The ongoing miniaturization of semiconductor manufacturing technologies has enabled the integration of tens to hundreds of processing cores on a single chip. Unlike frequency-scaling where performance is increased equally across the board, core-scaling and hardware thread-scaling harness the additional processing power through the concurrent execution of multiple processes or programs. This approach of mingling or interleaving process executions has engendered a new set of security challenges that risks to undermine nearly three decades’ worth of computer architecture design efforts. The complexity of the runtime interactions and aggressive resource sharing among processes, e.g., caches or interconnect network paths, have created a fertile ground to mount attacks of ever-increasing acuteness against these computer systems. One such class of attacks is cache side-channel attacks. While caches are vital to the performance of current processors, they have also been the target of numerous side-channel attacks. As a result, a few cache architectures have been proposed to defend against these attacks. However, these designs tend to provide security at the expense of performance, area and power. Therefore, the design of secure, high-performance cache architectures is still a pressing research challenge. In this thesis, we examine the viability of self-aware adaptive caches as a defense mechanism against cache side-channel attacks. We define an adaptive cache as a caching structure with (i) run-time reconfiguration capability, and (ii) intelligent built-in logic to monitor itself and determine its parameter settings. Since the success of most cache side-channel attacks depend on the attacker’s knowledge of the key cache parameters such as associativity, set count, replacement policy, among others, an adaptive cache can provide a moving target defense approach against many of these cache side-channel attacks. Therefore, we hypothesize that the runtime changes in certain cache parameters should render some of the side-channel attacks less effective due to their dependence on knowing the exact configuration of the caches. / 2020-06-03T00:00:00Z

This master’s degree thesis exists in two parts: a critical introduction and an original novella entitled Side Effects. The critical introduction introduces and explains the theories on, literature surrounding, and literary uses of dystopian fiction, the novella format, and drug-based psychotherapy. Current opinion on dystopian fiction sees it characterized by a seemingly perfect societal setting that ultimately contains hidden or suppressed moral flaws. The ultimate purpose of dystopian fiction is commentary on contemporary society through a defamiliarized setting. The novella format is shown to exist in a middle-ground state between the short story and the novel, yet the format manages to maintain positive literary elements of both. Finally, a discussion on drug-based psychotherapy illustrates the use of chemical compounds to treat or cure psychological conditions, a topic of much debate amongst current psychology practitioners. The section on drug-based psychotherapy focuses largely on memoirs for purposes of first-hand experience and character creation for the original novella. The novella, entitled Side Effects, follows the character Edward, a middle-aged man who creates and tests serums that suppress by mandate the emotions that his society deems toxic to the human condition. Edward remains ignorant of any life outside the symmetry and order of the Company, the corporation responsible for the maintenance of the society. That is, until a chance encounter with a young woman named Gabrielle causes Edward to explore a world outside the confines of his carefully crafted city and lifestyle. She introduces him to a community of people who reject the mandates of the Company and exist as the extreme opposition to its ideals. As Edward spends more time with this group, known as Splicers, he must confront his long-held standards and finally choose for himself what life he will live.

No description available.

No description available.

No description available.

No description available.

Sweden has since a few years back suffered from a bottleneck in the power transmission lines stemming from a lack of “space” on the grid that transports the power that is generated in the northern parts to the middle and southern parts of the country. A long-term solution would be to increase the grid's capacity by building more transmission lines, however in the meantime, a short-term solution would be to increase our power flexibility. Meaning that the energy consumption gets moved from hours of high demand to hours with low demand and by shutting off equipment. This is called power flexibility and has been researched more in recent times in projects like this. This project will mainly focus on demand-side flexibility which is about how the consumers use their electrical power. In collaboration with Uppsala Arenor och Fastigheter, power usage in one of the company’s properties will be made more flexible. This is achieved by mapping how much power each part of the property utilizes and estimating how much power the electrical equipment in that area consumes. Then determining if that piece of equipment can be either rescheduled to avoid hours of high-powerdemand in Uppsala or turned off without causing major consequences to the property. From this it will be decided if it can be added as a flexibility resource and after that the total power flexibility of the facility will be tallied. Lastly, suggestions on how to implement the power flexibility resources and recommendations for further improvements in future projects will be made. The results of the project were that the selected property Studenternas had an average power usage of 185kW during the day and of that 38% could be utilized as a power flexibility resource. The total power flexibility was therefore 71.2 kW which comes from the laundry rooms and the arena lighting while the other places in the facility contributed with a neglectable amount.

Deep Neural Networks (DNNs) have become ubiquitous due to their performance on prediction and classification problems. However, they face a variety of threats as their usage spreads. Model extraction attacks, which steal DNN models, endanger intellectual property, data privacy, and security. Previous research has shown that system-level side channels can be used to leak the architecture of a victim DNN, exacerbating these risks. We propose a novel DNN architecture extraction attack, called EZClone, which uses aggregate rather than time-series GPU profiles as a side-channel to predict DNN architecture. This approach is not only simpler, but also requires less adversary capability than earlier works. We investigate the effectiveness of EZClone under various scenarios including reduction of attack complexity, against pruned models, and across GPUs with varied resources. We find that EZClone correctly predicts DNN architectures for the entire set of PyTorch vision architectures with 100\% accuracy. No other work has shown this degree of architecture prediction accuracy with the same adversarial constraints or using aggregate side-channel information. Prior work has shown that, once a DNN has been successfully cloned, further attacks such as model evasion or model inversion can be accelerated significantly. Then, we evaluate several mitigation techniques against EZClone, showing that carefully inserted dummy computation reduces the success rate of the attack.

This research developed a double-side power electronics packaging scheme for high temperature applications exemplified by 1200 V, 150 A silicon devices. The power modules, based on both quarter and half-bridge topologies, were assembled using sintered silver device attachment rather than conventional solder alloys. Thermomechanical stresses in the double-side architecture were mitigated with a compliant layer fabricated from elliptical silver tubes. This research presents an introduction to conventional packaging techniques and their weaknesses. These shortcomings provide the basis for a module design which improves upon module thermal management while also addressing electrical and reliability requirements. The optimum package design enhances heat dissipation with the addition of a substrate bonded to the top electrical pads of the semiconductor devices. The use of sintered silver also increases the useful application temperature by avoiding the creep failure mechanisms of solder alloys. The modules were characterized extensively to quantify thermal and electrical performance. In the case of thermal characterization, the double-side architecture required multiple testing configurations to fully understand the parallel heat flow paths. These results were compared to models constructed using finite element analysis (FEA). The FEA models were also utilized for measurement of strains in multiple package designs to better determine the effects of increased compliance on the relative package cycling lifetime. These lifetimes were then assessed, in part, using experimental passive and cycling tests on functional double-side packages. The resulting power modules exhibited significant decreases in thermal resistance when they are cooled, as designed, from both sides of the module. Even single sided cooling options reveal significant advantages and transient thermal impedance was found to be significantly lower. Power module models revealed the compliant layer was successful in reducing the device shear stresses which was experimentally validated through the use of DC power stage testing. It was found, through double pulse testing and electrical modeling, that parasitic inductances were reduced by utilizing planar bonding and planar symmetrical traces. Finally, modeling of the double-side package with added tube compliance revealed a decrease in plastic and shear strains when compared to other single and double-side package designs. This reduction directly translates to increased cycling lifetime using well known strain based fatigue models. / Ph. D.

Over half (57%) of the women diagnosed with breast cancer are age 65 and older. Treatment for breast cancer may exacerbate current chronic illnesses and/or cause multiple treatment related side effects such as insomnia, fatigue, decreased physical functioning, alterations in body image, poorer quality of life, and changes in psychosocial health. While many women with breast cancer experience these changes, research suggests that older women have different needs than younger women and may not always benefit from interventions. The purpose of this integrative review of literature was to evaluate interventions designed to improve treatment related side effects in female breast cancer survivors age 65 years and older. This review of literature was conducted using CINAHL, PsycINFO, and MEDLINE databases using various key terms. Inclusion criteria consisted of peer reviewed research articles, women who have experience breast cancer, interventions directed at decreasing side effects, and research articles written in the English language.While using these search criteria, no interventions were found therefore, the age group was lowered to include women 50 and older. Eleven studies met the inclusion criteria. Interventions addressed a variety of treatment related side effects and were delivered in multiple formats. The findings indicate that interventions resulted in a significant improvement in sleep and fatigue, physical function, perception of body image, psychosocial health, and quality of life for older women with breast cancer. While these findings are positive, the literature did not break down results based on developmental stage or "older" age groups. Currently, there is limited literature that examines interventions in women age 65 and older, this may limit nurses' ability to suggest successful interventions to some of our oldest cancer survivors.