CloudBroker Platform Presentation Wibke Sudholt CloudBroker GmbH Technoparkstrasse 1, CH-8005 Zurich, Switzerland Phone: +41 44 633 79 34 Email: wibke.sudholt@cloudbroker.com Web: http://www.cloudbroker.com SCI-BUS is supported by the FP7 Capacities Programme under contract no. RI-283481 CloudBroker Platform • Web application store for the deployment and execution of scientific and technical software in the cloud • On demand, pay per use, browser / programmatic / command-line access, cross-domain • Uses infrastructure as a service (IaaS) from cloud providers • Offers platform as a service (PaaS) for software vendors and software as a service (SaaS) for end users • Easy to use, speeds up time to market, no need for own high performance computing infrastructure, OpEx instead of CapEx © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 2 Platform Architecture End Users, Software Vendors User Tools REST Web Service API Biology Applications Health Applications Engineering Applications CloudBroker Platform Amazon Cloud © CloudBroker GmbH All rights reserved. IBM Cloud … Applications CloudBroker Integration Chemistry Applications Java Client Library CLI Web Browser UI … Cloud SCI-BUS Summer School 2012 05.07.2012 / 3 Platform Functionality Process Manager Users Tools Web Browser UI Web Service API Application Manager Process Monitor User Manager Queuing System Accounting Module Resource Manager Storage Manager Billing Module Payment Module Image Manager Scalability and Fault Tolerance Handler Cloud Provider Access Manager Amazon Adapter IBM Adapter … Adapter Security Frame: Transport Layer Security, Access Rights Security Amazon Cloud © CloudBroker GmbH All rights reserved. IBM Cloud … Cloud SCI-BUS Summer School 2012 05.07.2012 / 4 Platform Security Customer Client Browser or Application Corporate IT Corporate Security Policies and Standards SSL Secured Connection Authentication CloudBroker CloudBroker CBP Platform . Industry Standard Application Security Technology Industry Standard Server Security Technology Industry Standard Secure Data Center SSL Secured Connection SSL secured connection Authentication to Cloud Authentication to VM Cloud Provider © CloudBroker GmbH All rights reserved. Cloud CBP Instances Dedicated, Secured . and Restricted Virtual Machines Security Certified Compute and Storage Cloud Technology Security Certified Data Center SCI-BUS Summer School 2012 05.07.2012 / 5 User Management • User roles: admin, advanced, standard • Software visibility: private, protected, public © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 6 Typical Job Submission Lifecycle 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Prepayment (user) or monthly invoice (platform) Software and resource selection and job creation (user) Data file upload (user) to cloud storage (platform) Job submission (user) Compute instance startup or reuse (platform) Data file upload from cloud storage to master node (platform) Computations on worker nodes (platform, application) Data file download from master node to cloud storage (platform) Compute instance shutdown or reuse (platform) Data file download (user) from cloud storage (platform) Accounting and billing (platform) © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 7 Dynamic Cluster in the Cloud • Communication to and between nodes via SSH • Instance information in dynamically generated node file (MPICH / MPICH2 / Open MPI format, more formats can be added upon request) • File sharing via NFS can be enabled © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 8 Application Requirements Item Requirements / possibilities Type Scientific and technical applications Domain All License Open source, free, commercial Character Compute-intensive, batch-oriented Interface Command line, non-interactive Runtime Hours, days, weeks Operating system Linux Parallelization Single-threaded / single-core, multi-threaded / multi-core, embarrassingly parallel / high-throughput, parallel across nodes (via MPI or other methods) Workflows Single jobs, pipeline jobs, parameter sweeps (via API) © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 9 Ported Applications Application Domain Remarks BLAST DNA and protein sequence alignment Available in public platform AutoDock Protein-ligand docking Available in public platform Gromacs Molecular dynamics simulations Available in public platform X! Tandem Mass spectrometry data matching Available in public platform OpenFOAM Computational fluid dynamics Available in public platform Rosetta Protein modelling Only with own license TransAT Computational fluid dynamics Only with own license Siesta Material science Only with own license T-Coffee DNA and protein sequence alignment FoldX Protein modelling R Mathematics and statistics More applications continuously to be added Applications can also be added by users © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 10 Typical Software Deployment Lifecycle 1. 2. 3. 4. Generation of installation script (vendor) Creation and configuration of software (vendor) Adding of executables (vendor) Adding of installation package (vendor), including installation script and software package, if needed 5. Deploying of software (vendor) 6. Checking of software and installation script (manager) 7. Generation of deployments (manager) 8. Running of special deployment jobs (platform) 9. Checking and activation of deployments (manager) 10. Activation of software (manager/vendor) 11. Publication of software (vendor) 12. Running of jobs (user) © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 11 Clients and APIs CLI • Command line interface • Job submission Java • Application programming interface • Java client library REST • Representational state transfer • Web service interface © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 12 Public Platform • https://platform.cloudbroker.com • First preview version published in fall 2010, current version 1.0.4 • Available clouds: – Amazon EC2 and S3 (computing and storage) – IBM SmartCloud Enterprise (only computing currently, storage upcoming) • Available applications: – BLAST, AutoDock, Gromacs, X! Tandem, OpenFOAM – Additional applications privately or upon request • Software checking before deployment • Standard pricing: – Cloud resource costs + software costs + CloudBroker surcharges • Production version, monitored and maintained • Help desk under platform@cloudbroker.com • Also available as hosted and in-house version © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 13 Protein Modelling Example Case • Setup: – Collaboration between IBM, CloudBroker and the Institute of Molecular Systems Biology (IMSB) at ETH Zurich – Goal is to better understand the mechanisms of infectious diseases – For the example of the Streptococcus bacterium, predict the 3D structures of important proteins from their 1D sequences by means of the Rosetta software – Use the CloudBroker Platform to manage the calculations on top of IBM SmartCloud Enterprise compute and storage resources © CloudBroker GmbH All rights reserved. • Results: – 249 target proteins modeled – More than 1’000 CPUs used in parallel – 36’000 jobs automatically managed – Almost 250‘000 CPU hours consumed – 2.3 Mio 3D protein structure models created – Less than two weeks runtime SCI-BUS Summer School 2012 05.07.2012 / 14 Further Examples • ASCOMP TransAT (http://www.ascomp.ch) – Computational fluid dynamics – Collaboration with HSR – Microsoft Silverlight-based web browser UI • KNIME (http://www.knime.org) – Konstanz Information Miner – Workflow framework – Computational chemistry example case • SCI-BUS (http://www.sci-bus.eu) – SCIentific gateway Based User Support – EU FP7 project – 11+6 user communities from different domains © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 15 CloudBroker in SCI-BUS © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 16 Goals in SCI-BUS • Provide connection to open and commercial cloud infrastructures for the scientific gateways • Allow scientific gateways to charge for their services via the CloudBroker Platform • Integrate CloudBroker Platform as new distributed computing infrastructure (DCI) in WS-PGRADE / gUSE © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 17 Features to Be Implemented • Already available: – Development and testing platform for SCI-BUS – Extended CloudBroker Platform REST and Java APIs – Integration of CloudBroker Platform in WS-PGRADE / gUSE under DCI Bridge, security and job settings – Possibility to configure and run CloudBroker jobs from within WS-PGRADE / gUSE workflows • In preparation: – Commercial components in WS-PGRADE / gUSE for user mapping, billing and cost display – Possibility to register own cloud resources in the CloudBroker Platform – Adapters to open cloud infrastructures (OpenStack, OpenNebula, Eucalyptus) © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 18 Acknowledgements • CloudBroker GmbH: – Maxim Malgin, Ann Mikhailova • ScaleTools AG: – Nicola Fantini, Andrey Sereda • MTA SZTAKI: – Peter Kacsuk, Zoltan Farkas, Akos Hajnal • … and many others © CloudBroker GmbH All rights reserved. SCI-BUS Summer School 2012 05.07.2012 / 19 Thank you! Questions? Please add yourself to the tutorial user list! SCI-BUS is supported by the FP7 Capacities Programme under contract no. RI-283481