Im a System Engineer, but really called the "computer guy", or the "you that knows how this works" guy to be honest.
That means that whenever the individual researcher or whole research groups needs computer help I am there for them supposed to do my best of the situation.
Together with my colleague Mattias Borell I am responsible for the server structure that contains a whole lot of servers and storage!
As the organization is growing rapidly and we had to take the challenge in optimize the server farm so that the resources is as optimized as possible to the researchers.
To fulfill the requirements we implemented a shared filesystem that is about 500TB total in storage and is based on lustre with total redundancy all from electricity to fibre and network. We also optimized the server avaibility by interconnect the servers into a cluster to become calculation powerhouse instead of dedicated resource for specific tasks. (Together we now have about 2+TB ram and 200+Cpu cores available for calculations and simulations)
Together with this we are responsible for the backup of all the servers, not only the local but also servers running external solutions, like LIMS and report systems. This is a great task due to the extreme amount of data but is manageable and works quite well. The Tivoli Server Manager from IBM is OK for this type of jobs as long as you optimize the performance of the backups.
When not checking the servers, logs and other daily duties I develop the web platforms for ANDIS, ANDIU and DIREVA. All three are based on the same consept, but differs in data types and content. It contains account manager, log system, revision archive system, a web based registration tool used by the clinical staff all around in Skåne, Uppsala and Vasa (in Finland) where it handles phenotype registration (including family history). In-house it contains a "mini" LIMS to keep track of what blood samples arrived from where, who and when to the researchers and project managers. It keeps track of the results for GADa, CPEPTID, DnaExt, Glucose and HBA1C where applied. It also contains a logic genetic data storage solution for SNP results and above all this a logic pattern match system to classify the patients.
I also work with virtual machines, Linux os configurations, clusters, general PHP and BASH script and web solutions when requested.
$if ($readagain is true); then goto line 1
Last updated: November 13, 2013
Website contact: LUDC webteam