Qatar Foundation Annual Research Forum Volume 2012 Issue 1

The rapid growth of large data processing has meant the implementation of the MapReduce programming model as a widely accepted solution. The simple map and reduce stages have introduced convenience to programmers in order that they may quickly compose and design complex solutions for large-scale problems. Due to the ever-increasing complexity of execution logic in real-life applications, more and more MapReduce applications involve multiple correlated jobs encapsulated and executed according to a defined order. For example, a PageRank job involves two iterative MapReduce sub-jobs; the first job joins the rank and linkage table and the second one calculates the aggregated rank of each URL. Two non-iterative MapReduce sub-jobs for counting out-going URLs and assigning initial ranks are also included in the PageRank job. Besides this, MapReduce programming model lacks built-in support and optimization when the input data are shared. The performance will benefit when the shared and frequently accessed files are read from local instead of from distributed file system. This paper presents Concurrent MapReduce; a new programming model built on top of MapReduce while maintaining optimization and scheduling for big data applications that are composed of large number of shared data items. Concurrent MapReduce has three major characteristics: (1) Unlike traditional homogeneous map and reduce functions, it provides a flexible framework, which supports and manages multiple yet heterogeneous map and reduce functions. In other words, programmers are able to write many different map and reduce functions in a MapReduce job. (2) It launches multiple jobs in a task-level concurrency and a job-level concurrency manner. For job-level concurrency, the framework manages the shared data by replicating them from HDFS to the local file system to ensure data locality. For task-level concurrency, it is the programmers' responsibility to define the data to be shared. (3) We have evaluated the framework using two benchmarks: Single Source Shortest Path and String Matching. Results have demonstrated up to 4X performance speedup compared to traditional non-concurrent MapReduce.

Pipe inspection is a critical process in many industries, including oil and gas. Conventional practice relies on a range of Non-Destructive Testing (NDT) approaches such as ultrasonic and magnetic flux leakage methods. While these approaches can provide high accuracy wall thickness measurements, which can be used to monitor corrosion, they provide poor visualizations, and are typically unable to provide full pipe coverage. Moreover, they cannot be used to localize where in the pipe a defect is without expensive and possibly restricted sensors such as Inertial Navigation Systems. We have developed an automated vision-based approach that builds high-resolution 3D appearance maps of pipes and provides vehicle localization. These maps include both structure and appearance information, and can be used for direct metric measurement of pipe wall thickness, or as input to automatic corrosion detection algorithms. They may also be imported into 3D rendering engines to provide effective visualization of a pipe network. Our most recent system uses a wide angle of view fisheye camera which enables full pipe coverage and is sufficiently compact for practical applications. Our approach to mapping and localization builds from state-of-the-art visual odometry methods and extends them to deal with (visually) feature poor engineered environments. We present the results of this work using image datasets collected within our constructed pipe network. A range of empirical results are presented to validate the approach.

Arabic is a morphologically rich language. This morphological complexity results in a high out-of-vocabulary rate. That is why a lookup table for pronunciation modeling is not efficient for large vocabulary tasks. In previous research, graphemic modeling was proposed by approximating pronunciation modeling to be graphemes rather than actual phonemes. In this research, we have proposed a hybrid acoustic and pronunciation modeling approach for Arabic large vocabulary speech recognition tasks. The proposed approach benefits from both phonemic and graphemic modeling techniques, where two acoustic models are fused together. The hybrid approach also benefits from both vocalized and non-vocalized Arabic resources, which is useful because the amount of non-vocalized resources is always higher than vocalized ones. Two speech recognition baseline systems were built: phonemic and graphemic. The two baseline acoustic models were combined after two independent trainings to create a hybrid model. Pronunciation modeling was also hybrid by generating graphemic variants in addition to phonemic variants. Three techniques are proposed for pronunciation modeling: Hybrid-Or, Hybrid-And, and Hybrid-Top(n). In Hybrid-Or, either graphemic or phonemic modeling is applied for any given word. In Hybrid-And, a graphemic pronunciation is always generated in addition to existing phonemic pronunciations. Hybrid-Top(n) is a mixture of Hybrid-Or and Hybrid-And by applying Hybrid-Or on the top n high frequency words. Experiments were conducted in the large vocabulary news broadcast speech domain with a vocabulary size of 250K. The proposed hybrid approach has shown a relative reduction in WER of 8.8% to 12.6% depending on pronunciation modeling settings and the supervision in the baseline systems. In large vocabulary speech domains, acoustic and pronunciation modeling is a common problem among all Arabic colloquial varieties. Thus, for future work, the proposed approach is currently being extended and evaluated with different Arabic colloquial varieties (e.g. Qatari, Egyptian, Levantine, etc.). Moreover, the proposed technique can be applied with other morphologically rich languages like Turkish, Finnish, Korean, etc. This work was funded by a grant from the Qatar National Research Fund under its National Priorities Research Program (NPRP) award number NPRP 09-410-1-069. Reported experimental work was performed at Qatar University in collaboration with University of Illinois.

Relaying technologies have been designated as major new enabling technologies for next generation wireless broadband systems, such as 3GPP LTE-Advanced. The practical deployment of decode-and-forward (DF) relaying technologies as supported by the current standard is based on repetition coding, meaning that the relay regenerates the same codeword generated by the source node. This scheme is suboptimal. Nevertheless, it is preferred in practice due to its simple implementation. The optimal relaying schemes called cooperative coding are difficult to construct in practice and require heavy computation load at the receiver. Therefore, they are rarely implemented despite their prospected performance gains. As a simple and practical alternative, we propose a novel relaying scheme that provides a superior performance similar to the advanced cooperative coding techniques, but it is less complex to implement. Our novel scheme is called "decode-partial-forward" (DPF) because its fundamental concept of operation is based on making the relay forwards one part of the source message such that the receiver at the destination node can rely on the direct channel with the source to obtain the other part of the message. The DPF scheme is based favorably on repetition coding and maximal ratio combining techniques, which are standardized techniques with low-complexity and low computation load. Nevertheless, our novel scheme performs very close to the optimal bounds for the supported rates over relaying links. More specifically, the increase in the supported reliable transmission rates (bits/sec) using the proposed DPF scheme may exceed 30% of the supported rates using the conventional repetition coding DF scheme. Another major advantage of our proposed scheme is that it is easy to adapt flexibly based on the channel conditions of the three links of the relay channel (source-destination, source-relay, relay-destination), by adjusting the power and rate allocation at the source and relay using simple closed-form analytic formulas. Therefore, we believe that the DPF relaying scheme is an excellent option for practical deployment in the telecommunication standards due to its simplicity, adaptability and high spectral efficiency gains.

The widespread deployment of varying networking technologies, coupled with the exponential increase in end-user data demand has led to the proliferation of multi-interface enabled devices. To date, these interfaces are mainly utilized independently based on network availability, cost, and user choice. While researchers have focused on simultaneously leveraging these interfaces by aggregating their bandwidths, these solutions however, have faced a steep deployment barrier and only focused on maximizing throughput while overlooking the energy awareness which is critical for mobile devices. We therefore developed OPERETTA, shown in Figure1, an optimal deployable energy efficient bandwidth aggregation system for mobile users. Our system does not require modifications to applications, legacy servers, network infrastructure, or client kernel. If legacy servers choose to adopt our system, however, OPERETTA dynamically leverages this to achieve higher performance gains. OPERETTA is built as a middle-ware that is responsible for scheduling various connections and/or packets to different interfaces. This middleware estimates application and network interface characteristics and utilizes these estimates to take the most appropriate scheduling decisions. We formulate our scheduling problem as a mixed integer programming problem that has a special structure allowing it to be efficiently solved. This formulation allows users to achieve a desired throughput with minimal energy consumed. We evaluate OPERETTA via prototype implementation on the Windows OS, as well as via simulation, and compare the results to the optimal achievable throughput and energy consumption. Our results show that, with no changes to the current legacy servers, OPERETTA can achieve up to 150% enhancement in throughput as compared to the current operating systems, with no increase in energy consumption. In addition, with only 25% of the servers being OPERETTA-enabled, the system performance reaches the throughput upper-bound. We ultimately demonstrate that OPERETTA achieves the goals of being optimal, energy-efficient, as well as easily and incrementally deployable.

In wireless cellular systems, uneven traffic load among the cells increases call blocking rates in some cells and causes low resource utilisation in others and thus degrades user satisfaction and overall performance of the cellular system. Various centralised or semi-centralised Load Balancing (LB) schemes have been proposed to cope with this time persistent problem, however, a fully distributed Self Organising (SO) LB solution is still needed for the future cellular networks. To this end, we present a novel distributed LB solution based on an analytical framework developed on the principles of nature-inspired SO systems. A novel concept of super-cell is proposed to decompose the problem of "system-wide blocking minimization" into the local sub-problems in order to enable a SO distributed solution. Performance of the proposed solution is evaluated through system level simulations for both macro cell and femto cell based systems. Numerical results show that the proposed solution can reduce the blocking in the system close to an Ideal Central Control (ICC) based LB solution. The added advantage of the proposed solution is that it does not require heavy signalling overheads.

Background and Objectives: Radio frequency identification (RFID) has paved the way for a plethora of monitoring applications in the field of oil and gas. Degradation in liquefied petroleum gas (LPG)/ liquefied natural gas (LNG) quality and pipe infrastructure can be a nuisance for the oil and gas industry in Qatar. Hence, a human centric cyber-physical-system (CPS) utilizing hybrid wireless technologies including RFIDs and other sensor motes can detect and prevent such hazards. CPS technique can be used for the oil and gas sector in Qatar with customized framework architecture, event detection and decision algorithms. The objective of this research is to allow maintainers and administrators to perceive and decide on top of the monitoring system to increase the performance and efficiency of the whole monitoring application. Methods: The sensors collect the data and send it to the base station through collaborating wireless technologies. At the base station the data is processed and algorithms were run to detect an event such as presence of moisture, abnormal pressure, temperature and defects in a pipe's infrastructure health. Human interaction will help to further refine the data for possible false alarms. Mobile applications can be used by the users/administrators to send details of a perceived event directly to the base station. Results: Experiments were performed on the moisture detection in Wireless Research Laboratory in Texas A&M University at Qatar. On the similar note, other sensors can also be associated with the RFIDs and their data can be relayed to the server. A framework architecture was proposed for the human centric approach of the detection and monitoring system. Conclusion: We propose a system where these RFID active tags with the sensors such as pressure, temperature, flow, strain etc., in collaboration with other wireless technologies, are able to send information regarding the gas quality and pipe's infrastructure health. The collaboration of RFIDs and other technologies helps us to create a smart human centric monitoring system. This will enhance maintenance and event detection such as the presence of moisture or strain on pipe's structure. It will additionally assist in the automatic adjustment of the valve's and pump's properties according to the detected events.

The proliferation of mobile devices equipped with communication technologies such as WiFi, 3G, and 4G, coupled with the exponential growth of online social networking, have increased the demand for social-context-aware systems. These systems leverage social information provided by users with contextual awareness, particularly location, to provide real-time personalized services. With Euromonitor International indicating a 37.7 % growth in mobile phone penetration in the past five years--ICTQatar recently reporting that each household owns 3.9 devices on average--Qatar is positioned as a strong candidate not only for the consumption of such services, but also for researching, building, and testing solutions related to context-aware system challenges. We address some of these challenges and propose pragmatic solutions. Our contributions fall into the following two categories: (i) We design and implement SCOUT, a context-aware, ubiquitous system that enables real-time mobile services by providing contextually relevant information. This information can either be generated reactively based on user request, or proactively created and disseminated to potentially interested users. Our SCOUT prototype consists of an android-based mobile client interfacing with facebook's API and a load-balancing profile-matching server that interacts with a localization engine. (ii) Since mobile users may not all experience reliable mobile-to-core connectivity due to contention, cost, or lack of connectivity, we therefore extend the reach of context-aware services to disconnected mobile users by proposing a novel communication paradigm that leverages opportunistic communication. We first send the intended information to the smallest subset of connected users deemed sufficient to reach the disconnected destinations. This subset then selectively forwards, based on social profiles, this information to nodes that are more likely to meet the destinations. Our evaluation, via simulation, shows that our algorithm achieves an improvement of 25% to 80% compared to current communication paradigms, while reducing overhead by as much as 50%. With SCOUT currently operational, and based on the simulation results, our ongoing work includes integrating our communication paradigm into the real system. We are also working on integrating real-time group recommender systems that identify groups based on user location coupled with social information to provide real-time contextual group recommendations.

Background & Objectives: As the electromagnetic spectrum resource is becoming highly scarce, improving spectral efficiency is becoming extremely important for the development of future wireless communication systems. Integrating cooperative relaying techniques into wireless communication systems sheds new light on better spectral efficiency. By taking advantage of the broadcast nature of wireless communications, cooperative transmission can be used in improving communication reliability, and enhancing power and spectrum efficiency. Moreover, comparing to other emerging techniques that could achieve similar performance advantages, such as multiple-input multiple-output (MIMO) technique, cooperative communication is superior in hardware feasibility and deployment flexibility. The important advantages of cooperative communication make it one of the promising techniques for future wireless communication systems. Recently, many cooperative communication schemes have been included in different cellular standards, such as WiMAX and LTE-Advanced. However, the promised throughput and diversity gain may be lost in the presence of interferences, and hence, interference management is very important for exploiting the benefits of cooperation. This requires the need to find proper methods to prevent the interference problems, which is the main target of our current research. Methods: In this study we introduce an efficient cooperative communications scheme which maximizes the received signal-to-noise ratio (SNR) while keeping the interference levels below a certain threshold. The introduced scheme is based on two relay selection methods: max (link 2) which is based on maximizing the SNR of the second hop, and max (e2e) which aims to maximize the end-to-end SNR of the relayed link. In this method, the perfect decoding-and-forward (DF) relaying protocol is used by the relays, and the maximum ratio combining (MRC) receiver is used to combine the direct and the best selected relay links. We derive exact closed form expressions for the probability density function (PDF) of the SNR, outage probability, average capacity, and average bit error probability for the introduced cooperative schemes. Results & Conclusion: Simulations are used to validate the analytical results and an agreement is observed. The results confirm the advantage of the introduced cooperation schemes in enhancing the wireless communication system and in interference management.

Geographic information systems (GIS) (also known as Geospatial information systems or Geotechnology) are computer software and hardware systems designed to capture, store and manipulate all types of geographical data, as well as analyze, manage, and display geographic information for informing decision making. Users of GIS range from communities, research institutions, environmental scientists, health organisations, land use planners, businesses, and government agencies at all levels. Numerous examples of applications of GIS are available in many different journals and are frequent topics of presentations at conferences in the natural and social sciences. For a long time, GIS has been a well-established and independent discipline throughout American and European Universities; however, there is no single University in the Arab World or even in the Middle East--as far as the author knows--that offers an undergraduate program for this important discipline. Qatar can be the pioneer in this field in the region by offering such a program through Qatar University or any higher education institute in the state. The benefits of opening such a program are numerous to be calculated. Qualified human resources in the field of GIS are in high demand not only in the region but at international level as well. The U.S. Department of Labor has designated Geotechnology (GIS) as one of the three "mega-technologies" of the 21st century—right up there with Nanotechnology and Biotechnology. Opening GIS undergraduate program in Qatar will make the state The "Kaaba" of this science in the region and will attract students from different states, which will generate scientific and financial revenue. Moreover, the host institute of this suggested program can establish an international research centre for GIS to carry out studies for the benefit of the region and beyond. Implementing such a proposal will have great impact on the Computing and Information Technology Research discipline, which is one of Qatar's core research areas.