Search

The new generation of data analytics frameworks require sub-second response times. This in turn requires low latency and high throughput schedulers to allow distributed real- time data analysis. Recent studies in resource allocation for datacenters have focused on probe-based distributed schedulers to improve throughput and jobs response time. However, reducing the search space by relying on probing causes poor response times, especially when the cluster load increases. As a result, high performance cluster schedulers operate at the expense of over provisioning with an increasing operational cost. This paper proposes an Adaptive Probe Size Estimation (APSE) algorithm for cluster management in datacenters. The algorithm efficiently estimates a minimum probe size that limits overprovisioning while keeping the region of high throughput and low response time intact. This approach is particularly effective in large clusters and under heavy loads. The algorithm has been implemented and tested in Sparrow with synthetic and trace-driven data. It has shown considerable improvement in terms of response time, load balancing, and provisioning cost. The algorithm can be used in probe-based schedulers such as Sparrow, Tarcil, Piper, and Hawk.

In probe-based distributed schedulers, little information is known about the state of the cluster. As a result, there is uncertainty about the underlying resource demand and usage. To efficiently leverage cloud datacenters' resources while maintaining the expected performance, one must address the question of how to achieve a good and accurate estimation of the cluster utilization in a stateless manner. We propose a scalable and efficient algorithm to estimate cluster load with a predetermined margin of error and confidence level. This algorithm can be used by cloud service providers to improve resource management systems and to estimate resource utilization. Due to its simplicity, the algorithm can be used in probe-based schedulers such as Sparrow, Tarcil, Piper, and Hawk.

Datacenters are core computational engines in cloud computing. They are de facto platforms to run computationally intensive applications and execute big data workload. They are expected to provide reliable and cost-effective cloud services. The exponential growth in cloud applications, the demand for more computational and communicative resources, and the need for energy and cost-efficient resource management have motivated researchers to introduce an array of resources management techniques to efficiently manage datacenters' resources. To address these challenges and facilitate future advancements in datacenters, it is important to synthesize, classify and assess the proposed resources management techniques for datacenter clusters. This article investigates how different architectural designs could affect the operational effectiveness of a datacenter. It presents key contributions and a taxonomy of resource management techniques that would guide future design and developmental efforts. It will help researchers and developers to design the most appropriate resource management framework for their datacenters and optimize certain design goals under different workloads.

Social network analysis has an increasing growth as an academic field which overlaps with popular interest in social networks. Search for an expert is one of the most important issues of mining of social networks which is finding the right person with the suitable skills and knowledge. The RLS algorithm exploited Q-Learning and referrals to find experts in social network to search expert in social network. Comparison of RLS with Simple Search Algorithm, Referral Algorithm and SNPageRank shows increase in both precision and recall. RLS learns to find new experts as old experts substitute their role with new ones due to changes in social network environment.

To support increased device functionality and higher data-rates in LTE-enabled systems, while improving user experience and usage time, there is a need to reduce RFIC size and power consumption without degrading performance, while maintaining backward compatibility with legacy 2G/3G systems [1]. This paper introduces a 13mm2, 40nm CMOS 2G/HSPA+/TDSCDMA/UE cat. 4 transceiver that consumes 36/65mA battery-referenced current in 3G/LTE20 modes (B1, −50dBm TX, −60dBm RX). This is achieved in part by employing a multiport single-core LNA with a multitap inductor and a current-mode-driven single-core transmit mixer. Baseband-assisted calibration techniques help achieve 60dBm IIP2 in all bands. To save on platform area and cost, the RFIC supports single-ended LNAs, 32kHz clock generation, and free-running XTAL operation. TX SAW filters are not required.

Last mile residential connectivity and the demand for increased indoor coverage and capacity in 3G cellular systems has led to the commercial introduction of femtocell (Home Node B) base stations. To achieve higher component integration, support additional functionality, and decrease power, there is a need to minimize component count, PCB area and power consumption without sacrificing performance, while also preserving backwards-compatibility with legacy 2G systems. In this paper, we present a 65nm CMOS 3G/HSPA+ femtocell transceiver with 350mW power consumption that eliminates the need for Tx SAW filters. The proposed solution supports 10 TX (Downlink) and 10+10 (Downlink+Uplink) RX UMTS bands from Band I to Band VI and Band VIII to Band XI, quad-band GSM sniffing and GPS band receive capability for soft GPS operation.

In multi-hop wireless mesh networks, the end-to-end throughput decays exponentially with the increase in hop counts. This is mainly due to the increase in successive spatial contentions. As a result, flows with longer hops suffer from bandwidth starvation and fair link access disparities along the path towards the gateway. In this paper, we first develop a theoretical reference model that can be used to measure throughput disparities among contending flows. Second, we introduce a fair access rate (FAR) algorithm, based on the end-to-end blocking probabilities of contending flows, that will offset throughput disparities. Finally, we simulate FAR over a set of 802.11s-based network topologies with different aggregate blocking probabilities. The performance of FAR, in terms of its throughput and fairness index, has been evaluated and compared with the performance of the default algorithm used in 802.11s.

In this article, individual's biometric index, radio frequency identification (RFID), and image processing are brought together in order to offer a new model of identity authentication. The suggested system has been designed in two phases: producing identity authentication card and identity confirmation. The individual's biometric image are put on a contact-less card equipped with an RFID tag and for each identity authentication, the data on the card are compared with the online biometric image. If the individual's identity is authenticated, then there will be no need for him/her to punch in the personal information since his/her information will be retrieved from the database through their electronic personal code (EPLC).

A call center is an organizational unit where inbound calls are received or outbound calls placed for the purposes of sales, support, advice and other business transactions. By applying knowledge management solutions, we can meet call centers’ challenges and gain benefits of reduced training costs, improved call handling and greater flexibility. This paper describes a solution to agent mediated knowledge management system in call centers and new search algorithms. We introduced two agent based search algorithms for finding the expert user in system. Searching an expert is one of the most critical features which each call center system should have. Our agent based search with agent based weighted graph works well in our prototype system. Our agent model can distribute and share knowledge in whole levels of organizations and call center.

Several factors greatly influence the quality of service (QoS) in multihop wireless mesh networks (WMNs). These include the transmission range, number of gateways, number of nodes served by each gateway, gateway location, relay load and access fairness. While finding an optimal solution to simultaneously satisfy the above constraints is known to be MV-hard, near optimal solutions can be found within the feasibility region in polynomial time using heuristic algorithms. In this paper, first we present a near optimal heuristics algorithm for gateway placement, and later we compare its performance with some previously known sub-optimal solutions.

A direct conversion, 65 nm CMOS multi-standard TV tuner is presented, which is compliant with several digital terrestrial, fixed and mobile TV standards (DVB-T, DVB-H, T-DMB and ISDB-T, including 1-3 segment). The tuner achieves a 3/3/2.8 dB NF at VHF, UHF and L-band respectively, while consuming less than 140 mW in DVB-T mode. By using integrated transformers, both single-ended and differential RF inputs are supported. Multi-standard capability is enabled by programmable channel-select filters and a Sigma-Delta fractional-N synthesizer.

Enterprise application integration (EAI) is a very complicated process of integrating multiple applications, which were independently developed, use incompatible technology, and remain independently managed. In this paper we are going to discuss about the evolution of EAI concepts and a new elastic framework for enterprise application integration (EAI). The new framework provides a service based architecture for seamless integration of enterprise applications which can develop from the bottom up, building on existing technology. The approach is demonstrated by building a real world application for EAI in the education service domain. Business drivers for and approaches to EAI are present first. We then present our implementation of the architecture and practical challenges encountered in EAI.

The main function of most current forecaster's workbench is data access. They are usually an integrative graphics system to show the data and derived data, and give some analysis tools, and graphics manipulation tools, we have three problems with such systems. 1. Although the system can provide many data (over 2 GB, several thousand fields one day), forecasters only use little of them (less than 1 %) in operational forecast. 2. Massive storage system is developing in Iran ,we will get many history data in operational forecast, how to make our system to use these data? 3. Many kinds of special observation system are developing, such as AWS, profiler, GPS/Met, storm lightning locator, etc. They are very useful for public service, but we must have a flexible data manager to use them in operational forecast. The most serious problem is the first one above. Data warehouse is a new generation of decision support system (DSS). It may provide a good solution for the problems.

Seamlessly roaming user applications in converging GSM-WLAN environments are creating the need for highly integrated, dual-mode transceivers, amenable to single-die integration on silicon. Based on the direct conversion principle, an efficient system architecture that can be switched between both standards is proposed. It is shown that, for implementing the dual-mode radio, the design of wideband RF cells and switches is called for, which is nontrivial. Appropriate design techniques based on the extensive usage of on-chip passives are proposed to address the design of the dual-mode RF circuitry and render their performance on silicon as predictable as possible. The development of a SiGe BiCMOS GPRS-WLAN transceiver is showcased. Measurement results are provided at the cell level, for a 1.8-2.5 GHz 24 dB-gain LNA and an embedded +24 dBm WLAN PA.

This paper describes the analog realization of fuzzifier and defuzzifier interfaces that take advantage of the integration capability afforded by LSI and VLSI technology. The proposed structure uses analog ramp generator (RG) blocks in the fuzzifier interface. Different membership functions are produced by changing the reference and control voltages. Triangular and trapezoidal membership functions can be generated. The defuzzifier interface uses a modified center of area (COA) method that allows the analog implementation of defuzzifier hardware.