Theoretical Computer Science is hardly represented in schools and often comes as a surprise to students of Computer Science with its use of formal and mathematical notation. It might prove useful to broaden the view of Computer Science in schools to give them a more realistic view of what studying Computer Science is about. This paper reports on a course held in a german secondary education school where the flipped classroom method was used to teach basics of automata theory. Although the results of the final course exam were good and a short questionnaire filled out by the students suggests interest in the method the results from the teacher's perspective are mixed.

HARNESS is a next generation cloud-computing platform that offers commodity and specialized resources in support of large-scale data processing applications. We focus primarily on application domains that are currently not well supported by today's cloud providers, including the areas of scientific computing, business-analytics, and online machine learning. These applications often require acceleration of critical operations using devices such as FPGAs, GPGPUs, network middleboxes, and SSDs. We explain the architectural principles that underlie the HARNESS platform, including the separation of agnostic and cognizant resource management that allows the platform to be resilient to heterogeneity while leveraging its use. We describe a prototype implementation of the platform, which was evaluated using two testbeds: (1) a heterogeneous compute and storage cluster that includes FPGAs and SSDs and (2) Grid'5000, a large-scale distributed testbed that spans France. We evaluate the HARNESS cloud-computing platform with two applications: Reverse-Time Migration, a scientific computing application from the geosciences domain, and AdPredictor, a machine learning algorithm used in the Bing search engine.

This paper presents a novel approach for allocating resources in systems with multiple accelerators. It has three main contributions. First, a new model based on Birkhoff's representation theory in capturing the ordering properties of resource allocation requests (RArs). Second, an effective technique for resource allocation based on this model, targeting systems with multiple accelerators. Third, the evaluation of the proposed approach for Maxeler MPC-X multi-accelerator systems, demonstrating time-efficiency and 30%-50% failure-rate decrease (FRD) on random input dataset.

In this paper, we investigate the applicability of parallel sequence spread spectrum (PSSS) modulation in high bandwidth millimeter wave communication systems using a 240GHz monolithic microwave integrated circuit (MMIC) based radio frequency (RF) frontend. The transmitter (Tx) and receiver (Rx) are connected directly via an attenuator in an incoherent back-to-back setup using two independent local oscillator (LO) sources. Although the system was restricted to phase modulated signals due to limited linearity when using conventional modulated signals, a transmission using a multivalent PSSS sequence with a spectral efficiency of 1 bit/s/Hz succeeded thanks to the introduced coding gain. The recovered data is evaluated in terms of eye opening and bit error rate (BER). At a gross data rate of 20 Gbit/s, an eye opening of 40% and a BER of 5.4 × 10-5 were observed.

The dependencies among the components of service-oriented software applications hosted in a mobile ad hoc network (MANET) are difficult to determine due to the inherent loose coupling of the services and the transient communication topologies of the network. Yet understanding these dependencies is critical to making good management decisions, since dependence data underlie important analyses such as fault localization and impact analysis. Current methods for discovering dependencies, developed primarily for fixed networks, assume that dependencies change only slowly and require relatively long monitoring periods as well as substantial memory and communication resources, all of which are impractical in the MANET environment. We describe a new dynamic dependence discovery method designed specifically for this environment, yielding dynamic snapshots of dependence relationships discovered through observations of service interactions. We evaluate the performance of our method in terms of the accuracy of the discovered dependencies, and draw insights on the selection of critical parameters under various operational conditions. Although operated under more stringent conditions, our method is shown to provide results comparable to or better than existing methods.

Alexander Wolf, President of the Association for Computing Machinery and Imperial College London, introduces the Ada Lovelace Symposium.

other Share on ACM's annual report for FY15 Author: Alexander L. Wolf View Profile Authors Info & Claims Communications of the ACMVolume 59Issue 1January 2016 pp 11–16https://doi.org/10.1145/2847218Published:21 December 2015Publication History 0citation2,943DownloadsMetricsTotal Citations0Total Downloads2,943Last 12 Months28Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

The management of large scale service based systems hosted in dynamic environments of hybrid wireless networks is a complex task. Such systems are designed as composite services integrating heterogeneous service components such as sensor, software and data services. In this paper, we propose a new situational aware framework which allows the network operators to effectively manage such complex systems. The framework integrates three key components. The Service Placement component and the Configuration Service which collectively arrive at decision about placements and compositions of the services. An instrumental element of the framework is the Policy Based Management System providing the network operators with a tool to dynamically control these mechanisms. 1

from the president I N AN EDITORIAL last month, ACM CEO John White reported on the outcomes of the November 2013 ACM Strategic Planning Retreat. The retreat generated a number of important new ideas in the areas of membership, conferences , publications, community, and practitioners. While they are likely to have a significant impact on ACM's activities going forward, I view them as defining only a relatively short-term agenda not fully addressing some deep issues facing ACM, issues we still need to understand and deal with. The challenges and opportunities of open access served as the original motivation for holding the retreat. Despite setting the modern standard for a liberal copyright policy in the digital age (over a decade ago), opening more content under the discretion of ACM authors and SIGs, fully embracing Green/Gold/Hybrid OA publishing, and complying with government mandates, there is a sense among a portion of our community that we have still not done enough. A sense that if an ACM publication sits behind any sort of paywall—regard-less of it also being freely available via an author's site, an author's institutional site, a SIG site, or even a conference site—we are somehow failing to meet our commitment to nurture the free flow of information. This position informed much of the retreat discussion. It led to our decision to encourage SIGs to open conference proceedings around the event and until its next occurrence, and it led to our decision to begin work on understanding the " article of the future , " in the broadest sense, along with the digital collection that might serve as host. But what it failed to do is lead us into a discussion of whether there can (even should be) a model for ACM in which publication revenue plays little or no role. ACM has a pretty straightforward business model. There are three major revenue streams: membership dues, conference registration fees, and publication subscription fees. Each has related sets of expenses. Membership runs at a loss—mainly because we subsidize students and members from developing countries. Conferences typically run at a surplus (not always—some conferences have lost considerable sums), but that surplus is retained by the SIGs and invested directly into serving and subsidizing their respective technical communities however they see fit. That leaves publications. Right now, publications overall (but not universally) generate a surplus. That surplus is used to underwrite the membership loss …

Cloud providers offer access to hardware accelerators, such as FPGAs and GPUs, according to a pay-per-use model. Data-parallel processing frameworks, such as MapReduce, make it easy for users to express parallel jobs, but an open challenge remains how such jobs can exploit accelerators in a cloud setting. While dedicated MapReduce frameworks for specific accelerators exist, users lack decision support on when accelerators can deliver substantial benefit for their jobs. We observe that the performance improvement attained by accelerators, if any, is dependent on both the data and the phase (map or reduce) of the MapReduce job. Based on a generalised model of shared-memory MapReduce, we describe HETMR, a MapReduce framework that can execute jobs in a hybrid fashion across CPUs and accelerators. While previous attempts at a hybrid solution simply allocate portions of job data to full MapReduce implementations residing on the CPU and accelerator, HETMR instead assigns a phase to each device and provides a means to move the data between the two. To avoid unnecessary accelerator costs and determine the best phase assignment, HETMR first profiles the MapReduce job on the CPU, yielding a measure of potential speedup. Depending on the outcome, it then deploys the job in either a CPU-only or hybrid fashion by drawing on a library of designs. Our experimental evaluation shows that HETMR’s hybrid execution model improves job completion by up to 6 times with an FPGA.

The combination of service-oriented applications, with their run-time service binding, and mobile ad hoc networks, with their transient communication topologies, brings a new level of complex dynamism to the structure and behavior of software systems. This complexity challenges our ability to understand the dependence relationships among system components when performing analyses such as fault localization and impact analysis. Current methods of dynamic dependence discovery, developed for use in fixed networks, assume that dependencies change slowly. Moreover, they require relatively long monitoring periods as well as substantial memory and communication resources, which are impractical in the mobile ad hoc network environment. We describe a new method, designed specifically for this environment, that allows the engineer to trade accuracy against cost, yielding dynamic snapshots of dependence relationships. We evaluate our method in terms of the accuracy of the discovered dependencies.

Abstract Fault localization in general refers to a technique for identifying the likely root causes of failures observed in systems formed from components. Fault localization in systems deployed on mobile ad hoc networks (MANETs) is a particularly challenging task because those systems are subject to a wider variety and higher incidence of faults than those deployed in fixed networks, the resources available to track fault symptoms are severely limited, and many of the sources of faults in MANETs are by their nature transient. We present a method for localizing the faults occurring in service-based systems hosted on MANETs. The method is based on the use of dependence data that are discovered dynamically through decentralized observations of service interactions. We employ both Bayesian and timing-based reasoning techniques to analyze the data in the context of a specific fault propagation model, deriving a ranked list of candidate fault locations. We present the results of an extensive set of experiments exploring a wide range of operational conditions to evaluate the accuracy of our method.

Early software environments have supported a narrow range of activities ( programming environments) or else been restricted to a single “hard-wired” software development process. The Arcadia research project is investigating the construction of software environments that are tightly integrated, yet flexible and extensible enough to support experimentation with alternative software processes and tools. This has led us to view an environment as being composed of two distinct, cooperating parts. One is the variant part, consisting of process programs and the tools and objects used and defined by those programs. The other is the fixed part, or infrastructure , supporting creation, execution, and change to the constituents of the variant part. The major components of the infrastructure are a process programming language and interpreter, object management system, and user interface management system. Process programming facilitates precise definition and automated support of software development and maintenance activities. The object management system provides typing, relationships, persistence, distribution and concurrency control capabilities. The user interface management system mediates communication between human users and executing processes, providing pleasant and uniform access to all facilities of the environment. Research in each of these areas and the interaction among them is described.