With the development of the Internet, new network services with strict network latency requirements have been made possible. These services are implemented as distributed systems deployed across multiple geographical locations. To provide low response time, these services require knowledge about the current network latency. Unfortunately, network latency among geo-distributed sites often change, thus distributed services rely on continuous network latency measurements. One goal of such measurements is to differentiate between momentary latency spikes from relatively long-term latency changes. The differentiation is achieved through statistical processing of the collected samples. This approach of high-frequency network latency measurements has high overhead, slow to identify network latency changes and lacks accuracy. We propose a novel approach for network latency estimation by correlating network paths to network latency. We demonstrate that network latency can be accurately estimated by first measuring and identifying the network path used and then fetching the expected latency for that network path based on previous set of measurements. Based on these principles, we introduce Sudan traceroute, a network latency estimation tool. Sudan traceroute can be used to both reduce the latency estimation time as well as to reduce the overhead of network path measurements. Sudan traceroute uses an improved path detection mechanism that sends only a few carefully selected probes in order to identify the current network path. We have developed and evaluated Sudan traceroute in a test environment and evaluated the feasibility of Sudan traceroute on real-world networks using Amazon EC2. Using Sudan traceroute we have shortened the time it takes for hosts to identify network latency level changes compared to existing approaches. / Med utvecklingen av Internet har nya nätverkstjänster med strikta fördröjningskrav möjliggjorts. Dessa tjänster är implementerade som distribuerade system spridda över flera geografiska platser. För att tillgodose låg svarstid kräver dessa tjänster kunskap om svarstiden i det nuvarande nätverket. Tyvärr ändras ofta nätverksfördröjningen bland geodistribuerade webbplatser, således är distribuerade tjänster beroende av kontinuerliga mätvärden för nätverksfördröjning. Ett mål med sådana mätningar är att skilja mellan momenta ökade svarstider från relativt långsiktiga förändringar av svarstiden. Differentieringen uppnås genom statistisk bearbetning av de samlade mätningarna. Denna högfrekventa insamling av mätningar av nätverksfördröjningen har höga overheadkostnader, identifierar ändringar långsamt och saknar noggrannhet. Vi föreslår ett nytt tillvägagångssätt för beräkningen av nätverksfördröjning genom att korrelera nätverksvägar till nätverksfördröjning. Vi visar att nätverksfördröjningen kan vara exakt uppskattad genom att man först mäter och identifierar den nätverksväg som används och sedan hämtar den förväntade fördröjningen för den nätverksvägen baserad på en tidigare uppsättning av mätningar. Baserat på dessa principer introducerar vi Sudan traceroute, ett Verktyg för att uppskatta nätverksfördröjning. Sudan traceroute kan användas för att både minska tiden att uppskatta fördröjningen samt att minska overhead för mätningarna i nätverket. Sudan traceroute använder en förbättrad vägdetekteringsmekanism som bara skickar några försiktigt valda prober för att identifiera den aktuella vägen i nätverket. Vi har utvecklat och utvärderat Sudan traceroute i en testmiljö och utvärderade genomförbarheten av Sudan traceroute i verkliga nätverk med hjälp av Amazon EC2. Med hjälp av Sudan traceroute har vi förkortat den tid det tar för värdar att identifiera nätverksfördröjnings förändringar jämfört med befintliga tillvägagångssätt.

With the rise of the usage of AI, the need for computing power has grown exponentially. This has made cloud computing a popular option with its cost- effective and highly scalable capabilities. However, due to its popularity there exists thousands of possible services to choose from, making it hard to find the right tool for the job. The purpose of this thesis is to provide a methodological approach for evaluating which alternative is the best for machine learning applications deployed in the cloud. Nine different instances were evaluated on a major cloud provider and compared for their performance relative to their cost. This was accomplished by developing a cost evaluation model together with a test environment for image recognition models. The environment can be used on any type of cloud instance to aid in the decision-making. The results derived from the specific premises used in this study indicate that the higher the hourly cost an instance had, the less cost-effective it was. However, when making the same comparison within an instance family of similar machines the same conclusion can not be made. Regardless of the conclusions made in this thesis, the problem addressed remains, as the domain is too large to cover in one report. But the methodology used holds great value as it can act as guidance for similar evaluation with a different set of premises. / Användingen av Artificiell Intelligens har aldrig varit så stor som den är idag och behovet av att kunna göra tyngre och mer komplexa beräkningar har växt exponentiellt. Detta har gjort att molnet, cloud, ett mycket populärt alternativt för sin kostadseffektiva och skalbara förmåga. Däremot så finns det tusentals alternativ att välja emellan vilket gör det svårt att hitta rätt verktyg för jobbet. Syftet med denna uppsats är att förse läsaren med en användbar metodik för att evaluera vilket instans som passar bäst för maskininlärnings applikationer som distribueras i molnet. Nio stycken olika instanser evaluerades på en molnleverantör genom att jämföra deras prestanda kontra deras kostnad. Detta gjordes genom att utveckla en kostnadsmodell tillsammans med en testmiljö för bildigenkänningsmodeller. Testmiljön som användes kan appliceras på flertal instanser som inte ingick i denna rapport för att tillåta andra att använda den för egna tester. Resultaten för studien var att de instanserna med högre timkostnad tenderar till att vara mindre kostnadseffektiva. Gör man samma jämförelse med endast instanser av samma typ som är anpassade för maskininlärning så är samma slutsats inte lika självklar. Oavsett slutsatser som ges i denna rapport så består problemet. Detta beror på att molnet berör så många olika faktorer som bör värderas i evalueringen, till exempel utvecklingstid och modellens förmåga att förutspå en bild vilket alla kräver sin egna tes. Men metodiken som används kan definitivt vara till stor nytta om man vill göra en liknande utvärdering med andra premisser.

At Halmstad University, the CatFish Project aims to measure and report on water quality. The System includes three components, one to measure, one to present, and one to handle data. This thesis explains the development of a cloud-based back-end solution created for the CatFish project. The solution connects IoT devices via the MQTT protocol. The devices are connected to collect and transfer data, later stored in a database, to be presented to a web application via a REST or WebSocket API. The solution is implemented with Amazon Web Services as a cloud service provider and is hosted on their platform. / Vid Högskolan i Halmstad ämnar CatFish projektet att mäta och rapportera om vattenkvalitet. Systemet innehåller tre komponenter, en för att mäta, en för att presentera och en för att hantera data. Denna rapport kommer detaljera utvecklingen av den molnbaserade lösning som skapats för projektet, i syfte att hantera data. Lösningen kopplar samman IoT enheter via MQTT protokollet, dessa enheter samlar och skickar sedan data till molnet, där denna data samlas i en databas, för att senare presenteras via en webbapplikation. Datan skickas till denna via REST eller WebSocket APIer. Lösningen implementeras med Amazon Web Services som plattform och det även på denna plattform som lösningen körs.

The majority of climate-health impacts are the result of extreme climatic events. In the Amazon region, hydrological extremes have become more frequent in recent years. Evidence exists about how these hydrological extremes affect the forest itself, yet little information is available on the impacts on human health. Hospitalisations for respiratory diseases are the leading cause of hospitalisations, excluding pregnancy related causes, for both Brazil, and the Brazilian Amazon. It has been shown elsewhere that during drought events and periods of intense fires there are statistically significant associations with respiratory health. Despite the increase in hydrological extremes and high rates of deforestation and fires observed annually in the Legal Amazon, there are limited studies linking such events and respiratory health. The lack of explicit spatial understanding about these connections restrains the ability of policymakers to plan and implement regional mitigation and adaptation policies in order to cope with predicted effects of climate change in the Amazon, one of Brazil’s poorest regions. Thus, this thesis explores the impacts of three large hydrological extremes: the 2005, and 2010 droughts and the 2009 flood, on children’s respiratory health in the Legal Amazon. The research is two-fold; firstly to establish how the extremes and associated human disturbance impact respiratory health in the region. A Geographically Weighted Poisson Regression is adopted which allows for local spatial data analysis to identify any relationships between selected variables and children’s respiratory health throughout the Legal Amazon. The second part explores local communities’ knowledge of respiratory health and the links to the environment which has assisted in creating recommendations to cope with respiratory health and environmental problems in the Legal Amazon.

This thesis examines how early warning signals perform when tested on climate systems thought to exhibit future tipping point behaviour. A tipping point in a dynamical system is a large and sudden change to the state of the system, usually caused by changes in external forcing. This is due to the state the system occupies becoming unstable, causing the system to settle to a new stable state. In many cases, there is a degree of irreversibility once the tipping point has been passed, preventing the system from reverting back to its original state without a large reversal in forcing. Passing tipping points in climate systems, such as the Amazon rainforest or the Atlantic Meridional Overturning Circulation, is particularly dangerous as the effects of this will be globally felt. Fortunately there is potential for early warning signals, designed to warn that the system is approaching a tipping point. Generally, these early warning signals are based on analysis of the time series of the system, such as searching for ‘critical slowing down’, usually estimated by an increasing lag-1 autocorrelation (AR(1)). The idea here is that as a system’s state becomes less stable, it will start to react more sluggishly to short term perturbations. While early warning signals have been tested extensively in simple models and on palaeoclimate data, there has been very little research into how these behave in complex models and observed data. Here, early warning signals are tested on climate systems that show tipping point behaviour in general circulation models. Furthermore, it examines why early warning signals might fail in certain cases and provides prospect for more ‘system specific indicators’ based on properties of individual tipping elements. The thesis also examines how slowing down in a system might affect ecosystems that are being driven by it.

<p>The contribution of surface evapotranspiration (ET) to moist convection, cloudiness and precipitation along the eastern flanks of the Andes (EADS) was investigated using the Weather Research and Forecasting (ARW-WRF3.4.1) model with nested simulations of selected weather conditions down to 1.2 km grid spacing. To isolate the role of surface ET, numerical experiments were conducted using a quasi-idealized approach whereby at every time step the surface sensible heat effects are exactly the same as in the reference simulations, whereas the surface latent heat fluxes are prevented from entering the atmosphere. </p><p>Energy balance analysis indicates that local surface ET along the EADS influences moist convection primarily through its impact on conditional instability, because it acts as an important source of moist entropy in this region. The energy available for convection decreases by up to ~60% when the ET contribution is withdrawn. In contrast, when convective motion is not thermally driven, or under conditionally stable conditions, latent heating from the land surface becomes secondary. At the scale of the Andes proper, removal of surface ET weakens upslope flows by increasing static stability of the lower troposphere, as the vertical gradient of water vapor mixing ratio tends to be less negative. Consequently, moisture convergence is reduced over the EADS. In the absence of local surface ET, this process operates in concert with damped convective energy, suppressing cloudiness, and decreasing daily precipitation by up to ~50% in the simulations presented here.</p><p>When the surface ET is eliminated over the Amazon lowlands (AMZL), the results show that, without surface ET, daily precipitation within the AMZL drops by up to ~75%, but nearly doubles over the surrounded mountainous regions. This dramatic influence is attributed to a dipole structure of convergence-divergence anomalies over the AMZL, primarily due to the considerable cooling of the troposphere associated with suppressed convection. Further examination of moist static energy evolution indicates that the net decrease in CAPE (Convective Available Potential Energy) over the AMZL is due to the removal of surface ET that is only partially compensated by related regional circulation changes. Because of the concave shape of the Andean mountain range, the enhanced low-level divergence promotes air mass accumulation to the east of the central EADS. This perturbation becomes sufficiently strong around nightfall and produces significant eastward low-level pressure gradient force, rendering wind currents more away from the Andes. Moisture convergence and convection over the EADS vary accordingly, strengthened in the day but attenuated at night. Nocturnal convective motion, however, is more widespread. Analytical solutions of simplified diagnostic equations of convective fraction suggest that reduction of lower troposphere evaporation is the driving mechanism. Additional exploratory experiments mimicking various levels of thinning and densification of AMZL forests via changes in surface ET magnitude demonstrate that the connection between the AMZL ET and EADS precipitation is robust.</p> / Dissertation

The western Amazon is known to be one of the most biologically diverse regions in the world, yet information about the spatial distribution of that biodiversity and the processes governing its distribution remains scarce. An improved understanding of those biogeographic patterns and processes can inform conservation and development planning in areas where anthropogenic landscape change is ongoing. Spatial components of biodiversity are known to be influenced by historical and present-day physical and human geographic processes. There is evidence that major Amazonian rivers form the boundaries of biological regions, at least for birds. Other factors that may influence bird species composition include the dispersal limitations of individual species, forest plant species composition and structure, topography, forest fragmentation, and hunting. Sites where bird species composition was measured in this study represented mature, upland forest on both sides of the Amazon River, and a range of non-flooded forest types, as indicated by soil and plant surveys. Bird species compositional variation was closely correlated with variation in plant species composition, human disturbance associated with forest fragmentation, and position north or south of the Amazon River. The strongest differences were between opposite sides of the river, even though local environments, including plant composition, were not different on the two sides. This strongly suggests that historical biogeographic factors, rather than present-day environmental gradients, are responsible for bioregional boundaries at Amazonian rivers. The difference between plant and bird distributions at this scale underscores the pressing need to re-evaluate general notions of bioregional complexity and pattern in the Amazon basin. Locally, the influence of habitat fragmentation on animal communities, including reduced species richness, was confirmed. The influence of local floristic variation is of particular importance due to its ubiquity across western Amazonia. Thus, understanding the distributions of soils and vegetation is critical for explaining Amazonian animal diversity. The use of these factors to model bird community heterogeneity contradicts assumptions that the processes shaping Amazonian animal community diversity are too complex to measure efficiently, and their use contributes a new understanding of the dimensions of that diversity. / text

Geography and the Environment / Studies of the Amazon drainage network have primarily focused on the Western Basin and the Amazon Cone, but they have neglected the integration between these areas. Data presents a time gap in the Amazon’s development and the forces responsible for the organization of the drainage network are poorly understood. A key element towards gaining an improved awareness of the Amazon is the Eastern Amazon River Valley. The focus of this study is an 80,000 km² portion of this area. An integrated method is adopted that combines terrain information derived from a digital elevation model with geologic data. The interpretation of DEM data is unique to this study. Seven distinct surfaces were identified, along with numerous erosional environments. This observation supports a geomorphologic record of numerous erosional events starting in the Miocene. This finding is significant as it rejects previous models for staircase-like terraces for the Amazon, and establishes a timeline for the development of geomorphologic landforms in the study area. In addition, neotectonics events provide an alternative explanation to the generation of topography in the study area. It was concluded that geomorphology in the study area is the result of physical and chemical weathering, and modified by neotectonics. These findings provide alternative means for Amazon landscape evolution. / text

Current processes of deforestation and development in the Amazon Basin continue historical trends that have devastated indigenous populations and drastically reduced their land rights. While protection of the Amazon ecosystem has become a worldwide concern, many indigenous and folk groups employ forest management strategies that utilize natural resources without causing permanent degradation. This paper considers historical, political and socioeconomic circumstances that threaten the survival of indigenous groups and their sustainable forms of forest use. The paper argues that discrepant cultural models and attitudes contribute to the differences in land use between traditional Amazon residents and newcomers. The problems and possibilities entailed by efforts to protect traditional land rights are also discussed.

This dissertation is aimed at analyzing how ethnic hierarchies question the environmentalization of indigeneity, which is the foundation of the Colombian state's multicultural policy. In particular, the dissertation develops a comparative ethnographic approach to the way in which the "multicultural turn" of 1991 impacted three indigenous communities located at San José del Guaviare, a colonization frontier in the Colombian Amazon: the Nükak, the Jiw and the Tucano. Against the assumption of multicultural policy that indigenous communities form a vast mass of people radically diferent from mainstream (even portrayed as anti-modern), in San José there is an unequal distribution of the Nükak, Jiw and Tucano in different positions inside local ethnic hierarchies. For some, Nükak incarnate what Hale (2004) label as a "good ethnicity", that serves to promote Guaviare as an eco-touristic destination, the Jiw are a "bad ethnicity" that annoys White people in San José, while the Tucano are portrayed as "civilized Indians". Thus, the dissertation states how these ethnic hierarchies contradict some of the core assumptions of multicultural policies that are based on an essentialized understanding of indigenous peoples as "ecologically noble savages." The dissertation argues that the analysis of contemporary experiences on indigeneity in an Amazonian context such as San José, could be better understood if it observes a set of processes and actors including: the historical transformation of senses on otherness, the production of forests as a field of domain under state regulations, the economic crossroads affecting indigenous peoples on their "resguardos" (indigenous lands) and the intervention of state laws, NGOs, indigenous political organizations, settlers, foreign governments and state officials. The analysis of such a variety of processes and actors shaping contemporary experiences on indigeneity in the Colombian Amazon follows the environmentality approach (Agrawal, 2005). From that perspective, I discuss the following ideas: a) indigenous resguardos were designed as governmentalized localities in multicultural policy to regulate and control how indigenous peoples manage natural resources; b) those communities portrayed as followers of the ecological nobility script act as regulatory communities; c) the technologies for governing the ecological realm do not necessarily assure the formation of environmental subjectivities.