Program Day 1

Er astronomi FAIR?

Indenfor astronomien er der en lang tradition for åben og international deling af forskningsdata og resultater. Dette fokus er vokset endnu mere inden for de seneste år, men har egentlig været praktiseret i årtier. Så er astronomien FAIR, uden måske helt at vide det?

Jeg vil bl.a. give eksempler fra NASA missionerne Kepler og TESS, som vi på Aarhus Universitet har været dybt involveret i. Disse sattellitmissioner blev opsendt med formålet at finde planeter omkring andre stjerner (exoplaneter) og studere svingninger i stjernerne. Men det viser sig at data også kan bruges til en masse andet videnskab - så måske er en åben data politik en god idé!

MeluXina - building-up a modern world-class supercomputer

LuxProvide is the High Performance Computing Centre of the Grand Duchy of Luxembourg and operates the MeluXina supercomputer, one of the first systems in the EuroHPC network. Designed to tackle a large variety of complex workloads, MeluXina has been built with modularity and efficiency in mind, arriving in 2021 as the greenest supercomputer in the EU, 4th in the world, and 36th most powerful. This talk will highlight key lessons learned from the planning phase to the daily operation of the facility.

BPF (Berkeley Packet Filter): Safely observe and extend your operating system

Have you ever been frustrated by the time it takes to get a new feature into your operating system? Does the OS kernel seem like an inscrutable black box? With BPF, this no longer has to be the case! BPF is a new technology pioneered by the Linux kernel which allows safely injecting arbitrary code into the kernel to extend functionality and observe the running system, while preserving high performance. In this talk, we'll give an overview of the technology, and dive into how this capability can be leveraged to implement novel functionality without having to change the kernel itself, with a special focus on networking use cases.

IBM Quantum – hvordan ser kvante computer landskabet ud nu og i fremtiden og hvor er vi på vej hen – hvorfor er det vigtigt at kende til kvante computer også i HPC sammenhænge?

Kom og hør IBM’s CTO omkring Quantum Computing og hvilke nye problematikker vi er i stand til at løse og hvordan IBM ser Data Management i HPC sammenhænge.

IBM Quantum Computing

• Kvante computer i HPC miljøer
• Status hvor er vi i dag indenfor Quantum Computing

IBM Data Management

• Data indsigt og tilgængelighed
• Intelligent Data tiering

Hvordan arbejder Sundhedsdatastyrelsens med Sundhedsdata og sikkerhed

Få svar på spørgsmål som:

• Hvordan samarbejder sundhedssektoren omkring arbejdet med cyber- og informationssikkerhed?
• Hvilke værktøjer bruger DCIS sund og sektoren til at samarbejdet?
• Hvilke samarbejdsformer har vist sig at være gode?
• Hvad er næste skridt for samarbejdet?

SSH Open Marketplace – View into SSH research practice

The Social Sciences and Humanities Open Marketplace - marketplace.sshopencloud.eu  - is a discovery portal which pools and contextualises resources for Social Sciences and Humanities research communities: tools, services, training materials, datasets, publications and workflows. The Marketplace highlights and showcases solutions and research practices for every step of the SSH research data life cycle.

The SSH Open Marketplace was built as part of the SSHOC project and acts as a thematic entry door into the EOSC. The discovery service is now funded and maintained by three European Research Infrastructures Consortia - DARIAH, CLARIN and CESSDA - as part of the SSH Open Cluster.

This talk presents how the SSH Open Marketplace can provide insights into the use of tools, methods and standards in the SSH research community, and how it can increase serendipity in the discovery of new methods and standards, by interlinking the records and describing workflows.

FAIR data and software training – TU Delft experiences

TU Delft Library published its vision for Research Data & Software management training in 2019. The training programme outlined in this document aims at enabling TU Delft researchers to work more efficiently and reproducibly, and following the FAIR principles, with the research data and code of their projects. In this presentation, the speaker will share her experiences and lessons learned when creating, setting up and running some of the more relevant courses within this vision.

CodeRefinery - a hub for FAIR Software practices

CodeRefinery is a project which supports students and researchers across all academic disciplines by advancing the FAIRness of research software and development practices so that they can collaboratively develop, review, discuss, test, share, and reuse their codes.  This initiative teaches the software engineering practices which are needed for modern collaborative computational research but are often missed in a traditional university curriculum.

Thus far, over 2000 students and researchers have been trained.  The project currently focuses on the Nordic/Baltic countries, but aims to expand beyond this region. CodeRefinery aims to operate as a community project with support from academic organizations.

In this contribution we will share our experiences and lessons learned from developing and evolving the lesson curriculum and course format, from organizing over 50 events and from growing a community of instructors and research software engineers with the goal of improving software engineering practices and thus the quality of computational research and engineering. We will summarize which development tools and practices we have found to be essential and offer recommendations on where to start and how to progress.

Running AI/ML in container on LUMI

Hvordan kan man på en nem og sikker måde kører AI/ML på LUMI ved at bruge f.eks tensorflow elle pythorch. Demo af container-builder.

Sundhedsdatastyrelsens erfaringer med DCIS

Få svar på spørgsmål som:

• Hvad er en DCIS og hvad laver man i DCISSund?
• Hvilket ansvar har DCIS og de andre myndigheder?
• Hvad var det første skridt?
• Hvad er en ”operativ kapacitet”?
• Hvad har vi lært og hvor mener vi at vi er gode nu?
• Hvad ville vi have gjort anderledes?

Trends Currently Shaping the Future of HPC

In this session we look at Microsoft’s view of what is currently shaping the future of HPC. We will discuss the following topics:

• Artificial Intelligence/Machine Learning in HPC
• HPC as a Service
• Data Storage and Data Management
• Cloud (and security)

ISO27001 - i NGC (Nationalt Genom Center)

NGC har etableret et ledelsessystem for informationssikkerhed, og er i Juli måned 2022 blevet certificeret efter standarderne ISO 27001 Digital informationssikkerhed og ISO 27701 for privatlivsbeskyttelse – Der er efter NGC's oplysninger kun 3-4 offentlige organisationer i Danmark som er certificerede efter begge standarder. Forud for det er gået et meget stort og komplekst arbejde med at etablere em sikkerhedsmodel, som omfatter hele organisationen, alle relationer og hele HPC platformen.

ssh-CA - a simple, yet elegant method of linking ssh-login to the home institution's login

Det er muligt (og ikke særligt kompliceret) at koble login med ssh til login på brugerens hjeminstitution (via WAYF). Kom og se det spændende proof-of-concept og hør om perspektiverne.

Rigsarkivets arbejde i forhold til FAIR langtidsbevaring og tilgængeliggørelse af forskningsdata

Der redegøres i oplægget kort for Rigsarkivets formål og arbejde med FAIR-principperne, herunder også i fht. forskningsdata. Dernæst behandles selve langtidsbevaringen- og tilgængeliggørelsen af data, og der listes en række grundlæggende udfordringer i forbindelse med dette.

Context, approach, and progress of the EOSC-A Long Term Data Preservation Task Force

Trustworthy long-term preservation repositories ensure continued access to the data and metadata necessary to deliver a sustainable EOSC. Trust in scientific outputs and processes depends on the ability to reproduce, replicate and re-use scientific results depends on the long-term. The FAIRness of the underlying data depends on FAIR-enabling repositories. The EOSC Strategic Research and Innovation Agenda (SRIA) underlines the importance of long-term data preservation, but an explicit strategy has not been formulated.

The Long-Term Data Preservation Task Force (LTDP-TF) will provide recommendations on the vision and sustainable implementation of long-term data preservation policies and practices, as well as suggestions for strategy execution. The roles, responsibilities and accountability of the preservation and related actors within EOSC, including the financial aspects and possible business models all form part of the task force scope. The LTDP-TF will also provide recommendations on engagement through a European Trusted Digital Repository (TDR) network to define, evaluate and increase the maturity of preservation and FAIR enabling practices across disciplines and geographies. 

The presentation will highlight the context, activities and ongoing issues surrounding the EOSC-A Long Term Data Preservation Task Force.

Joint Communication and Computing, a path towards sustainable Internet of Things

Internet of Things (IoT) has shown great potential for different verticals, whereas it has also imposed many critical technical challenges. The ever-increasing, massive IoT data not only stress communication bandwidth and storage capacity, but also cause excessive energy consumption. Furthermore, the cost to extract the values and knowledge from, and provide end-to-end information security for, vast amount of data becomes non-trivial. IoT ecosystem calls for novel, sustainable solutions for sensor data acquisition, transmission, computing and storage. Conventionally, communication and computing are two theoretical disciplines that have developed independently. For example, data compression and analytics are usually designed independently. In this talk, a holistic design of data compression and analytics, and a joint data compression and encryption solution will be introduced as examples. The research outcomes have shown interesting opportunities to make IoT ecosystem more efficient and sustainable. 

Brug af Microsoft Azure i det lokale HPC anlæg

HPC regnecentre blive mere og mere udbredte, og lige som med generelle IT-systemer er det interessant at udforske, hvad "skyen" kan bidrage med omkring HPC.  På DTU Fysik kører vi et pilot-projekt med at integrere VM-servere i Azure skyen som compute-noder i Linux clusteret "Niflheim".

Med en VPN-tunnel har vi koblet Niflheim netværksmæssigt tæt sammen med Azure, og Slurm batch-køsystemet allokerer nu VM'er i Azure efter behov.  Azure har mulighed for NFS Blob storage af brugernes data i dette system.

Projekt 5 status

Brug AIOps til at forbedre de studerendes oplevelse af netværket

Brug AIOps til at forbedre de studerendes oplevelse af netværket BYOD, IoT enheder, online forelæsninger, sikkerhed m.m. stiller stadig stigende krav til det underliggende netværk. Antallet af IT-ressourcer følger sjældent med i samme tempo. Derfor skal infrastrukturen simplificeres og AIOps skal implementeres for at sikre, at både de studerende og alle andre klienter på netværket har en god oplevelse.

LUMI overview. Status and demos.

Hvor er vi henne i forhold til LUMI og hvad kan vi forvente? Vi DEMOer forskellige ting og snakker lidt om status.

Store, manage, and curate your unstructured data – anywhere and forever

• No more platform lock-in or complexity – scale your storage with flexibility and simplicity
• Manage your data at the edge, in the data center, or in the cloud platform of your choice
• Unlock real-time data awareness, predictive cloud monitoring, and industry leading customer support