Browsing by department "Tietojenkäsittelytieteen osasto"

The suffix array is a space-efficient data structure that provides fast access to all occurrences of a search pattern in a text. Typically suffix arrays are queried with algorithms based on binary search. With a pre-computed index data structure that provides fast access to the relevant suffix array interval, querying can be sped-up, because the binary search process operates over a smaller interval. In this thesis a number different ways of implementing such an index data structure are studied, and the performance of each implementation is measured. Our experiments show that with relatively small data structures, one can reduce suffix array query times by almost 50%. There is a trade-off between the size of the data structure and the speed-up potential it offers.

Data compression is one way to gain better performance from a database. Compression is typically achieved with a compression algorithm, an encoding or both. Effective compression directly lowers the physical storage requirements translating to reduced storage costs. Additionally, in case of a data transfer bottleneck where CPU is data starved, compression can yield improved query performance through increased transfer bandwidth and better CPU utilization. However, obtaining better query performance is not trivial since many factors affect the viability of compression. Compression has been found especially successful in column oriented databases where similar data is stored closely in physical media. This thesis studies the effect of compression on a columnar storage format Apache Parquet through a micro benchmark that is based on the TPC-H benchmark. Compression is found to have positive effects on simple queries. However, with complex queries, where data scanning is relatively small portion of the query, no performance gains were observed. Furthermore, this thesis examines the decoding performance of the encoding layer that belongs to a case database, Fastorm. The goal is to determine its efficiency among other encodings and whether it could be improved upon. Fastorm's encoding is compared against various encodings of Apache Parquet in a setting where data is from a real world business. Fastorm's encoding is deemed to perform well enough coupled with strong evidence to consider adding delta encoding to its repertoire of encoding techniques.

Organizations are adopting cloud technologies at an increasing rate. Significant share of growth of cloud deployments is coming from application migrations to cloud computing. Migrating existing legacy applications to cloud computing platform is not a trivial task. A migration methodology will help migrating applications to cloud more effectively and with lower risk than doing it by trial and error. A part of the cloud migration process is the selection and execution of a migration strategy amongst the possible, situational and commonly used options. The migration strategy defines many of the migration process activities since they depend on cloud architecture and service and deployment models, which are implicitly set by the migration strategy. Many of the existing cloud migration methods don’t specify the factors that lead to migration strategy selection. The migration strategy selection is a critical part of migration planning involving multiple organisations and several individuals. This thesis presents categories of migration strategy factors derived from a cloud migration methodology and process framework review and validates the factors by doing a deductive thematic analysis against qualitative case study interview data. By having a clarity and a way to address the migration strategy factors, will increase the migration success rate and reduce planning time.

Ohjelmointikielen valinta on tärkeä osa ohjelmistoprojektien toteutusta. Vaikka ohjelmointikielet uudistuvat nopeaan tahtiin, nykypäivänä on yhä tavallista valita ohjelmiston toteutukseen C-ohjelmointikieli, joka on standardoitu yli 30 vuotta sitten. Tutkielmassa tutkitaan syitä, miksi C on nykypäivänä vieläkin laajassa käytössä uudempien ohjelmointikielten sijaan. Tutkielmassa C:hen verrattaviksi ohjelmointikieliksi valitaan Ada, C++, D, Go sekä Rust. Kaikki viisi kieltä ovat tehokkaita. Tämän lisäksi jokaisen kielten historiassa on ollut tavoitteena korvata C:n käyttö. Ohjelmointikieliä verrataan C:hen suorituskyvyn, muistinkäytön sekä C-yhteensopivuuden osalta. Tämän lisäksi tutkielmassa selvitetään tärkeimpiä C:n ominaisuuksia sekä C:n kehitettävissä olevia ominaisuuksia. Tuloksia käytetään uuden Purkka-ohjelmointikielen suunnitteluun. Muut ohjelmointikielet todetaan suorituskykymittauksissa C:tä hitaammiksi. Tämän lisäksi muiden ohjelmointikielten ominaisuudet, kuten automaattisen muistinhallinnan, todetaan aiheuttavan ongelmia C-yhteensopivuudelle. C:n tärkeimmiksi ominaisuuksiksi nousevat esiin yksinkertaisuus, tehokkuus sekä alustariippumattomuus. Nämä ominaisuudet otetaan huomioon Purkka-kielen suunnittelussa, jossa painotetaan näiden lisäksi yhteensopivuutta C-ohjelmointikielen kanssa. Tutkielmaa varten kehitetty Purkka-kieli on suunniteltu C:n kaltaiseksi ohjelmointikieleksi, jossa on muutettu C:n syntaksia yksinkertaisemmaksi ja johdonmukaisemmaksi. Suorituskykymittauksissa todetaan, että Purkan muutokset C:hen eivät aiheuta suoritusaikaisia rasitteita. Koska Purkka-kieli käännetään C:ksi, se on mahdollisimman yhteensopiva nykyisten kääntäjien kanssa.

Online gaming is more popular than ever and many video game companies are reliant on the cash flow generated by online games. If a video game company wants its game to be successful, the game has to be resilient against cheating, the presence of which can ruin an otherwise successful game. Cheating in a video game can bankrupt an entire company as the non-cheating players leave the game because of unscrupulous individuals using cheats to gain an unfair advantage. Cheating can also involve criminal activity where maliciously acquired in-game items are traded against real money online. Commercial cheat programs are sold on online black markets and are available even to players who have no deep technical knowledge. The widespread availability and easy accessibility of cheats compounds the issue. This thesis will categorize different anti-cheat techniques and give a brief history of anti-cheat starting from the early 1980s. The history section describes how the fight against online cheating began and how it has evolved over the years. This thesis will compare different anti-cheat methods, both on the client-side and server-side, and draw conclusions about their viability. It will also look at scenarios where different anti-cheat methods are combined to create more powerful systems. All the anti-cheat methods will be evaluated based on five different criteria on a scale of 1 to 4, with one being the lowest score and four the highest. The thesis will use a custom-built client-server game as an example to illustrate many of the anti-cheat techniques. Requirements of different types of games, such as first-person shooters and strategy games, will also be considered when reviewing the anti-cheat techniques. Lastly, the thesis will look into the future of anti-cheat and introduce video game streaming and the use of machine learning as possible new solutions to tackle cheating. The conclusion will summarize the advantages and disadvantages of different methods and show which techniques are preferable based on the analysis.

Computer programs are common in our daily environment and new robotic devices emerge constantly. One of the key skills for the future is understanding the general purpose as well as intricate details of both concepts. Programs written in conventional programming languages are purely text with syntax cryptic to an unexperienced observer and errors can easily be made in the writing process, which is far from ideal basis to introduce programming or robotics to children on elementary school level. We set to investigate how a visual learning environment could reduce this chasm between the skills of the young and the perceived difficulty of computer technology and with what assortment of technologies and working methods such platforms, both programming and robotic, could be produced in a project with multiple parties. A joint endeavor was started between two research groups in different universities and the education services of a municipality in Finland. In the project, a visual programming environment was created where the users can assemble programs by stacking and manipulating colorful blocks to form a command flow. The set of functions are traditional with calculations, variables and loops, but also include commands to pilot a robotic platform with wheels, arms and a facial expression indicating display. This robot is a product of the project also, based on a commercial design. The venture can be concluded a success. The learning environment and the roving robot were created in the course of the project and architectural design and selected technologies proved sensible. The empirical exploration accumulated knowledge for academic research.

A recent machine learning technique called federated learning (Konecny, McMahan, et. al., 2016) offers a new paradigm for distributed learning. It consists of performing machine learning on multiple edge devices and simultaneously optimizing a global model for all of them, without transmitting user data. The goal for this thesis was to prove the benefits of applying federated learning to forecasting telecom key performance indicator (KPI) values from radio network cells. After performing experiments with different data sources' aggregations and comparing against a centralized learning model, the results revealed that a federated model can shorten the training time for modelling new radio cells. Moreover, the amount of transferred data to a central server is minimized drastically while keeping equivalent performance to a traditional centralized model. These experiments were performed with multi-layer perceptron as model architecture after comparing its performance against LSTM. Both, input and output data were sequences of KPI values.

Popularity of mobile instant messaging applications has flourished during the last ten years, and people are using them to exchange private and personal information on daily basis. These applications can be freely installed from online marketplaces, and average users may have several of them installed on their devices. The amount of information available from these messaging applications for a third-party eavesdropper via network traffic analysis has therefore grown significantly as well. Security features of these applications have also been developing over the years, and the communication between the applications and the background server infrastructure nowadays practically always employs encryption. Recently, more advanced end-to-end encryption methods have been developed to hide the content of the exchanged data even from the messaging service providers. Machine learning techniques have successfully been utilized in analyzing encrypted network traffic, and previous research has shown that this approach can effectively be used to detect mobile applications and the actions users are performing in those applications regardless of encryption. While the actual content of the messages and other transferred data cannot be accessed by the eavesdropper, these methods can still lead to serious privacy compromises. This thesis discusses the present state of machine learning-based identification of applications and user actions, how feasible it would be to actually perform such detection in a Wi-Fi network and what kind of privacy concerns would arise.

Context: The advent of Docker containers in 2013 provided developers with a way of bundling code and its dependencies into containers that run identically on any Docker Engine, effectively mitigating platform and dependency related issues. In recent years an interesting trend has emerged of developers attempting to leverage the benefits provided by the Docker container platform in their development environments. Objective: In this thesis we chart the motivations behind the move towards Containerized Development Environments (CDE) and seek to categorize claims made about benefits and challenges experienced by developers after their adoption. The goal of this thesis is to establish the current state of the trend and lay the groundwork for future research. Methods: The study is structured into three parts. In the first part we conduct a systematic review of gray literature, using 27 sources acquired from three different websites. The sources were extracted for relevant quotes that were used for creating a set of higher level concepts for expressed motivations, benefits, and challenges. The second part of the study is a qualitative single-case study where we conduct semi-structured theme interviews with all members of a small-sized software development team that had recently taken a containerized development environment into use. The case team was purposefully selected for its practical relevance as well as convenient access to its members for data collection. In the last part of the study we compare the transcribed interview data against the set of concepts formed in the literature review. Results: Cross-environment consistency and a simplified initial setup driven by a desire to increase developer happiness and productivity were commonly expressed motivations that were also experienced in practice. Decreased performance, required knowledge of Docker, and difficulties in the technical implementation of CDE’s were mentioned as primary challenges. Many developers experienced additional benefits of using the Docker platform for infrastructure provisioning and shared configuration management. The case team additionally used the CDE as a platform for implementing end to end testing, and viewed the correct type of team and management as necessary preconditions for its successful adoption. Conclusions: CDE’s offer many valuable benefits that come at a cost and teams have to weigh the trade-off between consistency and performance, and whether the investment of development resources to its implementation is warranted. The use of the Docker container platform as an infrastructure package manager could be considered a game-changer, enabling development teams to provision new services like databases, load-balancers and message brokers with just a few lines of code. The case study reports one account of an improved onboarding experience and points towards an area for future research. CDE’s would appear to be a good fit for microservice oriented teams that seek to foster a DevOps culture, as indicated by the experience of the case team. The implementation of CDE’s is a non-trivial challenge that requires expertise from the teams and developers using them. Additionally, the case team’s novel use of containers for testing appears to be an interesting research topic in its own right. ACM Computing Classification System (CCS): Software and its engineering →Software creation and management →Software development techniques

Slack is an instant messaging platform intended for the internal communications of companies and other organizations. For organizations that use Slack extensively it may provide an interesting source of insight, but as such the data is difficult to analyze. Topic modeling, primarily latent Dirichlet allocation (LDA), is commonly used to summarize textual data in a meaningful way. Instant messages tend to be very short, which causes problems for conventional topic modeling methods such as LDA. The data sparsity problem can be tackled with data expansion and data combination techniques. For instant messages, data combination is particularly attractive as the messages are not independent of each other, but form implicit, and sometimes expicit, threads as the participants reply to each other. Most of the threads in the Slack data are not explicit, but must be ’untangled’ from the message stream if they are to be used as a basis for a data combination scheme. In this thesis we study the possibility of detecting implicit threads from a slack message stream and leveraging the threads as a data combination scheme in topic modeling. The threads are detected using a hierarchical clustering algorithm which uses word mover’s distance, latent semantic analysis, and metadata to compute the distances between messages. The clusters are then concatenated and used as the input for LDA. It is shown that on a dataset gathered from the Gofore Oyj Slack workspace, the cluster-based model improves on the message-based model, but falls short of being practical.