Communication and computation systems

Dependable network architectures

The focal concept of the “Dependable network architectures” direction is dependability of a computing system – an integrated property of a computing system jointly characterized by its availability, reliability, security and maintainability. It is recognized in the ICT research community that an evaluation of dependability properties is currently done a posteriori, after particular algorithms are designed and implemented as a software operating on top of already deployed network architecture.  This situation led to several known serious accidents where the failure of the computing system caused severe economical damages and even human casualties.

Our research focuses on creating a systematic framework for engineering of dependable communication systems where the dependability processes are assessed already at the design stage.

In order to enable such a framework requires modularization of the existing communication solutions and identification of atomic functional components. These components should further be characterized and classified with respect to addressing particular dependability properties in different scenarios. We research on dependability principles in LTE, Mesh and sensor networks.

Algorithms

The research aims at constructing fast and memory efficient algorithms and programs. Our current work concerns mainly how to organize and search large data sets and geometry objects. In particular, we have designed efficient algorithmic solutions for retrieval of important information from large datasets such as the numeric characterization of some key properties of the data.


Dependable software engineering

Our research in dependable software engineering is mainly concerned with the following questions:

• How should a programming language that allows convenient programming with event-driven components be defined?

• How does one implement such a programming language on platforms that range from desktop clusters to tiny embedded devices?

• Can this programming language be formalized in a way that allows platform-independent reasoning of even its timing properties?

• Would this programming language make it possible to raise the abstraction level of programming "in the small" as well as "in the large"?

Most of our efforts to date are channeled into the Timber language, which is partially developed by members of the computer science group. Our current research activities involve work on supercompilation, formal semantics, abstract interpretation, incremental garbage collection, graphical representation & transformation, and polymorphic subtyping.
 

Contacts

Groupleader: Valeriy Vyatkin

Research areas:
Pålitliga nätverksarkitektur: Evgeny Osipov
Pålitliga algoritmer: Jingsen Chen
Pålitliga mjukvara utveckling: Johan Nordlander