Catálogo de publicaciones - libros

Predictive -persistent CSMA protocol is an important MAC solution developed for networked sensor/control applications and used in Local Operating Networks (LonWorks) technology. The protocol uses a built-in network load prediction to support collision avoidance. The paper presents an analytical study of prediction efficiency for a channel with collision detection. The approach based on Markov chains is applied. The procedure of performance analysis includes the definition of transition probabilities of Markov chain for a specified load scenario, calculation of stationary distribution of contention window, and the probabilities of successful/unsuccessful transmission. It is shown that the predictive -CSMA protocol manages to control the size of a competition window in order to guarantee the sustained probability of a successful transmission. The simulative validation of analytical results is provided.

The mobile internet environment which is in the limelight as the important platform of the ubiquitous environment gets accomplished by the intimate relation with user. In order to realize the interaction between device and user, it is considered that resource of exterior/interior user information which can be collected by mobile device and the situation-aware (SA) personalization is suggested by applying the context set of collected current situation to the concept of situation-aware. Such a SA personalization is designed to offer advanced personalization using Link Retrieving Algorithm which is emphasized on prospecting. And the Markov Chain Model, prospecting matrix system, is used to support the SA personalization. Using SA personalization system, the custom service which is well-matched on the ubiquitous era and founded on user’s current situation will be offered.

Currently, there are a lot of e-learning and collaborative platforms to support distance and collaborative learning, however, all of them were designed just like an application without considering the network infrastructure below. Under these circumstances when the platform is installed and runs in a campus, sometimes it has very poor performance. This paper presents a network and data link layer infrastructure design that classifies and prioritizes the voice and video traffic in order to improve the performance and QoS of the collaborative systems applications. This infrastructure has been designed taking in consideration a typical network of a university campus, so that in this way it can be implemented in any campus. After making the design we have made some tests in a laboratory network demonstrating that our design improves 70-130% the performance of these real time collaborative systems which transmit voice and video.

Recently, the demands on information services have been increasing significantly. This is mainly due to the popularization of computer and mobile telecommunication devices and the rapid improvements on wireless communication technology. Specially, information services and their corresponding management for mobile devices, such as Location Based Service (LBS) and Telematics, become more and more important. However, the standard for geographical space data has not been finalized. Many commercial monitoring systems are using their own independent geographical information without making them compatible to others. Much efforts and resources have been wasted on managing and operating those different monitoring systems’ geographical Information System (GIS) databases. Accordingly, a standard format called GML, based on the most commonly used geographical data format such as DXF, DWG and SHP, has been emerged. In this paper, our work on GML’s visualization in Trace Monitoring Systems (TMS) is described fully. The details on how to trace and manage data moving among different mobile terminals are presented as well. 

Denial-Of-Service (DOS) attack is recognized as a biggest threat against the operation of large-scale wireless sensor networks (WSN). Especially, high-mobility radio jamming like vehicles carrying radio jamming device can cause a serious damage in performance of WSNs. Because of resource-constraint design of sensor node, it is hard to provide enough protection against high-mobility jamming attack. Therefore, large-scale WSNs are extremely vulnerable to that type of DOS attack. Recognizing the importance of the problem, we conducted a simulation study to investigate the impact of radio jamming on the performance of a large-scale WSN. Based on the simulation results, the moving speed of radio jamming source has the most conspicuous effects on the WSN performance such as packet delivery success ratio and delay. As the speed changes from 8 m/sec to 1 m/sec, the success ratio drops by up to 10%. On the other hand, the delay increases by up to 55%.

RFID (Radio Frequency Identification) is recently becoming popular, promising and widespread. In contrast, RFID tags can bring about traceability that causes user privacy and reduces scalability of RFID. Guaranteeing untraceability and scalability at the same time is so critical in order to deploy RFID widely since user privacy should be guaranteed. A large number of RFID protocols were designed in the open literature, but any known protocols do not satisfy untraceability and scalability at the same time to the best of our knowledge. In this paper, we suggest a RFID authentication protocol that guarantees untraceability and scalability together; needless to say preventing several known attacks: replay, spoofing, desyncronization, and cloning by eavesdropping. Our protocol supports ownership transfer and considers multi-tag-reader environment; a reader receives messages from the tags what a reader wants in our protocol. In addition, we address the reason why the item privacy is important, and a way to keep it securely.

In this paper, we analyze the security of the RFID authentication protocol proposed by Choi at SecUbiq 2005. They claimed that their protocol is secure against all possible threats considered in RFID systems. However, we show that the protocol is vulnerable to an impersonation attack. Moreover, an attacker is able to trace a tag by querying it twice, given the initial information from + 1( ≈ ℓ + 2) consecutive sessions and 2 · ( ≈ 2(ℓ + 1)) consecutive queries, where ℓ is the length of secret values (in binary).

Due to the low-cost nature of sensor network nodes, we cannot generally assume the availability of a high-performance CPU and tamper-resistant hardware. Firstly, we propose a reliable broadcast message authentication working under the above-mentioned circumstances. The proposed scheme, although based on symmetric cryptographic primitives, is secure against anyone who knew the message authentication key as well as the malicious router nodes in multi-hop networks. The proposed scheme consists of three steps; (i) reliable broadcast of a message, (ii) legitimate acknowledgments from all the nodes in the network, and (iii) disclosure of the message authentication key. Secondly, we propose a way to reduce the amount of the stored information until the disclosure of the key, in which the server transmits the message integrity code of a message before transmitting the message. Finally, we consider the characteristic and the security issues of the proposed schemes.

In distributed sensor networks, the researches for authentication in sensor network have been focused on broadcast authentication. In this paper, we propose a message and its origin authentication protocol for data aggregation in sensor networks, based on one way hash chain and Merkle tree authentication with pre-deployment knowledge. Proposed protocol provides not only for downstream messages but also for upstream messages among neighbors, and it solves the secret value update issue with multiple Merkle trees and unbalanced energy consumption among sensor nodes with graceful handover of aggregator. In treating compromised node problem, our protocol provides an equivalent security level of pair-wise key sharing scheme, while much less memory requirements compared to pair-wise key sharing scheme.

In this paper, we describe the design and implementation of a new security protocol based on Elliptic Curve Cryptosystems (ECC) for securing Wireless Sensor Networks (WSNs). Some public-key-based protocols such as TinyPK and EccM 2.0 have already been proposed in response. However, they exhibit poor performance. Moreover, they are vulnerable to man-in-the-middle attacks. We propose a cluster-based Elliptic Curve Diffie-Hellman (ECDH) and Elliptic Curve Digital Signature Algorithm (ECDSA) for efficiency and security during the pairwise key setup and broadcast authentication phases, respectively. We have implemented our protocol on 8-bit, 7.3828-MHz MICAz mote. The experimental results indicate the feasibility of our protocol for WSNs.