SURFsara Data Services

Transcription

1 SURFsara Data Services SUPPORTING DATA-INTENSIVE SCIENCES Axel Berg

2 Dutch Scientific Challenges From High Energy Physics to atomic and molecular physics (DNA); Life sciences (cell biology); Human interaction (all human sciences from linguistics to even phobia studies); and from the big bang; to astronomy; science of the solar system; earth (climate and geophysics); into life and biodiversity. Picture courtesy of UvA FNWI

3 High Energy Physics, discovery of Higgs boson SURFsara and NIKHEF are a tier 1 site for the Large Hadron Collider at CERN

4 LOFAR, discovery of two new pulsars Using computing resources provided by SURFsara, which is part of the Dutch and European Grid Infrastructure, Coenen and the team needed only a month to search through a set of LOFAR images that would have occupied a single computer for more than a century.

5 twinl: searching NL tweets Prof. Antal van den Bosch Dr. Erik Tjong Kim Sang

6 twinl usage

7 THE DATA EXPLOSION Production Cartesius Projects COSMO GRID (2009) RUMC (2013) OWLS (2006) ESSENCE (2006) ENTRAIN (2011) ITAMOC (2011) EAGLE (2013) TITAN (2014) HPC ARCHIVE (1993) LOFAR Volume 105TB 90TB 70TB 41TB 26TB 25TB 2,5TB (600TB) (1PB) 1172TB 12PB Community LOFAR LHC/ATLAS LHC/LHCb BBMRI MOLEPI LHC/ALICE ILDG DANS OTHERS Volume 5,8PB 3.9PB 1,4PB 115TB 78TB 74TB 43TB 18TB 50TB Start LHC

8 The world of the many Many different users (well organised (international) user communities, research groups, universities, research institutes, individual researchers, etc.) Many different use cases (central data, distributed data, small files, large files, static data, dynamic data, etc.) Many different user requirements (storage, meta data, data searching, persistent identifiers, long-term preservation, privacy, visualisation, data analytics, data sharing, data re-use, semantic annotation, work flows etc. etc. etc. etc.) Many different V s of Big Data (varieties, volumes, velocities, veracities, validities, volatilities)

9 CREATING DATA: designing, planning consent, collection and management, capturing and creating metadata CREATING DATA" RE-USING DATA: for follow-ups, new research, research reviews, scrutinising, teaching & learning RE-USING DATA" PROCESSING DATA" PROCESSING DATA: entering, transcribing, checking & validating, anonymising and describing TRUST ACCESS TO DATA: distributing, sharing, controlling access, promoting GIVING ACCESS TO DATA" ANALYSING DATA" ANALYSING DATA: interpreting, deriving, producing outputs & publishing, preparing for sharing Ref: UK Data Archive: PRESERVING DATA" PRESERVING DATA: migrating, backing-up, storing, creating metadata and documentation, archiving Research Data Life Cycle

10 CREATING DATA" Data ingest service PID Service Trusted Digital Repository Beehub B2SAFE (EUDAT) PID service Trusted Digital Repository B2SAFE (EUDAT) RE-USING DATA" GIVING ACCESS TO DATA" SURFsara services supporting the Research Data Life Cycle PROCESSING DATA" ANALYSING DATA" Grid Storage Beehub SURFdrive Grid Storage Beehub SURFdrive Hadoop HPC Cloud Remote Visualisation Cartesius, Lisa cluster Ref: UK Data Archive: PRESERVING DATA" Central data archive PID Service Trusted Digital Repository B2SAFE (EUDAT) Supporting the complete Research Data Life Cycle Collaboration with DANS Collaboration within EUDAT

11 SURFsara data strategy SURFsara is providing integrated ICT services (e.g. computing, visualisation, data storage, data analysis) for the scientific research community inthe Netherlands SURFsara is going to provide a Trusted Digital Repository to: - to enable long-term archiving of research data at SURFsara - improve the quality of research data with metadata and persistent identifiers enrich access with open standard protocols - improve search and discoverability via harvesting SURFsara is going to serve the Long Tail Data (e.g. community clouds) SURFsara is providing services for data analysis (Hadoop, HPC Cloud, Grid, remote visualisation) Collaborate with Universities, Institutes and research groups on data management issues (e.g. community clouds) and data management plans Engage with funding agencies and data owners to come to transparent models on preservation, policies, costs, security and privacy of research data and services

12 SURFsara (Big) Data Centric infrastructure services

13 SURFsara (Big) Data Centric infrastructure services

14 SURFsara (Big) Data Centric infrastructure services

15 SURFsara Data Storage Services Central Archive GRID SE BEEHUB Usage HPC, External GRID, HPC Cloud, Hadoop, Workflows, External Collaboration platform, Define your own access rules Preservation Medium, Long term Medium, Long term Short, Medium Media Disk + Tape Disk + Tape Disk Capacity (current) Bandwidth (aggregated) 240TB + 2.5PB 5PB + 20PB 80TB 1GB/s + 1GB/s >16GB/s + 2GB/s 1GB/s Latency Direct + High (50s) Low + High (100s) Low Objects Medium, Large (60M) Large (30M) Small, Medium Protocols NFS, GridFTP, (hpn-) SCP, Rsync SRM, GridFTP, HTTP, Webdav, NFS, Xrootd, dcap Webdav

16 SURFsara Data Services Data Ingest Service Easy way to upload large data from disk onto SURFsara facilities Upload data from 45 disks in parallel Used for archiving large sequencing datasets EPIC/Handle PID service Persistent references to physical locations to make data objects findable and citable PID is comparable to SBN of Books EPIC/Handle system is comparable to DNS for Data Objects EPIC consortium: Providing a sustainable service for storing an maintaining large volumes of PIDs

17 Trusted Digital Repository Data repository service to deposit data sets and objects Long-term preservation of research data Provides quality to data sets and objects via metadata descriptions Makes data sets and objects citable and findable in the future via Persistent Identifiers (EPIC PID service) Makes data sets discoverable via metadata harvesting Improves data access and re-usability of research data Ensures trust to researchers via regular auditing against standardized certifications (e.g. Data Seal of Approval, ISO16363, DIN 31644) Trusted Digital Repository service is currently under construction at SURFsara Collaboration with DANS and Universities to provide service to manage research data

18 Grid infrastructure and usage National Grid infrastructure (NGI_NL): - Operated in collaboration with Nikhef and RUG-CIT - Grid compute clusters at Nikhef, RUG-CIT and SURFsara - Grid Services Cluster at SURFsara - Life Science Grid (operated by SURFsara) - Grid Front-End storage at Nikhef and SURFsara (total capacity 5.2 PB) - Part of European Grid Infrastructure (EGI) Grid Usage: - In 2013 more than 53 million core hours (6050 core years) - Scientific projects/users: Number of e-infra projects: 34 - Number of users: 226

19 Life Science Grid sites LUMC Leiden WUR Wageningen NKI Amsterdam AMC Amsterdam Radboud Universiteit Nijmegen ErasmusMC Rotterdam Technische Universiteit - Delft Rijksuniversiteit Groningen Keygene Wageningen Each of these clusters has 128 cores and 40 Tb of storage space. SURFsara 3000 cores and large tiered storage (disk and tape) NIKHEF: 3300 cores, RUG-CIT: 1000 cores

20 Grid computing example UMCG, LUMC, EMC, VUMC and UMCU jointly work on the Genome of the Netherlands DNA sequencing data of 750 individuals, sequenced in China Information is stored and processed on the Life Science Grid at SURFsara

21 HPC Cloud You can clone your laptop or server Create an image and upload it into the cloud And expand it Largest VM goes up to 32 cores and 256 GB RAM Or clone it to multiple copies, create your own cluster Start multiple worker nodes when needed But remember: you are your own sys admin

22 HPC Cloud Current HPC Cloud infrastructure 30 HPC Nodes: 32 cores, 256 GB RAM, 1 TB local disk 10 "light User Service Nodes: 8 cores/node, 64 GB RAM, 6 TB local disk 1250 cores total 400TB shared storage (ISCSI, NFS, CIFS, CDMI...)

23 HPC Cloud usage num_of_users Number of cloud users actively starting VMs

24 HPC Cloud usage in core hours (2013) Red: reserved capacity based on users applications Blue: actual usage

25 Memory-intensive computing High memory compute nodes 2 TB RAM - 40 core machines 8TB disk storage 1 as part of the Life Science Grid 1 as part of the HPC cloud Service available Q Added value: Particular in Life Sciences there are still many applications which use few cores and scale in RAM. (e.g. de novo assembly of genomes)

26 Hadoop & nosql HADOOP can be used to process large datafiles in a distributed way. It comes with an easy to use programmer interface and is able to handle huge datafiles, in structured, semistructured or unstructured form. Hadoop infrastructure cores (4x capacity of 2012) - 2 PB of HDFS storage (2x capacity of 2012) Usage 2013: - 35 different users and user groups CPU-years of processing - 160TB data stored New: nosql service - Schemaless databases, not everything fits into rows and columns easily - Faster development, easy to scale

27 Cloud storage for research (Beehub) Capacity: 80 TB Nr of users: 544 users (69 use SURFconext) Usage: 21 TB in user folders, 2 TB in group folders, 23 TB stored Interface: webdav

28 Visualisation: SURF Collaboratorium

29

30 Remote visualisation infrastructure Remote visualisation, benefits for users: - Remote visualisation cluster with GPUs next to data in SURFsara datacenter - Visualisation application runs on remote high-end GPU cluster (possibly in parallel) - User accesses application through remote desktop (VNC) - No datasets, only pixels are transferred Remote Visualisation Service: - 9 render nodes, per node: 2x 6-core Intel Xeon E5-2620, 2x Nvidia Geforce GTX780Ti, 64 GB RAM, 800 GB scratch - 33 TB home file system - Software stack maintained by SURFsara: ParaView, VisIt, VTK, VMD, MIPAV, Blender, any other OpenGL application - Expert visualisation support

31 Remote visualisation SURFsara E.P. van der Poel & R.O. Monico - University of Twente - Rayleigh-Bénard convection simulation J. van der Dussen - TU Delft - Numerical simulation of stratocumulus cloud breakup - Institute for Marine and Atmospheric research Utrecht - High-resolution movie generation of oceanographic simulation - Streaming from RVS cluster to tiled display at IMAU

32 Personal cloud storage (SURFdrive) Trusted community cloud for personal storage Collaboration between SURFsara, SURFnet and Dutch universities Advantages: Trusted community solution for personal cloud storage Hosted, managed and served by community, data stored at SURFsara Login through SURFconext Specifications and service determined by end-users (universities) Initial capacity: 200 TB, 100 GB storage capacity per user Based on owncloud Operational: April 1 st, 2014

33 SURFsara national datacenter HPC services have datacenter requirements that are not commonly served by commercial datacenters: - High energy density per rack - Advanced cooling techniques - Relatively low redundancy European procurement resulted in a partnership agreement between SURFsara and TelecityGroup for the development of a national datacenter for higher education at Science Park Amsterdam New SURFsara national datacenter operational medio 2016 Datacenter in a datacenter 800m 2 and on-demand power 1.8MW Low PUE (1.2) à sustainable and low costs Available for national e-infrastructure services, for higher education and for SMEs

34 SURFsara provides user support on (nearly) all levels

35 SURF Outreach & Support (SOS) SURF s operating companies offer a world-class communication, computing and support infrastructure to facilitate scientific and scholarly research - data transfer - compute - (international) collaboration - visualisation - storage For these services we aim to provide a single point of contact for researchers and ICT staff SOS offers: - to raise awareness among researchers of the possibilities of SURF e- infrastructure for scientific research - to foster collaboration and to enable research projects

36 What SURF can do for researchers Activities for researchers and research support officers in 2014 SURF Roadshow visit universities and university medical centers Integrated services portfolio description of services and support at Central information point direct advice and support via

37 Questions? Axel Berg,