Electronic systems of the future require a very high bandwidth communications infrastructure within the system. This way the massive amount of compute power which will be available can be inter-connected to realize future powerful advanced electronic systems. Today, electronic inter-connects between 3D chip-stacks, as well as intra-connects within 3D chip-stacks are approaching data rates of 100 Gbit/s soon. Hence, the question to be answered is how to efficiently design the communications infrastructure which will be within electronic systems. Within this paper approaches and results for building this infrastructure for future electronics are addressed.

Network Coding is a new possibility to transmit data through a network. By combining different packets instead of simply forwarding, network coding offers the opportunity to reach the Min-Cut/Max-Flow capacity in multicast data transmissions. However, the basic schemes are vulnerable to so-called pollution attacks, where an attacker can jam large parts of the transmission by infiltrating only one bogus message. In the literature we found several approaches which aim at handling this kind of attack with different amounts of overhead. Though, the cost for a specific secure network coding scheme highly depends on the underlying network. The goal of this paper is on the one hand to describe which network parameters influence the efficiency of a certain scheme and on the other hand to provide concrete suggestions for selecting the most efficient secure network coding scheme considering a given network. We will illustrate that there does not exist “the best” secure network scheme concerning efficiency, but all selected schemes are more or less suited under certain network topologies.

Probabilistic-Write/Copy-Select (PWCS) is a novel synchronization scheme suggested by Nicholas Mc Guire which avoids expensive atomic operations for synchronizing access to shared objects. Instead, PWCS makes inconsistencies detectable and recoverable. It builds on the assumption that, for typical workloads, the probability for data races is very small. Mc Guire describes PWCS for multiple readers but only one writer of a shared data structure. In this paper, we report on the formal analysis of the PWCS protocol using a continuous-time Markov chain model and probabilistic model checking techniques. Besides the original PWCS protocol, we also considered a variant with multiple writers. The results were obtained by the model checker PRISM and served to identify scenarios in which the use of the PWCS protocol is justified by guarantees on the probability of data races. Moreover, the analysis showed several other quantitative properties of the PWCS protocol.

Abstract The current Internet of Things solutions for simple measurement and monitoring tasks are evolving into ubiquitous sensor networks that are constantly observing both our well being and the conditions of our living environment. The oncoming omnipresent wireless infrastructure is expected to feature artificial intelligence capabilities that can interpret human actions, gestures and even needs. All of this will require processing power on a par with and energy efficiency far beyond that of the current mobile devices. The current Internet of Things devices rely mostly on commercial low power off-the-shelf micro-controllers. Optimized solely for low power, while paying little attention to computing performance, the present solutions are far from achieving the energy efficiency, let alone, the compute capability requirements of the future Internet of Things solutions. Since this domain is application specific by nature, the use of general purpose processors for signal processing tasks is counterintuitive. Instead, dedicated accelerator based solutions are more likely to be able to meet these strict demands. This thesis proposes one potential solution for achieving the necessary low energy, as well as the flexibility and performance requirements of the Internet of Things domain in a cost effective manner using reconfigurable heterogeneous processing solutions. A novel graphics processing unit-style accelerator for the Internet of Things application domain is presented. Since the accelerator can be reconfigured, it can be used for most applications of the Internet of Things domain, as well as other application domains. The solution is assessed using two computer vision applications, and is demonstrated to achieve an excellent combination of performance and energy efficiency. The accelerator is designed using an efficient and rapid co-design flow of software and hardware, featuring ease of development characteristics close to commercial off-the-shelf solutions, which also enables cost-efficient design flow. / Tiivistelmä Esineiden internet tulee muuttamaan tulevaisuudessa elinympäristömme täysin. Se tulee mahdollistamaan interaktiiviset ympäristöt nykyisten passiivisten ympäristöjen sijaan. Lisäksi elinympäristömme tulee reagoimaan tekoihimme ja puheeseemme sekä myös tunteisiimme. Tämä kaikkialla läsnä olevan langaton infrastruktuuri tulee vaatimaan ennennäkemätöntä laskentatehokkuutta yhdistettynä äärimmäiseen energiatehokkuuteen. Nykyiset esineiden internet ratkaisut nojaavat lähes täysin kaupallisiin "suoraan hyllyltä" saataviin yleiskäyttöisiin mikrokontrollereihin. Ne ovat kuitenkin optimoituja pelkästään matalan tehonkulutuksen näkökulmasta, eivätkä niinkään energiatehokkuuden, saati tulevaisuuden esineiden internetin vaatiman laskentatehon suhteen. Kuitenkin esineiden internet on lähtökohtaisesti sovelluskohtaista laskentaa vaativa, joten yleiskäyttöisten prosessoreiden käyttö signaalinkäsittelytehtäviin on epäloogista. Sen sijaan sovelluskohtaisten kiihdyttimien käyttö laskentaan, todennäköisesti mahdollistaisi tavoitellun vaatimustason saavuttamisen. Tämä väitöskirja esittelee yhden mahdollisen ratkaisun matalan energian kulutuksen, korkean suorituskyvyn ja joustavuuden yhdenaikaiseen saavuttamiseen kustannustehokkaalla tavalla, käyttäen uudelleenkonfiguroitavia heterogeenisiä prosessoriratkaisuja. Työssä esitellään uusi grafiikkaprosessori-tyylinen uudelleen konfiguroitava kiihdytin esineiden internet sovellusalueelle, jota pystytään hyödyntämään useimpien laskentatehoa vaativien sovellusten kanssa. Ehdotetun kiihdyttimen ominaisuuksia arvioidaan kahta konenäkösovellusta esimerkkinä käyttäen ja osoitetaan sen saavuttavan loistavan yhdistelmän energia tehokkuutta ja suorituskykyä. Kiihdytin suunnitellaan käyttäen tehokasta ja nopeaa ohjelmiston ja laitteiston yhteissuunnitteluketjua, jolla voidaan saavuttaa lähestulkoon kaupallisten "suoraan hyllyltä" saatavien prosessoreiden kehitystyön helppous, joka puolestaan mahdollistaa kustannustehokkaan kehitys- ja suunnittelutyön.

Modern database systems have to process huge amounts of data and should provide results with low latency at the same time. To achieve this, data is nowadays typically hold completely in main memory, to benefit of its high bandwidth and low access latency that could never be reached with disks. Current in-memory databases are usually columnstores that exchange columns or vectors between operators and suffer from a high tuple reconstruction overhead. In this paper, we present the indexed table-at-a-time processing model that makes indexes the first-class citizen of the database system. The processing model comprises the concepts of intermediate indexed tables and cooperative operators, which make indexes the common data exchange format between plan operators. To keep the intermediate index materialization costs low, we employ optimized prefix trees that offer a balanced read/write performance. The indexed tableat-a-time processing model allows the efficient construction of composed operators like the multi-way-select-join-group. Such operators speed up the processing of complex OLAP queries so that our approach outperforms state-of-the-art in-memory databases.

<p>Spiking Neural Networks (SNNs) have recently attracted significant research interest as the third generation of artificial neural networks that can enable low-power event-driven data analytics. We propose ANN-SNN conversion using “soft re-set” spiking neuron model, referred to as Residual Membrane Potential (RMP) spiking neuron, which retains the “resid- ual” membrane potential above threshold at the firing instants. In addition, we propose a time-based coding scheme, named Temporal-Switch-Coding (TSC), and a corresponding TSC spiking neuron model. Each input image pixel is presented using two spikes with opposite polarity and the timing between the two spiking instants is proportional to the pixel intensity. We demonstrate near loss-less ANN-SNN conversion using RMP neurons for VGG-16, ResNet-20, and ResNet-34 SNNs on challenging datasets including CIFAR-10, CIFAR-100, and ImageNet. With the help of TSC coding, it achieves 7-14.5× less inference latency, and 30-60× fewer addition operations and memory accesses per inference across datasets compared to the state of the art (SOTA) SNN models. In the second part of the thesis, we propose a new type of recurrent neural network (RNN) architecture, named Os- cillatory Fourier Neural Network (O-FNN). We demonstrate that O-FNN is mathematically equivalent to a simplified form of Discrete Fourier Transform applied onto periodical activa- tion. In particular, the computationally intensive back-propagation through time in training is eliminated, leading to faster training while achieving the SOTA inference accuracy in a diverse group of sequential tasks. For instance, applying the proposed model to sentiment analysis on IMDB review dataset reaches 89.4% test accuracy within 5 epochs, accompanied by over 35x reduction in the model size compared to Long Short-Term Memory (LSTM). The proposed novel RNN architecture is well poised for intelligent sequential processing in resource constrained hardware.</p>

<p>Next-generation (Next-Gen) mobile applications, Extended Reality (XR), which encompasses Virtual/Augmented/Mixed Reality (VR/AR/MR), promise to revolutionize how people interact with technology and the world, ushering in a new era of immersive experiences. However, the hardware capacity of mobile devices will not grow proportionally with the escalating resource demands of the mobile apps due to their battery constraint. To bridge the gap, edge computing has emerged as a promising approach. It is further boosted by emerging 5G cellular networks, which promise low latency and high bandwidth. However, realizing the full potential of edge computing faces several fundamental challenges.</p> <p><br></p> <p>In this thesis, we first discuss a set of fundamental design challenges in supporting Next-Gen mobile applications via edge computing. These challenges extend across the three key system components involved — mobile clients, edge servers, and cellular networks. We then present how we address several of these challenges, including (1) how to coordinate mobile clients and edge servers to achieve stringent QoE requirements for Next-Gen apps; (2) how to optimize energy consumption of running Next-Gen apps on mobile devices to ensure long-lasting user experience; and (3) how to model and generate control-plane traffic of cellular networks to enable innovation on mobile network architectural design to support Next-Gen apps not only over 4G but also over 5G and beyond.</p> <p><br></p> <p>First, we present how to optimize the latency in edge-assisted XR system via the mobile-client and edge-server co-design. Specifically, we exploit key insights about frame similarity in VR to build the first multiplayer edge-assisted VR design, Coterie. We demonstrate that compared with the prior work on single-player VR, Coterie reduces the per-player network load by 10.6X−25.7X, and can easily support 4 players for high-quality VR apps on Pixel 2 over 802.11ac running at 60 FPS and under 16ms responsiveness without exhausting the finite wireless bandwidth.</p> <p><br></p> <p>Second, we focus on the energy perspective of running Next-Gen apps on mobile devices. We study a major limitation of a classic and de facto app energy management technique, reactive energy-aware app adaptation, which was first proposed two decades ago. We propose, design, and validate a new solution, the first proactive energy-aware app adaptation, that effectively tackles the limitation and achieves higher app QoE while meeting a given energy drain target. Compared with traditional approaches, our proactive solution improves the QoE by 44.8% (Pixel 2) and 19.2% (Moto Z3) under low power budget.</p> <p><br></p> <p>Finally, we delve into the third system component, cellular networks. To facilitate innovation in mobile network architecture to better support Next-Gen apps, we characterize and model the control-plane traffic of cellular networks, which has been mostly overlooked by prior work. To model the control-plane traffic, we first prove that traditional probability distributions that have been widely used for modeling Internet traffic (e.g., Poisson, Pareto, and Weibull) cannot model the control-plane traffic due to the much higher burstiness and longer tails in the cumulative distributions of the control-plane traffic. We then propose a two-level state-machine-based traffic model based on the Semi-Markov model. We finally validate that the synthesized traces by using our model achieve small differences compared with the real traces, i.e., within 1.7%, 4.9% and 0.8%, for phones, connected cars, and tablets, respectively. We also show that our model can be easily adjusted from LTE to 5G, enabling further research on control-plane design and optimization for 4G/5G and beyond.</p>

Network Coding is a new possibility to transmit data through a network. By combining different packets instead of simply forwarding, network coding offers the opportunity to reach the Min-Cut/Max-Flow capacity in multicast data transmissions. However, the basic schemes are vulnerable to so-called pollution attacks, where an attacker can jam large parts of the transmission by infiltrating only one bogus message. In the literature we found several approaches which aim at handling this kind of attack with different amounts of overhead. Though, the cost for a specific secure network coding scheme highly depends on the underlying network. The goal of this paper is on the one hand to describe which network parameters influence the efficiency of a certain scheme and on the other hand to provide concrete suggestions for selecting the most efficient secure network coding scheme considering a given network. We will illustrate that there does not exist “the best” secure network scheme concerning efficiency, but all selected schemes are more or less suited under certain network topologies.

Probabilistic-Write/Copy-Select (PWCS) is a novel synchronization scheme suggested by Nicholas Mc Guire which avoids expensive atomic operations for synchronizing access to shared objects. Instead, PWCS makes inconsistencies detectable and recoverable. It builds on the assumption that, for typical workloads, the probability for data races is very small. Mc Guire describes PWCS for multiple readers but only one writer of a shared data structure. In this paper, we report on the formal analysis of the PWCS protocol using a continuous-time Markov chain model and probabilistic model checking techniques. Besides the original PWCS protocol, we also considered a variant with multiple writers. The results were obtained by the model checker PRISM and served to identify scenarios in which the use of the PWCS protocol is justified by guarantees on the probability of data races. Moreover, the analysis showed several other quantitative properties of the PWCS protocol.

Nowadays, our daily live concerns the usage of Information Technology, increasingly. As a result, a huge amount of data has to be processed which is outsourced from local devices to data centers. Due to fluctuating demands these are not fully utilized all the time and consume a significant amount of energy while idling. A common approach to avoid unnecessary idle times is to consolidate running services on a subset of machines and switch off the remaining ones. Unfortunately, the services on a single machine interfere with each other due to the competition for shared resources such as caches after the consolidation, which leads to a degradation of performance. Hence, data centers have to trade off between reducing the energy consumption and certain performance criteria defined in the Service Level Agreement. In order to make the trade off in advance, it is necessary to characterize services and quantify the impact to each other after a potential consolidation. Our approach is to use random variables for characterization, which includes the fluctuations of the resource consumptions. Furthermore, we would like to model the interference of services to provide a probability of exceeding a certain performance criterion.