In large-scale graph processing, a fixpoint iterative algorithm is a set of operations where iterative computation is the core. The aim, in fact, is to perform repetitive operations refining a set of parameter values, until a fixed point is reached. To describe fixpoint iterative algorithms, template execution plans have been developed. In an iterative algorithm an execution plan is a set of dataflow operators describing the way in which parameters have to be processed in order to implement such algorithms. In the Bulk iterative execution plan all the parameters are recomputed for each iteration. Dependency plan calculates dependencies among vertices of a graph in order to iteratively update fewer parameters during each step. To do that it performs an extra pre-processing phase. This phase, however, is a demanding task especially in the first iterations where the amount of data is considerable. We describe two methods in order to address the preprocessing step of the Dependency plan. The first one exploits an optimizer which allows switching the plan during runtime, based on a cost model. We develop three cost models taking into account various features characterising the plan cost. The second method introduces optimizations that bypass the pre-processing phase. All the implementations are based on caching parameters values and so they are memory greedy. The experiments show that, while alternative implementation of Dependency plan does not give expected results in terms of per-iteration time, cost models are able to refine the existing basic cost model increasing accuracy. Furthermore, we demonstrate that switching plan during runtime is a successful strategy to decrease the whole execution time and improve performance.

Hadoop was developed as an open-source software framework that leveraged initially the MapReduce programming model and therefore was able to efficiently analyse and process large datasets. At the core of Hadoop is the Hadoop distributed file system or HDFS, which is used as the default storage across the cluster. Hadoop can also be used with other types of storage, with or without HDFS, such as Amazon S3, Windows Azure Storage Blobs, GlusterFS, Tachyon etc. This thesis focuses on Tachyon, a distributed file system that claims to enable reliable data sharing at memory speed across cluster computing frameworks. We benchmark and evaluate HDFS with and without Tachyon in regards to performance. To do so we used TestDFSIO as a benchmark to simulate different MapReduce workloads and an in-production Spark job from Spotify. Tachyon's different writetypes were also put to the test and evaluated. To see how cloud solutions compare, we perform the same evaluations of Tachyon over Google Cloud Storage.

Multipath TCP är en viktig  ändring förr TCP, där målet är att låsa upp fördelarna med multipath in ett nätverk genom att dela upp dataöverföring över subflöden, och samtidigt försöka bibehålla bakåtkompabilitet med nuvarande system och infrastruktur. Detta examensarbete är en omfattande säkerhetsutvärdering av Multipath TCP med en djupgående analys av dess främsta sårbarheter, speciellt sårbarheten ADD ADDR som påverkar version 0 av Multipath TCP. Efter en inledande fas bestående av en undersökning och testning av sårbarheten så bestod arbetet främst av utvecklingen av en rad fixar för sårbarheten ADD ADDR. Dessa fixar gjordes för Linuxkärnans implementation av Multipath TCP, i ett första steg mot version 1 av protokollet. Denna rapport innehåller en utvärdering baserad på experiment för varje steg i utvecklingsfasen, samt en slutgiltig prestandaanalys för den modifierade implementationen av Multipath TCP. I processen producerades dels ett förbättringsförslag till det officiella RFC-dokumentet som innehåller specifikationen för Multipath TCP, och även mindre kodändningar till närliggande open-source projekt, exempelvis packet-analyseraren Wireshark. Examensarbetet innehöll mycket interaktion med viktiga open-source grupper, framförallt gruppen för Linuxkärnan. Multipath TCP är ett fortlöpande projekt som är under ständig förändring, både kring protokollet och dess implementation. Detta arbete är ett steg i den övergripande distributionsprocessen, med fokus på säkerhet. / Multipath TCP is a major extension for TCP that aims at unlocking multipath benefits in the network by splitting data transferring across multiple subflows, keeping into consideration funda-mental compatibility goals in order to achieve a seamless deployment of the new protocol on current systems and infrastructures. This thesis work is an extensive security evaluation of Multipath TCP, with an in-depth analysis of its major vulnerabilities: in particular, the ADD ADDR vulnerability affecting Multipath TCP version 0 is addressed in this document. After a preliminary part that involves testing and studying the vulnerability itself, the core value of this thesis work concerns the development of a series of patches for the Linux Kernel implementation of Multipath TCP aimed at fixing the ADD ADDR vulnerability and marking the first step towards the new version of the protocol, Multipath TCP version 1. The paper contains an experimental evaluation of the various steps of the development phase, including a final performance analysis for the modified implemen-tation of Multipath TCP. In the process, contributions have been produced for the official RFC documentation containing the specifications for Multipath TCP, together with other minor code contributions for related open-source projects, like the famous packet analyzer Wireshark. The entire thesis work required strong interaction with important open-source communities, especially the Linux Kernel one. Multipath TCP is an ongoing project subjected to constant changes regard-ing the protocol and its implementations: this work is just a single step of the overall deployment process, with a main focus on security.

<p>QC 20150109</p>

With the evolution of technology the system becomes large and complex. The probability of faults to occur in the system also increases with the size of the system. It is almost impossible to make the system fault free thus Fault Management System (FMS) is required to make the system fault tolerant which keeps the system operational even in the presence of transient fault and recover the system. The FMS is quite complex and sophisticated for large systems and it requires smart technique to verify FMS. Fault injection technique is emerging technique to verify the system robustness and its fault tolerance. There are various types of fault injection techniques and each type has its own characteristics. In this thesis, Ericsson’s Digital Unit of Radio Base Station (RBS) is used as target system. This research work involves the study of RBS hardware, software and also various fault injection techniques. Previously each module of FMS is tested in lab separately. With this new approach the complete FMS is tested in lab while system being operational. Based on this investigation, Hybrid FI technique is proposed and implemented to verify its feasibility. This technique requires change in target’s hardware and fault injector software. Experiment is performed on RBS connected to client via network. The result of an experiment is compiled in the end of this report. This research work concludes that the FMS can be verified with FI technique. This technique not only reduces the verification time but also increase the reliability of the system by increasing the fault coverage.

There exist currently in production an immense number of applications that are considered safety critical, meaning that the execution of them is directly related to issues concerning the well being of people. A domain where these applications are particularly present is in the aeronautics industry. A piece of critical software that’s embedded into an airplane’s calculator cannot, under any circumstance, fail while the aircraft is in-flight. And this restriction becomes more and more severe when the priority of the application escalates. This situation also poses an inconvenient at the moment of testing software. Since for applications to be tested on their real environment (flight test) it is necessary to have certain guarantees that it won’t fail, other methods such as unitary tests and simulations have to be used. But none of these methods are sound, meaning that if some particular case is unintentionally left out of the executions, then the behavior of the program in such scenario is not contemplated in the performed analysis. But when we are talking about safety critical applications, these small cases could mean a very big difference. This is why more and more companies that produce this kind of software are starting to include in their verification process sound techniques to validate the absence of run-time errors on their programs. Particularly Airbus, one of the main aircraft manufacturers of the world, uses AstréeA, a static analyzer based on abstract interpretation, to prove that the programs embedded in their calculators cannot possibly fail. In the following report an investigation will be presented were AstréeA was used at Airbus to prove the absence of run-time errors on the ATSU. The introductory chapter presents a description of the software analyzed, an explanation of the objectives set for the project and its scope. Then, on chapter 2 all the necessary theoretical concepts will be presented. Sections 2.1 - 2.3 give an overview of the basics of abstract interpretation, while section 2.4 presents the analyzer used. Then chapters 3 and 4 describe in depth the solution given and how the investigation was carried out. Finally chapters 5 and 6 enter into the presentation and analysis of the results obtained in the period of study and the current state of the solution.

Global Positioning System (GPS) has revolutionized the way we navigate and get location-based information in the last decade. Unfortunately the accuracy of civilian GPS is still remaining in meter level and it does not work well in indoor environment, which is a major drawback for applications such as autonomous vehicle, robot machine and so on. UWB (Ultra-wideband) is one of the most promising technologies to solve this problem. The UWB technology has large bandwidth and it is quite robust to fading and multipath effect. Therefore, it is capable of high accuracy down to centimeters for positioning in both outdoor and indoor scenarios.  The IEEE 802.15.4a was released in 2007, which adopted UWB in this standard and specified its physical layer for accurate positioning in WPAN (Wireless Personal Area Network). Apart from the capability of accurate positioning, solutions based on this standard will have quite low power consumption and low cost.  In this thesis work a positioning system based on IEEE 802.15.4a has been designed. A few practical constrains have been taken into account in designing the system, such as performance, cost, power consumption, and governmental regulations and so forth. To reduce the system complexity and communication channel occupancy, TDOA (Time Difference of Arrival) has been chosen as the ranging protocol. The system has been designed accordingly. Main components have been selected and PCBs (Printed Circuit Board) has been designed as well. The design work covered both hardware and software. The proposed system is believed to be able to achieve a positioning accuracy of ±20 centimeters.

Hash functions are a type of mathematical function that takes an input of arbitrary length and gives an output of fixed length, called a hash value. Many consider hash values to be sort of a “fingerprint” of some data, since they can be assumed to have unique outputs for any possible input. This assumption does not make a lot of sense, because the size of the input space is infinite while the size of the output space remains finite! It is, nevertheless, defensible because of the infeasibility of finding two inputs that yield the same hash value. Hash functions are often used to store passwords in databases since it is not feasible either to go from a hash value to a preimage. By saving the hash value rather than the password and checking the hash value of the user’s input it is possible to check passwords without the need to store them, which is an advantage if one wants to control the damage of a possible data leakage. This work researches different cryptanalytic techniques for searching for preimages to hash values in a password-cracking context. A 27% increase in performance is gained using a time/memory tradeoff instead of naively iterating through password candidates. The attack is also demonstrated in practice, where it attains a 50% improvement. The data is then analyzed and discussed for the purpose of assessing the implementability of the attacks in already existing cracking implementations.

The Swedish Armed Force has a need to streamline the preparation of an exercise in tactical training for soldiers because the preparation of an exercise is currently laborious and time consuming when commanders manually share files to each computer that will be included in the exercise. The degree project was to create a file sharing program within the internal network of the Defence Forces’ international centre (Swedint). The program is used by the training management to effectively start up an exercise where soldiers train tactical training in a simulated combat environment. The training management use this program to upload files to a server which notifies all the computers on the network to download these files. The program is developed in C# which is an object oriented programming language that is a part of the Microsoft platform .NET and it will be used in Swedint/livgardet in Kungsängen and possibly in the facility in Kvarn and Hemvärnets stridsskola. In comparison with the previous method used to distribute files, this program streamlines the process of file transfer up to 80 times, and thus reduces the work for the training management in terms of preparation for an exercise. / Det finns ett behov hos Försvarsmakten att effektivisera förberedelserna av en övning i taktisk träning för soldater, då det idag är arbets- och tidskrävande för övningsledningen att manuellt dela filer till varje dator som ska ingå i övningen. Examensarbetet handlar om att skapa ett fildelningsprogram för det interna nätverket på försvarsmaktens internationella centrum (Swedint). Programmet används av övningsledningen för att på ett effektivt sätt ladda upp filer på en server som i sin tur initierar samtliga datorer på nätverket att ladda ner aktuella filer. Programmet är utvecklat i det objektorienterade programmeringsspråket C#, .NET och ska användas på Swedint/livgardet i Kungsängen och eventuellt även anläggningen i Kvarn och Hemvärnets stridsskola. I jämförelse med den tidigare använda metoden för att distribuera filer, effektiviseras övningsledningens arbete upp till 80 gånger vid användandet av det utvecklade programmet. På så sätt har övningsledningens arbete reducerats när det gäller förberedelse av en övning.

As "Voice over IP" has become more prevalent and many client and server applications have been designed for them, the VoIP industry has seen the need for faster, more capable systems to keep up. Traditionally, system speed-up has been achieved by increasing clock speeds but, conventional single-core CPU clock rates have peaked a few years ago due to very high power consumption and heating problems. Recently, system speed-up has been achieved by adding multiple processing cores to the same processor chip called multi-core processors. The existing VoIP applications cannot attain full benefit and efficiency of multi-core processors because of their sequential design. \VoIP paralleled client-server software for multicores" that can split up sequential code and run concurrently on multiple cores instead of trying to exploit single-core hardware is the solution. We have created a model of generic, open source paralleled VoIP-server (IOpen-VoIP) in C that suits multi-core and that can be used as a simulation tool. Furthermore, we have designed and implemented a tool for performance testing. It can be used for performance evaluation of IOpenVoIP and other SIP servers. The tool emulates thousands of communication sessions through a server. Performance testing can help developers to eliminate bottle necks in multi-core server design. On the other hand side, VoIP clients are not just used for voice and video communication over Internet. Along with audio and video they can carry other real time data i.e. patients ECG signals. Raw data is usually sent from one end and it is processed at other end which is a processor intensive task. We designed and implemented a graphical VoIP-Client which utilizes multi-core processors.