Dec 09, 2016 02:47 PM
Vabad lõputööde teemad / Available topics for BA and MA theses
Write a computer game that makes use of distributed memory parallel processing for either physics and/or graphics computations. email@example.com
QUIET LIQUID COOLED COMPUTERS
Computers can generate much heat when in use. This is especially true in data centers, but is also true for gaming machines. The object of the thesis will be to examine how a quite immersively cooled computer should be designed and operated. The thesis can comprise practical and theoretical aspects. firstname.lastname@example.org
PROGRAMMING HSA ARCHITECTURE
OpenCL and HSA (heterogeneous systems architecture) are two different methods of programming computers with CPUs and GPUs. Compare the performance and user-friendliness of the two programming methodologies. email@example.com
HPC AND HCI
High performance computing (HPC) has traditionally been viewed as a tool for experts. With personalized manufacturing enabled by advances in 3D printing, many more ordinary people will use HPC to help design their household items. The aim of the thesis would be to improve human computer interaction (HCI) for a program or set of firstname.lastname@example.org
PARALLEL FILE SYSTEMS
Examine performance and configurability of open source parallel file systems such as Lustre, Ceph and Glusterfs. High performance computing workloads are of greatest interest (typically large files), but search applications on many small files and cloud workloads can also be email@example.com
Perform a survey of available computer benchmarks. Run some of the bench- marks to see how reproducible the results are. Determine how they might influence your choice in buying/building a computer. Distinguish between full application benchmarks and kernel level benchmarks. firstname.lastname@example.org
Parallelisation of numerical solution of Helmholtz equation using Domain Decomposition method and implemented with numpy, scipy and mpi4py
GPS data Mining for Urban Map-Matching
With rapid increase in using GPS data in many application and devices a necessity to correct GPS error is a must, especially in urban areas. The purpose is to use raw GPS data collected from vehicles and try to match all the GPS data on the appropriate roads in near real time (This process is called Map-Matching). In order to make a good map-matching process, we have to take into account the logic of the trajectory and also the behavior of the drivers; for these reasons mining the data and analyzing it from spatio-temporal aspect will be very helpful step in the achievement of solving this problem.
Road Traffic Flow Estimation using Inductive Loop Data
The main idea behind this subject is to use Inductive loop data and estimate the traffic flow on the concerned roads and do a propagation of the flow via the roads in the city and evaluate the traffic situation through the prediction made by applying probabilistic algorithms or Monte Carlo Algorithm.
Trajectory Reconstruction using Call Detail Records Data (CDR).
The sparseness of Call detail Record data in time and space make it hard to recognize the real trajectory taken by the mobile phone users. Hence, the main purpose is to reconstruct the real trajectory by using the sparse information from the CDR data.
Spatio-temporal visualization framework to analyse mobile data
The visualization of mobile data helps to understand the behaviour or movement patterns from mobile users. The user's movement has a distribution in space and time. The goal is to develop a framework to show reactive and dynamics map representations from user mobile data by applying visual analytic techniques.
Previously proposed themes:
Geolocation-aware web-form submission technologies [taken!]
Survey and comparative analysis of existing web-form
solutions with support for geolocation data collection through supported
mobile and web browsers and related technologies. email@example.com
Bachelor + Master Level:
Personal budget managing app for android with geolocation support [taken!]
Developing a geolocation-enabled apps designing
platform with an application of easier management of personal finances
through a smart UI with location-triggered features and firstname.lastname@example.org, email@example.com
Asukohapõhiste veebisündmuste algoritmid (twitteri andmete näitel)
1. Kaardil tweetimise aktiivsuse kajastamine.
Antud on kaart + aja sliderid. Nende abil valides sobiva aja tekivad
kaardil tweetimise hotspotid. Tulemustes peaks olema siis tweetide arv
ning ka kasutajate info.
2. Andmekaevega seotud ülesanne
– otsida seoseid populaarsete asukohtade ning tweetimise aktiivsuse
vahel. Ehk siis uurida, kas asukohaga tweet on kasutaja teadlik valik (a
la olengi mingi monumendi juures ja tahan sellest teistele teada anda)
või on mul „location“ kogu aeg enabled. See võiks Twitterile endale ka huvi pakkuda.
Eerltoodud ülesanded peaksid toimima reaalajas, mis tähendab, et päringud peavad toimima suurte andmemahtude pealt.
Juhendaja: Elis Kõivumägi (elis.koivumagi (ät) gmail.com)
Automatiseeritud testklaster Linuxi kerneli testimiseks
Eesmärgiks on automatiseerida Linuxi kerneli arendusversioonide testimist hulga erinevate arvutite peal. Selleks tuleb alustuseks üles seada ja häälestada Jenkins või muu sarnane keskkond, ning siis selle külge liidestada erinevate nõuetega klientimasinad, neis automaatselt uut tarkvara kompileerida, testida erinevate testidega ning koguda vigade ja staatuse infot
Juhendaja: Meelis Roos
Fault Tolerant Networks
Taken by Anders Martoja
Juhendaja: Artjom Lind
Private Cloud Infrastructure on HPC Nodes
Taken by Andre Tättar
Juhendaja: Artjom Lind
Large-Scale Provisioning and Configuration Management
Taken by Dmitri Tšumak
Juhendaja: Artjom Lind
Overview of Apache Web Server Customizations
Taken by Annika Laumets
Juhendaja: Artjom Lind
In addition take a look at DS Seminar topics in Peer-to-Peer Computing and Computer Vision
Vt ka http://www.ut.ee/~tqnu/Teemad.html
NB! Teemasid on veel... Võib vabalt pöörduda ka oma teemaga. Aitame leida sobiva juhendaja!
Mõnda ülaltoodud teemadest on võimalik siduda osaliselt tasustatud tööga reaalses arvutiteadusega seotud projektis!
Varasemalt valitud teemasid / Some previously chosen topics
"Adapting machine learning algorithms on MapReduce" (Master thesis)
- Sergei Laada
Supervisor: Pelle Jakovits
"Comparing Spark and Hadoop MapReduce for iterative scientific algorithms", (B) (preliminary name)
- Lauri Taaleš
Supervisor: Pelle Jakovits
"Parallel scientific simulation using Bulk Synchronous Parallel model", (B) (preliminary name)
- Ilja Kromonov
Supervisor: Pelle Jakovits
"Logistic tracking web system using Cloud (web) services", (B) (placeholder name)
- Kaspar Lippmaa
Supervisor: Pelle Jakovits
Dynamic allocation tracking (B) - Viktor Karabut
Supervisor: Vladimir Shor
Arduino based tracking framework for object localization (B) - Steve Magi
Supervisor: Satish Narayana Srirama
Tracking human movement in a smart room (B) - Martti Marran
Supervisor: Satish Narayana Srirama
iOS application for tracking fuel usage of automobiles (B) - Triin Uudam
Supervisor: Satish Narayana Srirama
Supervisor: Satish Narayana Srirama
Smart Homes (B) - Tanel Lips
Supervisor: Satish Narayana Srirama
Comparison/review of concurrent programming languages (Clojure), (B) - Kristjan KeltIterative Solvers using GPGPU, (M) - Kait KasakReview of parallel programming framework Charm++, (B) - Tõnu Jaarma
2011/12 õppeaastal väljapakutud teemasid / Available Topics to choose from (2011/12 -- many of the topics are still on share - contact the supervisor:)
All of the following topics can be laid out either as BsC or as MsC thesis (advisors are shown in brackets).
B means Bachelor, M means Master. A number in front of one of these letters means that there are several theses offered in this topic
There are also some thesis topics specific to the work of Ulrich Norbisrath on this page.
Scientific Computing on the Cloud (SciCloud) (Satish Srirama)
Here are some of the Bachelors(B)/Masters(M) thesis topics associated with the SciCloud project. For detailed description of the project, click here
The cost of virtualization for scientific computing. (B/M) (Satish Srirama, Pelle Jakovits)
Description: What are the downsides of using platform virtualization for deploying applications on the cloud. Virtualization provides many benefits for clouds, but it also has disadvantages. Goal is to measure and quantify the disadvantages of virtualization for applications deployed on the cloud. Benchmark and measure the added latencies (network, IO, CPU, Memory, ...) and slow downs when running applications in virtual machine in comparison to running them directly on the same hardware. Use different visualization platforms (Xen, VMware, KVM, VirtualBox)
Adapting scientific computing problems to JPREGEL BSP framework. (B/M) (Pelle Jakovits)
Description: Bulk Synchronous Parallel is a distributed computing model and JPREGEL is a large scale graph computing framework based on Google's Pregel. The goal of this thesis is to study can scientific computing problems be adapted to this framework and how efficient and scalable the results are. Study should also include comparison of the results of algorithms implemented using MPI.
Adapting scientific computing problems to HAMA framework. (B/M) (Pelle Jakovits)
Description: Apache Hama is a pure BSP(Bulk Synchronous Parallel - distributed computing model ) computing framework on top of HDFS (Hadoop Distributed File System) for massive scientific computations such as matrix, graph and network algorithms. The goal of this thesis is to study can scientific computing problems be adapted to this framework and how efficient and scalable the results are. Study should also include comparison of the results of algorithms implemented using MPI.
Adapting scientific computing problems to HaLoop MapReduce. (B/M) (Pelle Jakovits)
Description: Previous results have shown that Hadoop MapReduce has serious problems with iterative algorithms. The goal of this thesis is to study alternative MapReduce framework, HaLoop for iterative (suggestion:Scientific) algorithms. Study should also include comparison of the results of algorithms implemented in both Hadoop and HaLoop MapReduce.
Utilizing Q array processing programming language and kdb+ database for scientific computing. (B/M) (Pelle Jakovits)
Description: Kdb+ is proprietary but kdb+ 32 bit is usable for non-commercial use. Adapt a number of scientific algorithms or applications, analyze and compare the results to other distributed computing solutions. (Hadoop, Twister, MPI, BSP, ...)
Eucalyptus private cloud health monitoring application. (B/M) (Satish Srirama, Pelle Jakovits)
Description: Study existing solutions and create a cloud health monitoring and status visualization system for SciCloud private cloud, according to the project requirements.
Adapting MPI based legacy parallel scientific computing applications to BSP. (B/M) (Pelle Jakovits)
Description: Study what are the difficulties of adapting existing MPI based parallel applications to the BSP model. Adapt different case study examples and analyze the results. Can choose BSP implementation freely.
Mesos, implementing multi-framework distributed computing platform on SciCloud. (B/M) (Satish Srirama, Pelle Jakovits)
Description: The goal of this thesis is to study creating Mesos eucalyptus images on SciCloud that include multiple distribute computing frameworks that share the same computing resources. The thesis should identify problems that may arise for frameworks that have to share resources with other distributed computing frameworks, and try to analyze the effects of these problems to the efficiency of applications running on these frameworks.
Implementing NAS Parallel Benchmarks (NPB) on MapReduce. (M) (Satish Srirama, Pelle Jakovits)
Description: The goal is to implement the NAS parallel benchmarks for MapReduce framework. And utilize these benchmarks to find better configuration options for different types of MapReduce applications.
Efficiency of Hadoop MapReduce applications. (M) (Satish Srirama, Pelle Jakovits)
Description: The goal is to study and document what affects the efficiency of applications written for Hadoop MapReduce framework. Implement some data heavy and processing heavy MapReduce applications, profile and measure the effects of different improvements to Hadoop MapReduce framework and different coding practices. Of the MapReduce model.
Adapting graph processing algorithms to MapReduce. (M) (Satish Srirama, Pelle Jakovits)
Description: The thesis should identify some graph processing algorithms that can be adapted to the Hadoop MapReduce algorithm. The thesis should also demonstrate several such application along with performance latencies and analyze the efficiency compared to existing non-MapReduce solutions. Focus should be on non Iterative MapReduce algorithms for processing very large graphs. Aim is to use MapReduce frameworks like Hadoop for graphs that would not fit into the collective memory of the cluster. Algorithms can be iterative by design but the number of iterations should be very low and should not depend on the size of the graph.
Using Hadoop ZooKeeper to manage tasks in distributed applications (Synchronization, configurations, worker grouping). (B/M) (Satish Srirama, Pelle Jakovits)
Description: ZooKeeper is a distributed configuration and management service for distributed applications. The thesis should research using ZooKeeper to manage concurrent tasks in distributed applications. Implement a distributed application that uses ZooKeeper for most of the management tasks like choosing the master (initially or when current master fails), synchronization/barriers, group membership (which concurrent nodes are available/active/free), etc.
Developing social networks and mash-up applications with cloud application services (Several Bachelor/Master theses) (Satish Srirama)
This topic gives the student to define and come up with some applications that can be deployed with the public/private cloud platforms.
Example: Keeping track of a researcher’s calendar and location.
Let Alice is a researcher who attends conferences regularly and presents her results. She wants to keep track of herself and want to have a map of her current and future conferences. The idea is to have a map of her travel on the Google maps with storage of presentations and probably some pictures on cloud storages like Amazon S3. The application should also have support, to store all the relevant data regarding each publication like versions, reviews etc. As an extension, the application should remind her via mail, 4 weeks before the future conference.
Example 2: Can you extend the application to all of the employees of her department or can you think of building a scientific community? Can you think of any social networks from this scenario? Something like say the researchers she met during the last conferences, and probably reminding of their reappearance at alternative conferences.
Implementing Cloud based Security Services for mobiles (B/M) (Satish Srirama, Pelle Jakovits)
Description: The goal is to implement security services on the cloud, which can be offered to mobile devices.
Example: Speech recognition service for identifying mobile device users. Mobile device owners speech patterns are stored in the cloud and is used to identify mobile device users. The user has to read a certain sentence (provided by the service) aloud, and the recording of the voice is sent to the cloud for speech recognition. If the recognition fails, then some of it's functionality, application or even the mobile device itself is not enabled.
Using Hadoop Hbase as a source for input/output data in Hadoop MapReduce applications. (B) (Satish Srirama, Pelle Jakovits)
Description: The goal it to study utilizing the Hbase (Hadoop distributed database) as input and output medium for MapReduce applications. For example Graph processing algorithms. Compare the results to other Hadoop MapReduce input/output sources. (HDFS, …)
Large-scale distributed data mining with Hadoop Hive. (B) (Satish Srirama, Pelle Jakovits)
Description: The goal of this thesis is to research which data mining algorithms can be adapted to Hive. The thesis should demonstrate several such algorithms and compare them to existing non-MapReduce solutions.
Using Hadoop ZooKeeper for cluster management/scaling. (M) (Satish Srirama, Pelle Jakovits)
Description: This thesis should research using the ZooKeeper service for dynamic cluster management and scaling. Nodes connect to ZooKeeper service and use it to elect a master. Master uses ZooKeeper service to manage the connected and active nodes and scale the dynamic cluster if needed.
Building mobile social network application (Several B/M) (Satish Srirama)
The idea of the thesis is to build interesting social network application using cloud services from facebook, twitter, Amazon web services and google maps.
CVS or SVN on the Eucalyptus cloud (B/M) (Satish Srirama)
The idea of the thesis is to establish CVS or SVN on the Eucalyptus cloud or come up with the best solution, which is already available in the cloud domain.
Implementing a scalable, efficient linear system solver using MapReduce. (M) (Satish Srirama, Pelle Jakovits)
Description: The goal is to research, study and implement an applicable linear system solver application using MapReduce. It should be scalable and efficient enough to be usable in other applications. Compare it to already existing solutions. (Sequential, MPI based, ...)
General topics (Satish Srirama)
Issue tracking system (B) (Satish Srirama, Jurmo Mehine)
Description: We are looking for an able programmer to create a web based issue tracking system/collaboration tool for managing different research projects within the distributed systems group. Details of the system are not yet fleshed out, but it should allow basic task management, be integrated with Git version control and have built in time-tracking.
Mobile Cloud (Satish Srirama)
Peer-to-Peer Mobile Photo Sharing: (Bachelor Thesis) (Satish Srirama, Huber Flores)
Increasing advances in mobile technologies have enable the handset becoming a service provider rather than a service consumer. Moreover, handsets are equipped with multiple variety of sensors that allow monitoring the activities of the user. In this thesis, such capabilities are explored by developing a mobile application that allows sharing and editing pictures in real-time between mobiles. The application makes use of the sensors (e.g. accelerometer) for implementing basic editing functions such as undo or redo (based on the movements of the user).
XMPP protocol implementation between the mobile and the cloud: (Master Thesis) (Satish Srirama, Huber Flores)
Communication between the cloud and the mobile phone usually follows the synchronization protocol (SyncML). However, in some critical scenarios is necessary to have real-time communication. Thus, the idea of this project is to implement XMPP in both, client and server sides.
Mining Sensor Data for the Recognition of User Activities: (Master Thesis) (Satish Srirama, Huber Flores)
Embedded technologies such as sensors within the smartphones enable mobile applications adapting dynamically in real-time to the user's context. The idea of this thesis is analyzing the information provided by the sensors (create algorithms to find patterns) and collected with real-time databases (Sqlite, tokyo cabinet, etc.) with the intention of inferring what the user is doing in order to share such status in a social network such as Twitter/facebook.
Real-time Hadoop for Mobile Devices: (Master Thesis) (Satish Srirama, Huber Flores)
Hadoop is a framework that allows distributing computing tasks in parallel. However, such implementation is not available in real-time for mobiles. This project consists of creating a middleware framework feature for providing parallel computing to the handsets.
Mobile Video Maker Using the Cloud: (Bachelor/Master Thesis)
Agents are entities that enables dynamic peer-to-peer communication between mobiles. This project consists of implementing a mobile application based on agents (hint: Jade is a framework for creating agents on Android) for generating a video sequence in the cloud. Each mobile device implements an individual agent that is listening for requests. When the application triggers the video maker feature the agent sends the requests to the other devices (agents) for sending the pictures to the cloud. Once the pictures are located at the cloud a Map/Reduce process joins the pictures for creating the video (Using MCM). A link of the video is sent back to the device who trigger the video maker event
Mobile Cloud Middleware Quality of Service: (Master Thesis) (Satish Srirama, Huber Flores)
Mobile Cloud Middleware is a framework that enables to perform data-intensive invocation of cloud services from the mobile phone. However, A queue middleware mechanism has to be implemented (with mathematical analysis/model) for providing QoS for the mobile users.
Drag and Drop Image Authentication for Social Network Using Mobiles: (Bachelor Thesis) (Satish Srirama, Huber Flores)
Ergonomic improvements in mobile devices enables the user to provide inputs to the mobile with considerable ease. However, inputs that involved an authentication process is relatively long, and thus introducing such long inputs from the mobile is hard for the user. This project propose to develop a mobile application that enables the user to create an image based on different sources (local or external – cloud) that can be use as an authentication input (password for social networks, e-amil, etc). Since the image is store locally, the user just have to drag and drop the picture on the login interface.
Generating an Indoor Schema for Mobile Contextualization using QR-codes: (Master Thesis) (Satish Srirama, Huber Flores)
The magnetic field is a sensor that enables inferring direction and orientation in the context of the user. QR-codes is the graphical representation of a resource that can be analyzed with the mobile. The idea of this thesis consists of mixing both elements for generating an indoor schema of the room based on the tracking of the mobile and the reading of QR-code resources.
Remotely Control of an Android Phone using VNC Server: (Bachelor Thesis) (Satish Srirama, Huber Flores)
The project consists of creating a Desktop application that allows connecting to the Android device remotely from a personal computer. Once connected, the application has to show all the information on the device (applications installed, files, etc.). Since the principal aim of the remote connection is the management of resource, the application has to allow installing apk files remotely.
Migration from native android applications to HTML5 applications (Master Thesis). (Satish Srirama, Huber Flores)
HTML5 is the prominent technology that will enable mobile platform in dependency. The idea of this thesis consist of creating an application that takes a native android application and transforms it into a HTML5 application.
Data Forensic in the Mobile Phone on the Cloud: (Master Thesis) (Satish Srirama, Huber Flores)
The idea consists of moving all the content from the physical device to the emulator so the emulator can be analyzed and several instances of the mobile can be executed in several machines running on the cloud.
Mobile Gallery Analysis based on Parallel Image Processing on the Cloud Using MapReduce: (Master Thesis): The project consists of creating several recognition algorithms for identifying objects, people, etc. from pictures taken by the mobile phone. The algorithms can be implemented using MapReduce.
Load Balancing on the Cloud for Analyzing Middleware Scalability: (Master Thesis) (Satish Srirama, Huber Flores)
Mobile Cloud Middleware (MCM) is a framework that enables to perform data-intensive invocation of cloud services from the mobile phone. This thesis consists of a comparison of Load Balancing techniques with multiple tools (HAProxy, etc) in order to analyze the Scalability of the MCM Framework. (Using tools such as Tsung, JMeter, etc.
SyncML Mobile Host: (Master Thesis) (Satish Srirama, Huber Flores)
Increasing advances in mobile technologies have enable the handset becoming a service provider rather than a service consumer. Mobile Host is an implementation that allows providing services from the mobile. This thesis creating extending a Synchronization engine for the communication peer-to-peer using mobile phones.
From IPv4 to IPv6 discovery mechanisms for Mobiles: (Master Thesis) (Satish Srirama, Huber Flores)
With the introduction of IPv6 at the end of 2012, the migration of IPv4 to IPv6 has to be transparent as possible. This thesis consist of analyzing the transition from Ipv4 to Ipv6 and implementing a discovery mechanism that enables the communication peer-to peer between mobiles.
Friend-to-Friend Computing (Spontaneous Desktop Grid) (Ulrich Norbisrath, Artjom Lind)
message splitting (B-M)bittorrent message transport (B-M)language and sandbox support in f2f (llvm, javavm)
python, c, java, fortran (1M, 3B)
better user interface for pidgin (B)implement otr (security) for f2f message exchange (B-M)plugins
new sip-communicator plugin (B)Microsoft Live (B)new Skype plugin (B)Trillian (B)Miranda (B)
build environment for f2f (M)upnp for routers (B)apple nat-pmp (B)
Information Retrieval and Knowledge Management
email parser in python (imap+all metadata) (Georg Singer, Ulrich Norbisrath) (1B)search experience tracking plugin (Georg Singer, Ulrich Norbisrath) (1M)distributing queries for huge graphs (Ilja Livenson, Ulrich Norbisrath) (1M)
Grid and Cluster Computing
Distributed Tracer-Framework (Marti Taremaa, Ulrich Norbisrath): General-purpose (single-application) trace tools can be used for debugging distributed systems, but for complex setups management overhead quickly becomes an issue. You will have to write (in Python) a helper framework for managing the trace sessions in distributed systems, including setting up and running them and collecting their results. Sound knowledge in remote command execution and some knowledge of tracing tools needed. (B-M)Using PS3 for cluster computing (Ulrich Norbisrath)
development/build environment (virtual machine) (B)stable distribution on whole cluster (unified distribution) (B)
possible keywords (Debian Distribution Managament, ROX, Google for Clustersystem admin, Orca, ...)
Optimizing a graph search algorithm for the PS3 cluster (1-2M)
implementing a cloud service (1-2B-M) (Satish Srirama)implementing a mashup application (1-2B-M) (Satish Srirama)implementing an application on Hadoop (1B) (Satish Srirama)Using HadoopFS for Cloud Storage implementation (1B-M) (Satish Srirama, Ilja Livenson)
Scientific and Parallel Computing (Oleg Batrashev, Eero Vainikko)
Comparison/review of concurrent programming languages/frameworks:
Compare Erlang/Scala/Oz/Clojure programming language constructs and techniques with one of widely used imperative languages (C, Java). [more]Actor-based and/or distributed algorithms in Scala (Oleg Batrashev)
Charm++ is a runtime library to let C++ objects communicate with each other efficiently. The programming model is thus very much like CORBA, Java RMI, or RPC; but it is targeted towards tightly coupled, high-performance parallel machines. It uses the ``single program, multiple data'' (SPMD) programming model made popular by MPI. (Oleg Batrashev)
ParFUM - framework in Charm++ for adaptive parallel programs that utilize unstructured meshes. One possibility is to explore how FEM is implemented in ParFUM. (Oleg Batrashev)
Chapel - an emerging parallel programming language. For example, implement preconditioned Conjugate Gradient. (Oleg Batrashev)
Review any formal model of concurrency (Oleg Batrashev, Eero Vainikko): Concurrent programs are difficult to write and debug, they expose many new problems like deadlocks, race conditions and network failures. The theory of concurrent systems, though not widely known, aids to tackle these problems with formal models of concurrency. [more]Web-crawling and PageRank problem Solution (Eero Vainikko): Web-crawling techniques are used in major search engines, like Google. The PageRank problem leads to a problem of solution of a large system of linear equations. In this topic you can do an interesting research in an area, which modern search engines are not giving out their source code so easily... [more]Comparison of speed of math-apps on Fortran/Java/Python/psyco/scipy/jython (Eero Vainikko, Ulrich Norbisrath): -Possible 2 theses- Speed comparisons of mathematical applications in different languages. [more]Speed Gains via Piped Communcation (Ulrich Norbisrath): The initial handshake for building up connections via the network is time-consuming. Would keeping connections alive or even busy increase the speed and latency in this case? This thesis will make tests to prove and find out, at which specific settings a connection between peers is optimal, regarding speed.Profiling DOUG with Paraver (Oleg Batrashev)Web DoS atack prevention with the help of distributed computing tasks (Eero)NFC (Near Field Communication) technology and possible security protocols (Eero)
Application development in high level shader languages (Ulrich Norbisrath): in-depth overview of currently available programming languages for programming the GPU (graphics processing unit) and their applications, development tools and patterns.Solve mathematical problems on GPU/multicore systems (Eero Vainikko): to utilize all available processing power of a personal coputer equiped with a modern graphics card one must balance the calculations carefully.GPGPU - General Purpose computing on Graphics Processing Units (Eero vainikko): re-implement known algorithms on GPUs to take advantage of its processing power.GPU-accelerated network IDS (Meelis Roos): build a network intrusin detection system like Snort and Suricata but implemented on GPU for greater bandwidth
Software Engineering for Distributed and Embedded Systems
Management tools for GRID workflows (Eero Vainikko): [more]
Visualization of Legacy Fortran Applications (Ulrich Norbisrath): -multiple theses- There are plenty of Fortran applications out there, which are very hard to maintain and to extend. To re-use these applications, you first have to get an overview about the whole system and its parts. These theses will develop different methods to visualize different aspects of legacy Fortran applications. One of the reviewed Fortran example will be DOUG, which was developed in the ds-groups here in Tartu.
Web-service for solving integral equations (Eero Vainikko, Ulrich Norbisrath): Web service for an existing integral equation solver, written in SciPy. [more]Development of Web Services-based GRID applications (WSGA) (Eero Vainikko): [more]
Support for debugging of distributed Java applications (Vladimir Shor, Ulrich Norbisrath): Evaluate and compare various possibilities of debugging distrubuted Java applications from different development environments. [more]Comparison of open source UML modeling tools (Vladimir Shor, Ulrich Norbisrath): Find and compare open source UML modeling tools and their support for different UML features. [more]
Efficient object age monitoring in the JVM (1M) (Vladimir Shor)Information about the age of objects can help in analysis and prediction of memory leaks. However, finding an efficient algorithms and methods for this task is a big challenge. To find the solution you will have to learn at least Java Virtual Machine Tooling Interface (JVMTI), Java Native Interface (JNI) and Byte code instrumentation. Topic requires good knowledge of both C/C++ and Java. Dynamic allocation tracking (B) (Vladimir Shor) - Taken by Viktor KarabutDevelop the Java/native agent that will track object allocations with the ability to dynamically specify the class of interest. You will might have to learn Java Virtual Machine Tooling Interface (JVMTI), Java Native Interface (JNI) and Byte code instrumentation. Topic requires knowledge of both C/C++ and Java.