La technologie de réseau cellulaire de 5ème génération (5G) sera le réseau supportant l'Internet des objets (IoT). Elle a introduit une fonctionnalité majeure, communications appareil-à-appareil (D2D), que permettent communications sans fil à consommation d'énergie restreinte en interagissant à proximité et à puissance d'émission plus faible. La coopération entre appareils suscit donc un intérêt considérable pour l'énergie, et peut être utilisé en conjonction avec la récupération d'énergie pour prolonger la durée de vie des appareils. Les programmes de coopération renforcent la mise en réseau d'un appareil à l'autre, ce qui accroît la nécessité d'exécuter des mécanismes de sécurité pour assurer la protection des données et les relations de confiance entre les nœuds du réseau.Ces mécanismes sont fondamentaux pour la protection contre les attaques malveillantes mais elles représentent aussi une importante consommation d'énergie, souvent négligée en raison de l'importance de la protection des données. L'établissement d'un canal securisé peut être coûteux en termes d'utilisation du CPU, la mémoire et la consommation d'énergie, surtout si les appareils sont limités en ressources. La confidentialité et l’intégrité des données ont un faible coût énergétique, mais sont utilisées en permanence. Il est donc nécessaire de quantifier la consommation d'énergie engendrée par la sécurité d'un appareil. Un modèle énergétique basé sur la sécurité est proposé pour répondre à cet objectif.Dans les réseaux composés d'équipements d'utilisateurs (UE), la mobilité est une caractéristique clé. Elle peut agir sur la connexion à proximité d'objets IoT, étendant la couverture 5G vers l'IoT via les UEs. Une solution d'authentification légère est présentée qui permet par l'authentification directe et des communications UE-IoT, d'étendre la couverture et réaliser des économies d'énergie potentielles importantes. Cette approche peut être particulièrement utile en cas de catastrophe où l'infrastructure réseau peut ne pas être disponible.La condentialité et l'authentification des données sont une source de consommation d'énergie importante. Les appareils équipés avec équipements de collecte d'énergie (EH) peuvent avoir un excédent ou un déficit d'énergie. La sécurité appliquée peut donc être ajustée en fonction de l'énergie disponible d'un appareil, en introduisant l'établissement de canal sécurisé qui tient compte de la consommation d'énergie. Après avoir étudié en profondeur les normes 5G, il a été constaté que les réseaux d'UE D2D utilisant ce type de norme dépenseraient une quantité importante d'énergie et seraient généralement moins sûr. Un mécanisme léger de recléage est donc proposé pour réduire les coûts liés cette adaptation. Pour compléter le concept de canal sécurisé prenant en compte l'énergie et le mécanisme de recléage, une méthode de bootstrapping des paramètres de sécurité est également présentée. Le méthode désigne le cœur du réseau (CN) comme responsable de la politique de sécurité, rend l'ensemble du réseau plus sûr et aide à prévenir les pannes de communication. L'adaptation susvisé requiert l'étude du compromis entre l’énergie et sécurité. À cette fin, un processus décisionnel de Markov (MDP) modélisant un canal de communication est présenté lorsqu'un agent choisit les éléments de sécurité à appliquer aux paquets transmis. Ce problème d'optimisation du contrôle stochastique est résolu par plusieurs algorithmes de programmation dynamique et d’apprentissage par le renforcement (RL). Les résultats montrent que l'adaptation susvisé peut prolonger de manière significative la durée de vie de l'équipement et de la batterie, et améliore la fiabilité des données tout en offrant des fonctions de sécurité. Une étude comparative est présentée pour les différents algorithmes RL. Puis une approche d'apprentissage Q-profond (DQL) est proposé que améliore la vitesse d'apprentissage de l'agent et la fiabilité des données. / The 5th Generation Cellular Network Technology (5G) will be the network supporting the Internet of Things (IoT) and it introduced a major feature, Device-to-Device (D2D) communications. D2D allows energy-constrained wireless devices to save energy by interacting in proximity at a lower transmission power. Cooperation and device-assisted networking therefore raise signicant interest with respect to energy saving, and can be used in conjunction with energy harvesting to prolong the lifetime of battery-powered devices. However, cooperation schemes increase networking between devices, increasing the need for security mechanisms to be executed to assure data protection and trust relations between network nodes. This leads to the use of cryptographic primitives and security mechanisms with a much higher frequency.Security mechanisms are fundamental for protection against malicious actions but they also represent an important source of energy consumption, often neglected due to the importance of data protection. Authentication procedures for secure channel establishment can be computationally and energetically expensive, especially if the devices are resource constrained. Security features such as condentiality and data authentication have a low energetic cost but are used constantly in a device engaged in data exchanges. It is therefore necessary to properly quantify the energy consumption due to security in a device. A security based energy model is proposed to achieve this goal.In User Equipment (UE) D2D networks, mobility is a key characteristic. It can be explored for connecting directly in proximity with IoT objects. A lightweight authentication solution is presented that allows direct UE-IoT communications, extending coverage and potentially saving signicant energy amounts. This approach can be particularly useful in Public Protection and Disaster Relief (PPDR) scenarios where the network infrastructure may not be available.Security features such as condentiality or data authentication are a significant source of consumption. Devices equipped with Energy Harvesting (EH) hardware can have a surplus or a deficit of energy. The applied security can therefore be adjusted to the available energy of a device, introducing an energy aware secure channel. After in depth analysis of 5G standards, it was found that D2D UE networks using this type of channel would spend a signicant amount of energy and be generally less secure. A lightweight rekeying mechanism is therefore proposed to reduce the security overhead of adapting security to energy. To complete the proposed rekeying mechanism, a security parameter bootstrapping method is also presented. The method denes the Core Network (CN) as the security policy maker, makes the overall network more secure and helps preventing communication outages.Adapting security features to energy levels raises the need for the study of the energy/security tradeoff. To this goal, an Markov Decision Process (MDP) modeling a communication channel is presented where an agent chooses the security features to apply to transmitted packets. This stochastic control optimization problem is solved via several dynamic programming and Reinforcement Learning (RL) algorithms. Results show that adapting security features to the available energy can signicantly prolong battery lifetime, improve data reliability while still providing security features. A comparative study is also presented for the different RL learning algorithms. Then a Deep Q-Learning (DQL) approach is presented and tested to improve the learning speed of the agent. Results confirm the faster learning speed. The approach is then tested under difficult EH hardware stability. Results show robust learning properties and excellent security decision making from the agent with a direct impact on data reliability. Finally, a memory footprint comparison is made to demonstrate the feasibility of the presented system even on resource constrained devices.

The subject of the diploma thesis are procedures in EPS network related to the operation of mobile terminals. In the introduction of my work I described the EPS system, including the quality of services and bearers. The following section focuses on control procedures in the EPS system such as attach network procedure, detach procedure, service request, handover or tracking area procedure. Further, three technologies for Internet of Things were described, namely EC-GSM-IoT, LTE-M and NB-IoT. All the procedures mentioned above were also captured in the experimental VUT EPS network and analyzed by TravelHawk using software M5 and Wireshark. Theoretical and practical findings were subsequently used out for setting of laboratory task for subject MKPM.

La course à la miniaturisation des appareils informatiques est en train de transformer notre relation avec ces derniers, ainsi que leurs rôles dans notre société. Le nombre d’ordinateurs miniatures contrôlés à distance augmente considérablement et ces objets connectés - comme ils sont communément appelés - sont de plus en plus sollicités pour effectuer des tâches à la place de l’Homme. La tendance actuelle consiste à créer une place dans Internet pour ces objets connectés, autrement dit, à construire des protocoles adaptés à leurs ressources limitées. Cette tendance est connue comme l’Internet desObjets - ou l’acronyme anglais IoT - qui est différent des protocoles destinés à une utilisation exclusivement par des humains dit Internet des Personnes ou IoP en anglais. Avec l’adoption de cette séparation conceptuelle, comment est-ce qu’une personne échangerait ses informations avec des objets sans sacrifier la sécurité ? Pour aider à réduire cet écart, on a besoin d’un intermédiaire et la mise en réseau de ces intermédiaires amène à construire le concept d’Internet des Services ou IoS en anglais. Les personnes et les objets sont connectés à travers les services. Le réseau dans son ensemble, incluant les personnes, les objets et les services est donc l’Internet des Personnes, des Objets et des Services. Notre travail se situe à l’intersection de ces trois domaines et notre contribution est double. Premièrement, nous assurons que la liaison entre l’identité d’une personne et de ses objets ne se fasse pas au détriment des propriétés de sécurité telles que l'Intégrité, l'Anonymat et la confidentialité. Et deuxièmement, nous abordons la gestion de la confidentialité des données avec les objets dits connectés. Dans la quête d’une meilleure intégration des objets connectés à Internet, nous avons contribué à la définition de protocoles autant sur la couche applicative que sur la couche réseau du modèle OSI, avec pour préoccupations principales les contraintes de l’IoT et la sécurité. / The constant efforts of miniaturization of computing machines is transforming our relationships with machines and their role in society. The number of tiny computers remotely controlled is skyrocketing and those connected things are now more and more asked to do things on human behalf. The trend consists in making room for these specific machines into the Internet, in other words, building communication protocols adapted to their limited resources. This trend is commonly known as the Internet of Things (IoT) which consist of appliances and mechanisms different from those meant to be used exclusively by humans, the Internet of Persons (IoP). This conceptual separation being adopted, how would a Person exchange information with Things ?Sorts of brokers can help bridging that gap. The networking of those brokers led to the concept of Internetof Services (IoS). Persons and Things are connected through Services. This global networking is called the Internet of Persons Things and Services (IoPTS). Our work is on the edge of these 3 Internet areas and our contributions are two fold. In the first hand, we tackle the secure biding of devices’ and persons’ identities while preserving the Integrity, Anonymity and Confidentiality security properties. On the other hand, we address the problem of the secrecy of data on constrained Internet-connected devices. Other mechanisms must be created in order to seamlessly bind these conceptual areas of IoP, IoT andIoS. In this quest for a better integration of Internet connected-devices into the Internet of Persons, our work contributes to the definition of protocols on application and network layers, with IoT concerns and security at heart.

IoT devices provide immense contributions in fields of education, communication, business, science, industrial zones, permeating various aspects of everyday life. Despite these benefits, their diversity, heterogeneity, and rapid development can introduce significant challenges, especially when the secure design has not been incorporated into their software lifecycle. Consequently, they can be targeted by malicious attackers, resulting in important security threats that need to be addressed. The main goal of this research is to explore the benefits of securing IoT devices after deployment, by examining, analyzing, and comparing default vs secure IoT device configurations. This will allow us to make assumptions regarding the differences in behavior, patterns, and motives of the attackers, as well as to measure the performance output between the two environments. To achieve our goal, we make extensive use of honeypot systems simulating the two different environments and collect log data to conclude meaningful information. As a secondary goal, we also explore any potential performance degradation for default vs secure configurations. The produced results suggest that there is a major difference both in terms of how the attackers approach a hardened IoT device/service, and how a device is affected in terms of operability and performance.

Internet of Things (IoT) devices often described as Smart products for consumers consists ofphysical things that inherit an Internet connection and therefore enable physical things to talkbetween each other and with people. IoT is a growing market with products existing in e.g.,consumers' homes, healthcare, or industries. These physical things have sensors that cangather information about users which later on can be used to adapt the behaviors of IoTdevices or create profiles of users. As with any device that is connected to the Internet, IoTdevices can fall victim to attacks from outside parties that try to steal private information orobserve users of the devices. With the growing market, it is of importance to understand what people know about IoTdevices and that a Smart home does not come with comfortability without its rough edges.This Bachelor thesis answers what people know about IoT, how they handle security issues,and the sharing of people’s personal information.The data was gathered through a survey that had 133 participants, the survey was shared inFacebook groups and spread through the snowball effect. After a participant was done withthe survey, they were urged to continue to share the survey with people they knew. The result shows that the majority of people have not heard the term IoT before and that it isnot common to take steps in protecting private information when using IoT devices. IoT is aterm that is unfamiliar to many, the result shows that the term Smart devices is morecommonly used than IoT devices among people.

Avec l'évolution exponentielle de l'Internet des Objets (IoT), assurer la sécurité des réseaux est devenue un grand défi pour les administrateurs réseaux. La sécurité des réseaux est basée sur de multiples équipements indépendants tels que Firewall, IDS/IPS, NAC dont le rôle principal est de contrôler les informations échangées entre le réseau de l'entreprise et l'extérieur. Or, l'administration de ces équipements peut s'avérer très complexe et fastidieuse si elle est réalisée manuellement, équipement après équipement. L'introduction du concept de Software Defined Networking (SDN) depuis ces dernières années, et du protocole OpenFlow, offre beaucoup d'opportunités pour l'amélioration de la sécurité des réseaux en proposant une administration centralisée et programmable.Dans le cadre de cette thèse, nous avons proposé une nouvelle approche de sécurisation des échanges dans un réseau en fonction des événements détectés et de manière automatisée. Cette solution basée sur l'approche SDN couplé avec un système de détection d'intrusion permet d’analyser, de détecter et de supprimer des menaces de sécurité dans un réseau et de manière automatisée. En implémentant cette solution, nous contribuons à faire évoluer la manière de sécuriser les échanges dans un réseau avec du SDN couplé avec un IDS à travers la mise en place d'une architecture réelle de cas d'usage. Ainsi, la gestion de la sécurité du réseau devient simplifiée, dynamique et évolutive. / With the exponential evolution of the Internet of Things (IoT), ensure the network security has become a big challenge for networkadministrators. Traditionally, the network security is based on multiple independent devices such as firewall, IDS/IPS, NAC where the main role is to monitor the information exchanged between the inside and the outside perimeters of the enterprises networks. However, the administration of these network devices can be complex and tedious with an independent manual configuration. Recently, with the introduction of the Software Defined Networking concept (SDN) and the OpenFlow protocol offers many opportunities by providing a centralized and programmable network administration.As part of this research work, we proposed a new approach to secure the network traffic flows exchanges based on a method of events detection, in an automated manner. This solution is based on the SDN approach coupled to an intrusion detection system which allows analyze, detect and remove security threats. With the implementation, we contribute to change the paradigm of secure the network traffic flows exchanges using the SDN principle, coupled with an IDS in a real use case architecture. In this way, the management of network security becomes simplified, dynamic and scalable.

With 5G around the corner and an overall increase in a faster and more stable internet connection, the future of Internet of Things (IoT) looks bright. There is a steady increase in the development of IoT devices, such as the smartwatch, and a high increase in usage of IoT, both by organizations and private citizens. Organizational managing of a smartwatch falls under the “Bring your own device” (BYOD) policy which allows employees to do work on their private devices. It appears to be a lack of knowledge in organizations on how to manage IoT devices both regarding policies and technical IT security. There has been an increase in malware attacks against IoT devices, and compromised smartwatches could be used to gain unauthorized access to organizations’ networks. The smartwatch is a common and powerful IoT device and will be used as an example in this thesis which purpose is to examine how organizations’ perceive and manage IoT devices, focusing on the smartwatch in order to gain insight regarding whether IoT devices such as the smartwatch is an area of concern within organizations. To understand the smartwatch, understanding IoT first will be important. The literature review delves into both IoT and smartwatch functionality and security. It looks at the BYOD policy and technical IT security solutions regarding the smartwatch. The review pointed to there being IT security issues with smartwatches and that implementing a BYOD policy increases productivity but increases the risk of malware attacks from and against the allowed devices. To fulfill the thesis purpose, qualitative interviews with high ranking IT security personnel at tech organizations were performed, thematized, and analyzed. The most prominent results are discussed; if the smartwatch is a threat and possible technical solutions for prevention, the organizations customer businesses IT security level, and BYOD policy. The results from the thesis showed that the organizations had a high awareness of the smartwatch and the IT security risks brought with it. They all had BYOD policies to restrict/limit access for the smartwatch’s access to their internal networks and a set of technical solutions to prevent breaches in the IT infrastructure and to detect if there had been a breach. The informants claimed that their organizations’ awareness regarding the smartwatch and the concerning IT security was higher than many of their customer businesses, which makes for an interesting subject for future research. How can these organizations reach the same level of awareness? / Med 5G runt hörnet och en generell ökning av både snabbare och stabilare internet så ser framtiden för Internet of Things (IoT) ljus ut. Det pågår en stadig ökning i utvecklingen av IoT-enheter såsom smartklockan, samtidigt som en användandet av IoT ökar både på företag och hos privatpersoner. En verksamhets hantering av smartklockan hamnar under policyn ”Bring your own device” (BYOD) vilket tillåter anställda att använda sina privata enheter i jobbrelaterat syfte. Det verkar finnas en kunskapsbrist hos verksamheter avseende hur man hanterar IoT-enheter, både gällande policy och teknisk IT-säkerhet. Det har skett en ökning av malware attacker (skadlig kod) mot IoT-enheter och en kompromissad smartklocka kan potentiellt användas för att få otillbörlig åtkomst till en verksamhets nätverk. Smartklockan är en vanlig och kraftfull IoT-enhet och kommer att användas som exempel i den här uppsatsen. Syftet med uppsatsen är att undersöka hur verksamheter uppfattar och hanterar IoT-enheter med fokus på smartklockan, för att ta reda på om IoT-enheter såsom smartklockan är ett område som verksamheter arbetar med. För att förstå smartklockan så är det viktigt att först förstå IoT. I litteraturstudien redogörs både IoT och smartklockors funktionalitet samt säkerhetsaspekter. Vidare beskrivs även BYOD policy och tekniska IT-säkerhetslösningar gällande smartklockan. Litteraturstudien pekade på att det existerar IT-säkerhetsproblem med smartklockan och att implementera en BYOD policy kan öka verksamhetens produktivitet men även öka riskerna med malware attacker, både mot och från de tillåtna enheterna. För att uppfylla uppsatsens syfte utfördes kvalitativa intervjuer med högt uppsatt IT-säkerhetspersonal på IT-orienterade verksamheter, som sedan tematiserades och analyserades. De mest relevanta resultaten diskuteras, avseende smartklockan som ett hot och de relaterade tekniska lösningarna, verksamheternas kundföretags IT-säkerhetsnivå och BYOD policyn. De empiriska resultaten från uppsatsen visade att verksamheterna som intervjuades hade en hög medvetenhet relaterat till smartklockan och de IT-säkerhetsproblem som den kan medföra. Alla verksamheterna hade en BYOD policy för att begränsa/förbjuda smartklockans åtkomst till deras interna nätverk, samt ett par tekniska lösningar för att förebygga intrång i deras IT-infrastruktur och för att upptäcka om ett intrång redan skett. Informanterna påstod att deras verksamheters medvetenhet kring smartklockan och den relaterade IT-säkerheten var högre kontra flera av deras kundföretags, vilket är ett relevant ämne för framtida forskning. Hur kan dessa verksamheter nå upp till samma nivå av medvetenhet?

This thesis project explores two well-known real-time web technologies: WebSocket and WebRTC. It explores the use of a mobile phone as a gateway to connect wireless devices with short range of radio links to the Internet in order to foster an Internet of Things (IoT). This thesis project aims to solve the problem of how to collect real-time data from an IoT device, using the Earl toolkit. With this thesis project an Earl device is able to send real-time data to Internet connected devices and to other Earl devices via a mobile phone acting as a gateway. This thesis project facilitates the use of Earl in design projects for IoT devices. IoT enables communication with many different kinds of “things” such as cars, fridges, refrigerators, light bulbs, etc. The benefits of IoT range from financial savings due to saving energy to monitoring the heart activity of a patient with heart problems. There are many approaches to connect devices in order to create an IoT. One of these approaches is to use a mobile phone as a gateway, i.e., to act as a router, between IoT and the Internet. The WebSocket protocol provides efficient communication sessions between web servers and clients by reducing communication overhead. The WebRTC project aims to provide standards for real-time communications technology. WebRTC is important because it is the first real-time communications standard which is being built into browsers. This thesis evaluates the benefits which these two protocols offer when using a mobile phone as a gateway between an IoT and Internet. This thesis project implemented several test beds, collected data concerning the scalability of the protocols and the latency of traffic passing through the gateway, and presents a numerical analysis of the measurement results. Moreover, an LED module was built as a peripheral for an Earl device. The conclusion of the thesis is that WebSocket and WebRTC can be utilized to connect IoT devices to Internet. / I detta examensarbete utforskas två välkända realtidsteknologier på internet: WebSocket och WebRTC. Det utforskar användandet av en mobiltelefon som gateway för att ansluta trådlösa enheter - med kort räckvidd - till Internet för att skapa ett Internet of Things (IoT). Det här examensarbetet försöker med hjälp av verktyget Earl lösa problemet med hur insamlandet av realtidsdata från en IoT-enhet skall genomföras. I det här examensprojektet kan en Earl-enhet skicka data i realtid till enheter med Internetanslutning, samt till andra Earl-enheter, med hjälp av en mobiltelefon som gateway. Detta projektarbete förenklar användandet av Earl i design-projekt ör IoT-enheter. IoT tillåter kommunikation mellan olika sorters enheter, så som bilar, kyl- och frysskåp, glödlampor etc. Fördelarna med IoT kan vara allt från ekonomiska - tack vare minskad energiförbrukning - till medicinska i form av övervakning av puls hos patienter med hjärtproblem. Det finns många olika tillvägagångssätt för att sammankoppla enheter till ett IoT. Ett av dessa är att använda en mobiltelefon som en gateway, dvs en router mellan IoT och internet. WebSocket-protokollet erbjuder effektiv kommunikation mellan web-servrar och klienter tack vare minskad överflödig dataöverföring. WebRTC-projektet vill erbjuda standarder för realtidskommunikation. WebRTC är viktigt då det är den första sådana standarden som inkluderas i webläsare. Det här examensarbetet utvärderar fördelarna dessa två protokoll erbjuder i det fallet då en mobiltelefon används som gateway mellan ett IoT och Internet. I det här examensprojektet implementerades ett flertal testmiljöer, protokollens skalbarhet och fördröjningen av trafiken genom mobiltelefonen (gateway) undersöktes. Detta presenteras i en numerisk analys av mätresultaten. Dessutom byggdes en LED-modul som tillbehör till en Earl-enhet. Slutsatsen av examensarbetet är att WebSocket och WebRTC kan användas till att ansluta IoT-enheter till Internet.

IoT-enheter blir allt populärare i takt med att de blir kraftfullare och skalbara. Därför var det viktigt att undersöka prestandan hos IoT-enheter när det kommer till maskininlärning. Huvudsyftet med detta arbete är att implementera två maskininlärningsmodeller på en bärbar dator och en Raspberry Pi 4 för att bestämma vilka maskininlärningsuppgifter som kan utföras på dessa enheter genom att sätta upp scenarier där vi kan testa låg-, medel- och högkrävande maskininlärning uppgifter på båda enheterna, som också möjliggör noggrann mätning. En bärbar dator användes som referenspunkt för att se om Raspberry Pi 4 kan prestera bra jämfört med en modern bärbar dator när de utför maskininlärningsuppgifter. Tensorflow valdes att användas för att implementera de två maskininlärningsmodellerna och för att mäta processen. Noggrannheten och träningstiden mättes för att bestämma prestandan för modellerna på de två enheterna. Tre datauppsättningar valdes ut för att användas för att träna och testa modellerna på de två enheterna, dessa datauppsättningar innehöll bilder, den första datauppsättningen bestod av mycket små bilder, och den andra bestod av lite större bilder, den sista datauppsättningen bestod av ännu större bilder, detta gjordes för att testa tre olika svårighetsgrader för modellerna på de två olika enheterna. Efter att träningen och utvärdering av modellerna slutförts på båda enheterna med hjälp av de tre datauppsättningarna analyserades de resulterande mätningarna och diskuterades sedan för att nå en slutsats. Slutsatsen från detta arbete var att endast lågnivåmaskininlärningsuppgifter är ett gångbart alternativ nu på grund av den extrema tid som krävs för att träna modellen för bildklassificering, men om tiden inte är en viktig faktor, skulle Raspberry Pi 4 efter en lång tid fortfarande uppnå samma noggrannhet som den bärbara datorn gjorde. / IoT devices are becoming increasingly popular as they are becoming more powerful and scalable. Therefor it was important to examine the potential of IoT devices when it comes to Machine Learning. The main objective of this work is to implement two machine learning models on a laptop and a Raspberry Pi 4 to determine what machine learning tasks that can be performed on these devices by setting up scenarios where we can test low, mid and high demanding machine learning tasks on both devices that also allows for accurate measurement being taken. A laptop was used a reference point to see if the Raspberry Pi 4 can perform well relative to a modern-day laptop when performing machine learning tasks. Tensorflow was chosen to be used to implement the two machine learning models and to measure the process. The accuracy and training time were measured to determine the performance of the models on the two devices. Three datasets were chosen to be used for training and testing the models on the two devices, these datasets contained images, the first dataset consisted of very tiny images, and the second one consisted of slightly larger images, the last dataset consisted of very large images, this was done to test three different levels of difficulty for the models. After training and evaluation of the models were completed on both devices using the three datasets, the resulting measurements were analyzed and then discussed to reach the conclusion. The conclusion from this work was that only low-tier machine learning tasks are a viable option now because of the extreme amount of time required to train the model for image classification, however if time isn’t an important factor, the Raspberry Pi 4 would after a long time still reach the same accuracy as the laptop did.

Internet of Things (IoT) är ett koncept som fått mycket uppmärksamhet från forskare och utövare av IoT världen över. IoT kan beskrivas som ett nätverk där vardagliga föremål kommunicerar med varandra eller med människor. IoT kommer att bygga ut internet genom att integrera föremål för kommunicering via inbäddade system, detta leder till ett distribuerat nätverk av enheter som kommunicerar med människor och andra IoT-enheter. Härnösand Miljö & Energi AB (HEMAB), är en av ägarna till ServaNet och är ett företag som vill utöka sitt testområde för IoT-sensorer som använder LoRa-teknik för att kommunicera. HEMAB är en av ägarna till ServaNet. Studien har konstruerat metoder för mätning av höjdskillnad från data av en barometer och en accelerometer. Resultatet jämför de två sensorerna med avseende på hur tillförlitligt de kan mäta höjdskillnad över tid. Accelerometern ger mer tillförlitliga resultat som ligger närmare verkligheten, på lägre höjder, med en kortare mätperiod i jämförelse med barometern. Däremot, visar barometern mer tillförlitliga värden på högre höjder, och mer tillförlitligt värde än accelerometern på den högsta testade höjden. Anledningen till att accelerometern ger mindre tillförlitliga resultat på högre höjder med längre mätperiod, beror på det sättet dubbelintegrering tar med sig felande och avstickande värden som blir större och större ju längre tid som integreras. Barometern ger osäkra värden på lägre höjd på grund av att sensorn ger spridda värden redan då den ligger stillastående, detta jämnas ut på högre höjder. LoRaWAN är en lämplig teknik för att kommunicera våghöjd, förutsatt att beräkningar görs på prototypen och endast våghöjden skickas över LoRa, inte sensordata. / Internet of Things (IoT), is a concept that has received a lot of attention from researchers and practitioners of IoT worldwide. IoT can be described as a network where everyday objects communicate with each other or with people. IoT will expand the internet by integrating objects for communication via embedded systems, this leads to a distributed network of devices that communicate with people and other IoT devices. Härnösand Miljö & Energi AB (HEMAB), is one of the owners of ServaNet and is a company that wants to expand its test area for IoT sensors that use LoRa-technology to communicate. HEMAB is one of the owners of ServaNet, as ServaNet is the city network in Härnösand. This study has constructed a method for each sensor that can measure height difference. The result compares a barometer and an accelerometer with respect to how reliably they can measure height difference over time. The accelerometer gives more reliable results that are closer to the reality, at lower altitudes, with a shorter measurement period compared to the barometer. On the other hand, the barometer shows more reliable measurements at higher altitudes, and more accurate measurement than the accelerometer at the highest tested altitude. The reason why the accelerometer performs worse at higher altitudes with longer measurement periods, is due to the way double integration stacks up errors that grows with time, resulting in larger errors for longer periods of measurements. The barometer gives no reliable measurements at lower altitudes, due to the sensor giving scattered values even when the sensor is stationary, this evens out at higher altitudes. LoRaWAN is a suitable technology for communicating wave height, provided that calculations are made on the prototype and only the wave height is sent over LoRa, not sensor-data.  The outcome of the benchmarks measurement showed that increasing the number of nodes would not result in superior performance. It was noted that an optimal number of nodes was found to be between 1 and 20 for all the tests performed. The study showed that no load balancer could be considered a clear winner, instead, different configurations of load balancers performed varyingly well at different tests.