Xin Lv

A Load Balancing Scheme for Distributed Key-Value Caching System in Cloud Environment

Distributed key-value caching system has been deployed in many kinds of clouds. The effect of load balancing between each node is crucial in key-value caching system in clouds. Invalidation is a feature of the data in the key-value caching system, and making some cache invalid is efficient instead of adjusting the load location when load balancing. It is worth studying that how to utilize the feature reasonable to reach maximum load balancing. Aiming at this point, a new cache-invalidation-scope model in key-value caching system is proposed. Combined with greedy algorithm, the scheme provides a better load balancing algorithm for different load cases (CLB). CLB algorithm utilizes entropy and the scope of invalid cache invalid as the evaluation basis of load balancing effect. Compared with existing algorithm, CLB load balancing algorithm lifts up the performance of key-value caching system more than one times.

A Novel Identity Authentication Scheme of Wireless Mesh Network Based on Improved Kerberos Protocol

The traditional Kerberos protocol exists some limitations in achieving clock synchronization and storing key, meanwhile, it is vulnerable from password guessing attack and attacks caused by malicious software. In this paper, a new authentication scheme is proposed for wireless mesh network. By utilizing public key encryption techniques, the security of the proposed scheme is enhanced. Besides, timestamp in the traditional protocol is replaced by random numbers to implementation cost. The analysis shows that the improved authentication protocol is fit for wireless Mesh network, which can make identity authentication more secure and efficient.

A Clustering-Based QoS Prediction Approach for Web Service Selection

With the number increasing of Web services, recommending and selecting the optimal Web services for consumers has become one of the most important challenges in the field of service computing. The goal of consumers is to discover and use services which lead to them experiencing the highest quality. The existing approaches of quality evaluation mostly assume the primary goal of consumer as being optimization of performance, so that the consumers are unable to effectively identify and engage with providers who deliver services that will best meet their needs. In order to solve this problem, we propose a clustering-based QoS prediction framework for Web services. In our framework, we employ the k-means clustering algorithm. We determinate the similarity among consumers based on their expectations. In terms of the relationship between their expectations and the rating of the services, we predict the rating of a service used to select desirable services. At last, the experiment shows that the approach achieves better prediction.

A New Verifiable Threshold Decryption Scheme Without Trusted Center

In the most of scenarios, the trusted center dose not existed. A new verifiable threshold decryption scheme without trusted center was proposed in this paper. In order to publicly verify the identity of decryption member, the scheme forces the member to submit a commitment at the beginning of the process. Moreover, the scheme has several advantages such as cheat-proof and dynamic member revoking.

Image scrambling based on life-like cellular automata

This paper presents an image scrambling method based on the life-like cellular automata (CA). In the scrambling process, a life-like CA with an initial random configuration is set to run for several generations to obtain scrambling matrices. The restoration process is the reverse process of scrambling. In order to achieve a good diffusion property, we analyze how the scrambling effect is influenced by different CA initial configurations, and give the results of which life-like rules generate the best scrambling effect.

Schnorr Ring Signature Scheme with Designated Revocability

Abstract Ring signatures enable a user to sign a message so that a ring of possible signers is identified, without revealing exactly which member of that ring actually generated the signature. In some situations, however, an actual signer may possibly want to expose himself, for instance, if doing so, he will acquire an enormous benefit. In this paper, we present a Schnorr ring signature scheme with designated revocability, providing a confirmation procedure, in which the real signer is able to convince a designated party that he is the one who generates the signature. The confirming proof is non-transferable, which only can be triggered by the signer. Meanwhile, the scheme satisfies unconditional anonymity, and has been proven to be existentially unforgeable under adaptive-chosen message attack in the random oracle.

A new identity-based threshold ring signature scheme

Aiming at the security problems existed in most identity-based threshold ring signature (ITRS) schemes, including poor security and inefficient in verifying, a new ITRS scheme is proposed. We significantly enhance the security of signature key by using bilinear pairings and secret sharing techniques. Our analysis indicates that the proposed scheme achieves unconditional anonymity and unforgeability. A security problem (ring member-changing attack) also can be solved. Compared to the congener schemes, ours is much more efficient in signing process.

Public Verifiable Threshold Decryption Scheme without Trusted Center

The paper proposes a publicly verifiable threshold decryption scheme with no center based on the situation that a trust center does not exist in many special occasions. The scheme asks decryption member for submitting commitment in the decryption process in order to make sure the identity is publicly verifiable. Moreover, the scheme has several advantages such as protecting the shadow without revealing and the members can be deleted dynamically.

Uncertain Frequent Itemsets Mining Algorithm on Data Streams with Constraints

Nowadays, many emerging applications in real-life can produce amount of uncertain data streams, while people are often interested in some aspects. To mine constrained frequent itemsets on uncertain data streams, this paper presents a method. First, determining the order of items in the transactions of data streams according to the properties of constraints; then, inserting items into the tree in order; finally, mining constrained frequent itemsets from the tree. Existing algorithms are compared with the proposed method and the performances are analyzed. Results indicate that the proposed method is effective and efficient, which mines constrained frequent itemsets when users request for the mining results and need no additional memory.

A Similarity Model Based on Trend for Time Series

This paper presents a time series similarity matching model based on trend meeting the peoples intuitive sense of trends characterize similarity. At the same time, the concept of similarity value is introduced in order to display the similarity of time series in a more intuitive form. In this model, the original time series are segmented according to the time series segmentation algorithm based on significant points. Each sub-section of the time series are mapped to a two-dimensional vector according to the slope and time span, and then symbolic the two-dimensional vector and calculate the distance between two time series of strings. Finally according to similarity calculation formula proposed, obtain the similarity value between the two time series. Experimental results show that the time series similarity matching model is good. In the aspect of similarity matching, the applicability, high efficiency.