The business of setting up automated teller machines (ATMs) for banks depends on many factors such as the price of buying or leasing an ATM, cost of deployment, cost of operation, and ATM characteristics to be deployed. ATM deployment is an intensive computational problem since it is analogous to file server placement, which is known to be as NP-complete problem. Also, ATM maintenance is an intensive service for the bank to sustain its competitive advantage and customer's satisfaction. We have formulated the ATM allocation problem as an optimization problem, where the objective function is to minimize deployment and operational costs subject to the customers' satisfaction and bank's requirements. Therefore, we have proposed a custom-made genetic algorithm to search for the best possible placement of ATMs with the lease cost. Our proposed method finds solutions under 15 minutes on a desktop computer.
In this paper an optimized k-means implementation on the graphics processing unit (GPU) is presented. NVIDIA's Compute Unified Device Architecture (CUDA), available from the G80 GPU family onwards, is used as the programming environment. Emphasis is placed on optimizations directly targeted at this architecture to best exploit the computational capabilities available. Additionally drawbacks and limitations of previous related work, e.g. maximum instance, dimension and centroid count are addressed. The algorithm is realized in a hybrid manner, parallelizing distance calculations on the GPU while sequentially updating cluster centroids on the CPU based on the results from the GPU calculations. An empirical performance study on synthetic data is given, demonstrating a maximum 14x speed increase to a fully SIMD optimized CPU implementation.
Digital Preservation deals with the long-term storage, access, and maintenance of digital data objects. In order to prevent a loss of information, digital libraries and archives are increasingly faced with the need to electronically preserve vast amounts of data while having limited computational resources in-house. However, due to the potentially immense data sets and computationally intensive tasks involved, preservation systems have become a recognized challenge for e-science. We argue that grid and cloud technology can provide the crucial technology for building scalable preservation systems. In this paper, we present recent developments on a Job Submission Service that is based on standard grid mechanisms and capable of providing a large cluster of virtual machines. The service allows clients to specify and execute preservation tools on large data sets based on dynamically generated job descriptors. This approach allows us to utilize a cloud infrastructure that is based on platform virtualization as a scaling environment for the execution of preservation workflows. Finally, we present experimental results that have been conducted on the Amazon EC2 and S3 utility cloud infrastructure.
Internet-based media streaming services are replacing more traditional broadcast services, resulting in increased demand for network resources. Multicast techniques may be used to improve bandwidth utilisation by servicing one or more clients with each network stream. In this paper, we re-evaluate the performance of the existing multicast patching technique. In particular, we build on the original server-centric performance evaluation to examine the effect of patching on an entire network. We also propose a novel approach to the implementation of patching using standardised transport control (RTSP) and multicast control (MLD) protocols. Our results highlight how network distance from the origin server effects the improvements gained by using patching.
