing the confidentiality and the integrity of transmitted data between two parties. However, there have been lots of vulnerabilities in TLS protocol and attacks exploiting them in aspects of protocol, implementation, and cryptographic tools. In spite of the lessons learned from the past experiences, various attacks on the network systems are being reported continuously due to the lack of care with regard to the proper TLS deployment and management. In this paper, we investigate TLS vulnerabilities in Korea’s top 100 websites selected from Alexa global top 500 sites and 291 Korea’s public enterprise websites. We compare the analysis results with those of Alexa global top 100 websites. Then, we discuss the lessons learned from this study. In order to analyze TLS vulnerabilities efficiently, we developed a TLS vulnerability scanner, called Network Vulnerabilities Scanner (NVS). We also analyze e-mail security of Korea’s top 3 e-mail service providers, which are supposed to be secured by TLS. Interestingly, we found that the e-mail service of them is not so secured by TLS as opposed to the analysis of Google’s transparency report.

Bitcoin is a cryptocurrency based in peer-to-peer network that uses a blockchain. To maintain the blockchain without trusted third parties, a player called a miner proves that he has completed a proof-of-work. As the difficulty of proof-of-work is increasing, mining pools, consisting of a number of miners, have become major players compared with solo miners. Most mining pools consist of a manager and miners. All miners who belong to a mining pool submit their shares to the manager and get paid in proportion to the amount of their shares. Therefore, the manager has to pay all miners fairly.However, many Bitcoin mining pools were ruined by an attack called the Duplicate Share Attack (DSA) in 2015. In this paper, we analyze DSA in multiple directions. First, we mathematically analyze DSA against one mining pool and multiple mining pools. As results of our analyses, we derive the optimal attacker’s strategy, which shows that DSA can give a large extra profit to an attacker with little computational power. Because the duplicate share vulnerability has been already fixed in a few large mining pools after DSA was introduced, DSA may not be considered a threat any more. However, we show that several small mining pools are still vulnerable to DSA and an attacker can unfairly earn a large extra profit using these unpatched small mining pools. In summary, we argue that honest miners in Bitcoin network are not yet free from DSA.

WiFi network traffics will be expected to increase sharply in the coming years, since WiFi network is commonly used for local area connectivity. Unfortunately, there are difficulties in WiFi network research beforehand, since there is no common dataset between researchers on this area. Recently, AWID dataset was published as a comprehensive WiFi network dataset, which derived from real WiFi traces. The previous work on this AWID dataset was unable to classify Impersonation Attack sufficiently. Hence, we focus on optimizing the Impersonation Attack detection. Feature selection can overcome this problem by selecting the most important features for detecting an arbitrary class. We leverage Artificial Neural Network (ANN) for the feature selection and apply Stacked Auto Encoder (SAE), a deep learning algorithm as a classifier for AWID Dataset. Our experiments show that the reduced input features have significantly improved to detect the Impersonation Attack.

Accurate and fair data charging in cellular networks is an important issue because of its large impacts on profits of operators and bills for users. In this study, we analyze the data charging policies and mechanisms for protocols and applications. The analysis shows that all operators in South Korea did not charge the payload of Internet Control Message Protocol (ICMP) echo request/reply messages, as well as the payload attached to Transmission Control Protocol (TCP) SYN and TCP RST packets. In addition, the operators only utilize IP addresses to verify whether the traffic comes from the expected application. By misusing the findings with consideration of Network Address Translator (NAT) in IPv4 cellular networks, we validate with empirical experiments the feasibility of free-riding attack, which enables an adversary to use the cellular data service for free, and propose effective countermeasures.

Manual vulnerability discovery and exploit development on an executable are very challenging tasks for developers. Therefore, the automation of those tasks is becoming interesting in the field of software security. In this paper, we implement an approach of automated exploit generation for firmware of embedded systems by extending an existing dynamic analysis framework called Avatar. Embedded systems occupy a significant portion of the market but lack typical security features found on general purpose computers, making them prone to critical vulnerabilities. We discuss several techniques to automatically discover vulnerabilities and generate exploits for embedded systems, and evaluate our proposed approach by generating exploits for two vulnerable firmware written for a popular ARM Cortex-M3 microcontroller.

As SSL/TLS has become the de facto standard Internet protocol for secure communication in recent years, its security issues have also been intensively studied. Even though several tools have been introduced to help administrators know which SSL/TLS vulnerabilities exist in their network hosts, it is still unclear whether the best security practices are effectively adopted to fix those vulnerabilities in real-world applications. In this paper, we present the landscape of real websites about SSL/TLS weaknesses through an automatic analysis of the possibilities of six representative SSL/TLS attacks—Heartbleed, POODLE, CCS injection, FREAK, Logjam and DROWN—on popular websites. Surprisingly, our experiments show that 45% and 52.6% of top 500 most popular global and Korean websites are still vulnerable to at least one of those attacks, respectively. We also observed several interesting trends in how websites were vulnerable to those attacks. Our findings suggest that better tools and education programs for SSL/TLS security are needed to help administrators keep their systems up-to-date with security patches.

This study integrates the representative information security certification systems such as ISMS, PIMS and PIPL in order to improve efficiency of information security management. It also suggests information security management assessment model for the financial sector by incorporating new control items derived from laws and regulations related to financial IT and information security into the integration model of information security certifications to reflect characteristics of financial industry. The findings have significance in that they solve problems related to duplication of previous information security certification systems and suggest the orientation of information security management system for financial industry enhancing the organizations’ ability to cope with security accidents. Moreover, the suggested methodology can be used in study on systematic and specific information security management standard for each industry.

Most journals and conferences adopt blind review process to ensure fairness through anonymization. Although the identity of an author is blinded in a manuscript, information about the author is known to the system when an account is created for submission. So, Information leak or the abuse from journal editor, who is able to access this information, could discredit the review process. Therefore, the triple-blind review process has been proposed to maximize anonymity through blinding the author, reviewer and also the editor. However, it has not been widely used compared to single- and double-blind review processes because there is difficulty in selecting the reviewers when the author is not known to the editor. In this paper, we propose a novel scheme to select the adequate reviewers in the triple-blind review process without any disclosure of author information to even the editor. This is done by using machine learning classification and a conflict of interest measuring method.

Recently years, vector map is used in many applications and on/off-line services widely. The cost of production of vector map data is very expensive but it is stolen or copied easily by pirates without permission from the original providers. Therefore, provider desires vector map data should be encrypted before storing and transmitting to ensure the access control and prevent the illegal copying of vector map. In this paper, we proposed a perceptual encryption scheme for vector map using geometric objects. The geometric objects of vector map data is extracted to compute features as bounding boxes and distance vectors. After that, we encrypted those features and use them to compute and obtain encrypted vector map data. Experimental results is verified that the entire vector map is changed after encryption process. The proposed method is very effective for a large of dataset, responsive to requirements of security.

Efficiency of the next generation pairing based security protocols rely not only on the faster pairing calculation but also on efficient scalar multiplication on higher degree rational points. In this paper we proposed a scalar multiplication technique in the context of Ate based pairing with Kachisa-Schaefer-Scott (KSS) pairing friendly curves with embedding degree $$k = 18$$k=18 at the 192-bit security level. From the systematically obtained characteristics p, order r and Frobenious trace t of KSS curve, which is given by certain integer z also known as mother parameter, we exploit the relation $$\#E({\mathbb {F}}_{p}) = p+1-t$$#E(Fp)=p+1-t mod r by applying Frobenius mapping with rational point to enhance the scalar multiplication. In addition we proposed z-adic representation of scalar s. In combination of Frobenious mapping with multi-scalar multiplication technique we efficiently calculate scalar multiplication by s. Our proposed method can achieve 3 times or more than 3 times faster scalar multiplication compared to binary scalar multiplication, sliding-window and non-adjacent form method.

Due to the advancement of side-channel attacks, leakage-resilient cryptography has attracted a lot of attention in recent years. Many fruitful results have been proposed by researchers. Most, if not all, of these results are theoretical in nature. Not much has been done to realize these schemes for practical use. In this work, we design and provide a leakage-resilient cryptographic system $$\mathcal {LRCRYPT}$$LRCRYPT with programming interfaces for users to build leakage-resilient cryptographic applications. $$\mathcal {LRCRYPT}$$LRCRYPT consists of a few fundamental building blocks that perform leakage-resilient public-key encryption, leakage-resilient signature, and leakage-resilient secret-key encryption, which can also be extended to many existing leakage resilience cryptographic primitives. We have conducted both a security analysis and a performance evaluation on $$\mathcal {LRCRYPT}$$LRCRYPT. To our knowledge, $$\mathcal {LRCRYPT}$$LRCRYPT is the first to work in this domain.

A group signature allows any group member to anonymously sign a message. One of the important issues is an efficient membership revocation. The scheme proposed by Libert et al. has achieved O(1) signature and membership certificate size, O(1) signing and verification times, and $$O(\log N)$$O(logN) public key size, where N is the total number of members. However the Revocation List (RL) data is large, due to O(R) signatures in RL, where R is the number of revoked members. The scheme proposed by Nakanishi et al. achieved a compact RL of O(R/T) signatures for any integer T. However, this scheme increases membership certificate size by O(T). In this paper, we extend the scheme proposed by Libert et al., by reducing the RL size to O(R/T) using a vector commitment to compress the revocation entries, while O(1) membership certificate size remains.

This paper is a position paper based on an invited talk at WISA’16 in Korea. We argue that Quantum-Safe Cryptography (QSC) will likely have a deep impact on the practice of IT professionals. We detail also in the second part a classical candidate for quantum-safe cryptography: multivariate cryptography. Finally, we conclude by presenting HFEBoost a real-life deployment of multivariate cryptography.

Integral cryptanalysis is one of the most powerful attacks on symmetric key ciphers. Todo proposed a novel technique named the division property to find efficient integral characteristics. In this paper, we apply this technique to the symmetric key block cipher KASUMI which was developed by modifying MISTY1. It has been used worldwide in the 3-rd generation mobile communication networks. As a result, we found new 4 and 5-round integral characteristics of KASUMI with FL and 6-round characteristics of KASUMI without FL for the first time. We show that 6-round KASUMI with FL is attackable with $$2^{57}$$257 data complexity and $$2^{58}$$258 encryptions. The attack of 6-round KASUMI by integral cryptanalysis is the best in terms of time complexity.

Some authors suggest to estimate the number of unbiased bits extractable from a stateless physical source by Shannon entropy, which can be justified asymptotically by the Asymptotic Equipartition Property. We show that this estimate, refereed to as the AEP heuristic, involves a heavy error term and makes the extracting process insecure.Suppose one wants to obtain k almost uniform bits from i.i.d samples $$X_1,\ldots ,X_n$$ X1,…,Xn. While the AEP heuristic gives $$k \approx \mathbf {H}(X)$$ k≈H(X) where H is the Shannon Entropy, we show that pseudoentropy of this sequence equals $$k = {H}(X) -\varTheta \left( \sqrt{n\log (1/\epsilon )}\right) $$ k=H(X)-Θnlog(1/ϵ)where $$\epsilon $$ ϵ is a user-defined security parameter that bounds distinguishing probability (typically $$\epsilon = 2^{-80}$$ ϵ=2-80).Implications of our result are as follows.(a)AEP heuristic is provably insecure in the information-theoretic sense(b)AEP heuristic is not provably secure in the computational setting(c)AEP heuristic is secure if the error term is addressed.Our proof uses tools from large deviation theory and hypothesis testing.

Integral attack is a powerful method to recover some round keys of block ciphers by exploiting the characteristic that a set of outputs after several rounds encryption has (integral distinguisher). Recently, Todo proposed a new algorithm to construct integral distinguisher with division property. However, the existence of integral distinguisher which holds in additional rounds can not be denied by the algorithm. On the contrary, our approach is to obtain the number of rounds which integral distinguisher does not hold. The approach is based on algebraic degree estimation. We execute a random search for a term which has a degree equals the number of all inputted variables. We propose two algorithms and apply them to PRESENT and RECTANGLE. Then, we confirm that there exists no 8-round integral distinguisher in PRESENT and no 9-round integral distinguisher in RECTANGLE. From these facts, it is infeasible to attack more than 11-round and 13-round of PRESENT and RECTANGLE, respectively.

Soft computing offers a number of interesting options how to solve many real world problems where security and cryptology domains are not exceptions. There, machine learning and various optimization techniques can play a significant role in finding new, improved solutions. Sometimes those methods are used to solve the problem itself, while sometimes they just represent a helper tool in a larger task. A more in-depth understanding of such techniques is always beneficial. Moreover, the research topics belonging to the intersection of the soft computing and the cryptology are rather demanding since usually neither of those two communities devotes much attention to the other area. In this paper, we briefly discuss three well-known applications of soft computing to the cryptology area where we identify main challenges and offer some possible future research directions.

In this paper we revisited the parallel implementations of LEA. By taking the advantages of both the light-weight features of LEA and the parallel computation abilities of ARM-NEON platforms, performance is significantly improved. We firstly optimized the implementations on ARM and NEON architectures. For ARM processor, barrel shifter instruction is used to hide the latencies for rotation operations. For NEON engine, the minimum number of NEON registers are assigned to the round key variables by performing the on-time round key loading from ARM registers. This approach reduces the required NEON registers for round key variables by three registers and the registers and temporal registers are used to retain four more plaintext for encryption operation. Furthermore, we finely transform the data into SIMD format by using transpose and swap instructions. The compact ARM and NEON implementations are combined together and computed in mixed processing way. This approach hides the latency of ARM computations into NEON overheads. Finally, multiple cores are fully exploited to perform the maximum throughputs on the target devices. The proposed implementations achieved the fastest LEA encryption within 3.2 cycle/byte for Cortex-A9 processors.

Multi-precision squaring is one of the most performance-critical operations for implementations of public-key cryptography, e.g. RSA, ECC as well as Diffie-Hellman key exchange protocols. In this paper, we propose novel techniques to push the speed limits of multi-precision squaring on embedded processors. The method reduces the number of memory access operations and improves the previous Sliding Block Doubling method by 4.1% on 8-bit RISC processor.

Biometric authentication is the use of unique human features to provide secure, reliable, friendly and convenient access to an environment or a computer installation. However, the use of biometrics as a means of authentication exposes legitimate users to security threats, privacy attacks and loss of identity. This paper proposes and implements a novel non-invertible transformation technique known as matrix transformation. Matrix transformation is a simple but powerful and effective method to achieve template revocability and prevent the recovery of original biometric data from secured templates. The approach provides a high level template security and user privacy. It is also robust against replay attack, cross matching and loss of identity.

Most of current text-based CAPTCHAs have been shown to be easily breakable. In this work, we present two novel 3D CAPTCHA designs, which are more secure than current 2D text CAPTCHAs, against automated attacks. Our approach is to display CAPTCHA characters onto 3D objects to improve security. We exploit difficulty for machines in rotating 3D objects to find a correct view point and in further recognizing characters in 3D, both tasks that humans can easily perform. Using an offline automated computer vision attack, we found that 82% of the new text reCAPTCHA characters were successfully detected, while approximately 60% of our 3D CAPTCHAs were detected only if characters were focused and zoomed from the direct view point. When CAPTCHAs are presented in slightly different views, the attack success rates against our approaches are reduced to almost 0%.

User behavior-based authentication, while providing convenience to the user, is not widely used in the real world due to its low accuracy. Keystroke dynamics is one of the user behavior-based authentication methods, and it has been studied for about 40 years. Conventional keystroke dynamics has used key timing features for the personal computer (PC) environment. Since the smartphone equipped with advanced sensors (e.g., accelerometer, gyroscope, and touchscreen sensor) was released, sensor-based features have been used to improve the accuracy of classifying users with key timing features. In this paper, we analyze the keystroke dynamics features in the literature and evaluate each feature to find efficient features. Based on tapping data collected from 12 participants, we evaluate the effectiveness of several features from the empirical data of a six-digit PIN. Our experimental results show that the feature Up-Up (UU), the time difference between releasing a key and the next key, and the min, max, and mean features extracted from motion sensor data have the best accuracy and efficiently classify each user.

A number of recent maritime accidents strongly imply the need of distributed smart surveillance. The maritime cloud, proposed as communications infrastructure of e-Navigation, is one of the most optimal infrastructure systems in the smart surveillance environment. To maintain the safe maritime environment, security in the distributed smart surveillance environment is critical, but research on security of the maritime cloud, which will be adopted as major communications infrastructure in the smart surveillance system, is still in the fledging stage. In this regard, this paper suggested a safe synchronization method of Almanac, which is necessary to provide unimpeded maritime cloud service. Almanac plays a role of a telephone directory and it should be shared in the latest version in communicating between vessels or a vessel and land. In other words, synchronization of Almanac between offshore and vessels is required to safely deliver major video information collected by the distributed smart camera. The method proposed in this paper enables geocasting based synchronization between vessels, which is suitable for maritime conditions, and does not expose information in the course of synchronization even in the case of broadcasting through an unsafe channel. In addition, the method ensures integrity based on block ID and supports delta update, thereby minimizing bandwidth and boosting performance.

In this paper, we provide a mechanism to be able to save time and human resources to predict the time of occurrence of accidents at sea this time of the incident are not clear and to extract the suspected vessel-related accidents at sea. The proposed mechanism, it save such amount of data and time to trajectory extraction by managing separate the control area in a grid pattern, and is characterized as possible through trajectory analysis for a particular area. It also can be reduced 30 times faster than existing marine accident analysis time using the playback function in VTS operating system, so it is effective in saving human resource and time.