WNoDeS – a Grid/Cloud Integration Framework Elisabetta Ronchieri (INFN-CNAF) for the WNoDeS Project

Slides:



Advertisements
Similar presentations
ITEC474 INTRODUCTION.
Advertisements

CREAM-CE status and evolution plans Paolo Andreetto, Sara Bertocco, Alvise Dorigo, Eric Frizziero, Alessio Gianelle, Massimo Sgaravatto, Lisa Zangrando.
1 Software & Grid Middleware for Tier 2 Centers Rob Gardner Indiana University DOE/NSF Review of U.S. ATLAS and CMS Computing Projects Brookhaven National.
S. Gadomski, "ATLAS computing in Geneva", journee de reflexion, 14 Sept ATLAS computing in Geneva Szymon Gadomski description of the hardware the.
Workload Management Workpackage Massimo Sgaravatto INFN Padova.
DataGrid Kimmo Soikkeli Ilkka Sormunen. What is DataGrid? DataGrid is a project that aims to enable access to geographically distributed computing power.
Tier-1 experience with provisioning virtualised worker nodes on demand Andrew Lahiff, Ian Collier STFC Rutherford Appleton Laboratory, Harwell Oxford,
Minerva Infrastructure Meeting – October 04, 2011.
The SAM-Grid Fabric Services Gabriele Garzoglio (for the SAM-Grid team) Computing Division Fermilab.
Pilots 2.0: DIRAC pilots for all the skies Federico Stagni, A.McNab, C.Luzzi, A.Tsaregorodtsev On behalf of the DIRAC consortium and the LHCb collaboration.
Copyright © 2010 Platform Computing Corporation. All Rights Reserved.1 The CERN Cloud Computing Project William Lu, Ph.D. Platform Computing.
 Cloud computing  Workflow  Workflow lifecycle  Workflow design  Workflow tools : xcp, eucalyptus, open nebula.
1 Evolution of OSG to support virtualization and multi-core applications (Perspective of a Condor Guy) Dan Bradley University of Wisconsin Workshop on.
:: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :: GridKA School 2009 MPI on Grids 1 MPI On Grids September 3 rd, GridKA School 2009.
Grid Workload Management & Condor Massimo Sgaravatto INFN Padova.
GILDA testbed GILDA Certification Authority GILDA Certification Authority User Support and Training Services in IGI IGI Site Administrators IGI Users IGI.
Grid Computing at The Hartford Condor Week 2008 Robert Nordlund
WNoDeS – Worker Nodes on Demand Service on EMI2 WNoDeS – Worker Nodes on Demand Service on EMI2 Local batch jobs can be run on both real and virtual execution.
And Tier 3 monitoring Tier 3 Ivan Kadochnikov LIT JINR
Issues Autonomic operation (fault tolerance) Minimize interference to applications Hardware support for new operating systems Resource management (global.
NA-MIC National Alliance for Medical Image Computing UCSD: Engineering Core 2 Portal and Grid Infrastructure.
Ruth Pordes November 2004TeraGrid GIG Site Review1 TeraGrid and Open Science Grid Ruth Pordes, Fermilab representing the Open Science.
Grid User Interface for ATLAS & LHCb A more recent UK mini production used input data stored on RAL’s tape server, the requirements in JDL and the IC Resource.
Conference name Company name INFSOM-RI Speaker name The ETICS Job management architecture EGEE ‘08 Istanbul, September 25 th 2008 Valerio Venturi.
6/23/2005 R. GARDNER OSG Baseline Services 1 OSG Baseline Services In my talk I’d like to discuss two questions:  What capabilities are we aiming for.
EGEE-II INFSO-RI Enabling Grids for E-sciencE EGEE Site Architecture Resource Center Deployment Considerations MIMOS EGEE Tutorial.
International Symposium on Grid Computing (ISGC-07), Taipei - March 26-29, 2007 Of 16 1 A Novel Grid Resource Broker Cum Meta Scheduler - Asvija B System.
Testing and integrating the WLCG/EGEE middleware in the LHC computing Simone Campana, Alessandro Di Girolamo, Elisa Lanciotti, Nicolò Magini, Patricia.
3/12/2013Computer Engg, IIT(BHU)1 CLOUD COMPUTING-1.
Workload management, virtualisation, clouds & multicore Andrew Lahiff.
DIRAC Project A.Tsaregorodtsev (CPPM) on behalf of the LHCb DIRAC team A Community Grid Solution The DIRAC (Distributed Infrastructure with Remote Agent.
EGI Technical Forum Amsterdam, 16 September 2010 Sylvain Reynaud.
Automatic testing and certification procedure for IGI products in the EMI era and beyond Sara Bertocco INFN Padova on behalf of IGI Release Team EGI Community.
CLOUD COMPUTING WHAT IS CLOUD COMPUTING?  Cloud Computing, also known as ‘on-demand computing’, is a kind of Internet-based computing,
OpenNebula: Experience at SZTAKI Peter Kacsuk, Sandor Acs, Mark Gergely, Jozsef Kovacs MTA SZTAKI EGI CF Helsinki.
Information Initiative Center, Hokkaido University North 11, West 5, Sapporo , Japan Tel, Fax: Management.
Tutorial on Science Gateways, Roma, Catania Science Gateway Framework Motivations, architecture, features Riccardo Rotondo.
INFN OCCI implementation on Grid Infrastructure Michele Orrù INFN-CNAF OGF27, 13/10/ M.Orrù (INFN-CNAF) INFN OCCI implementation on Grid Infrastructure.
Breaking the frontiers of the Grid R. Graciani EGI TF 2012.
Computing Facilities CERN IT Department CH-1211 Geneva 23 Switzerland t CF Cluman: Advanced Cluster Management for Large-scale Infrastructures.
Use cases and a practical example for Grid and Cloud integration Elisabetta Ronchieri, INFN CNAF EPIKH School, Beijing,
II EGEE conference Den Haag November, ROC-CIC status in Italy
INTRODUCTION TO GRID & CLOUD COMPUTING U. Jhashuva 1 Asst. Professor Dept. of CSE.
Claudio Grandi INFN Bologna Virtual Pools for Interactive Analysis and Software Development through an Integrated Cloud Environment Claudio Grandi (INFN.
DIRAC for Grid and Cloud Dr. Víctor Méndez Muñoz (for DIRAC Project) LHCb Tier 1 Liaison at PIC EGI User Community Board, October 31st, 2013.
INFN/IGI contributions Federated Clouds Task Force F2F meeting November 24, 2011, Amsterdam.
G. Russo, D. Del Prete, S. Pardi Kick Off Meeting - Isola d'Elba, 2011 May 29th–June 01th A proposal for distributed computing monitoring for SuperB G.
HLRmon Enrico Fattibene INFN-CNAF 1EGI-TF Lyon, France19-23 September 2011.
Farming Andrea Chierici CNAF Review Current situation.
WP5 – Infrastructure Operations Test and Production Infrastructures StratusLab kick-off meeting June 2010, Orsay, France GRNET.
EMI is partially funded by the European Commission under Grant Agreement RI Elisabetta Ronchieri, INFN CNAF Munich, 29 March 2012 WNoDeS Tutorial.
Building on virtualization capabilities for ExTENCI Carol Song and Preston Smith Rosen Center for Advanced Computing Purdue University ExTENCI Kickoff.
EGI-InSPIRE RI EGI-InSPIRE EGI-InSPIRE RI EGI Services for Distributed e-Infrastructure Access Tiziana Ferrari on behalf.
The EPIKH Project (Exchange Programme to advance e-Infrastructure Know-How) gLite Grid Introduction Salma Saber Electronic.
BaBar & Grid Eleonora Luppi for the BaBarGrid Group TB GRID Bologna 15 febbraio 2005.
Accessing the VI-SEEM infrastructure
Dynamic Extension of the INFN Tier-1 on external resources
Introduction to Cloud Technology
Workload Management Workpackage
Design rationale and status of the org.glite.overlay component
GWE Core Grid Wizard Enterprise (
StratusLab Final Periodic Review
StratusLab Final Periodic Review
Grid Computing.
Ákos Frohner EGEE'08 September 2008
Introduction to Cloud Computing
GGF15 – Grids and Network Virtualization
20409A 7: Installing and Configuring System Center 2012 R2 Virtual Machine Manager Module 7 Installing and Configuring System Center 2012 R2 Virtual.
Wide Area Workload Management Work Package DATAGRID project
Cloud Computing: Concepts
Presentation transcript:

WNoDeS – a Grid/Cloud Integration Framework Elisabetta Ronchieri (INFN-CNAF) for the WNoDeS Project

Overview Problem Statement WNoDeS as solution New features and why WNoDeS and IGI Details 2E. Ronchieri, EGI TF, September 2011, Lyon, France

Overview Problem Statement WNoDeS as solution New features and why WNoDeS and IGI Details 3E. Ronchieri, EGI TF, September 2011, Lyon, France

Requirements we started with Large computing centers used for a wide range of applications by multiple user communities – To determine proper and timely configuration of compute resources without sacrificing efficiency, flexibility and security User Communities – To allow accessing resources by using a number of different interfaces, authenticating via several methods, exploiting existing local and distributed infrastructures 4E. Ronchieri, EGI TF, September 2011, Lyon, France

In short To define a multi-layer framework to optimize the usage of computing center resources for multiple communities … starting from the INFN CNAF center (located in Bologna, Italy) that currently hosts about 10,000 computing cores, and 9 PB of disk space and 10 PB of tape space each day hosts about 50,000 jobs supports tens of international experiments of astro-particle physics (e.g., AMS2, Argo, Auger, Fermi/Glast, Magic, Pamela, Virgo) with about 15% of its computing resources supports the Italian Tier-1 for CERN-based LHC experiments of high energy physics (i.e, ATLAS, CMS, LHCb, Alice) and Tier-0/1 for several others adopts Platform LSF as batch system and IBM GPFS as shared File System 5E. Ronchieri, EGI TF, September 2011, Lyon, France

Overview Problem Statement WNoDeS as solution New features and why WNoDeS and IGI Details 6E. Ronchieri, EGI TF, September 2011, Lyon, France

The Worker Nodes on Demand Service, a.k.a. WNoDeS, in one slide Is a software framework created by INFN to integrate Grid and Cloud provisioning through virtualization – All resources (Grid, Cloud, or else) are taken from a common pool Is scalable and reliable – it is in production at several Italian centers, including the INFN Tier-1 (CNAF, Bologna) since November 2009 – Currently WNoDeS is managing about 2000 on‐demand Virtual Machines (VMs) there Is totally transparent for Grid users and for users of traditional Computing Centers batch systems Supports a native Cloud interface – OCCI (Open Cloud Computing Interface) compliant – A Cloud Web portal Integrates authentication, policy and accounting It is not required to convert an entire farm to WNoDeS: it can coexist and share resources with a traditional computing cluster not using virtual machines 7E. Ronchieri, EGI TF, September 2011, Lyon, France

Key Characteristics Use of Linux KVM to virtualize resources on-demand that are available and customized for – direct job submissions by local users [1-2] – Grid job submissions (with direct support for the EMI CREAM-CE and WMS components) [1-2] – instantiation of Cloud resources [6] – instantiation of Virtual Interactive Pools (VIP) [3] Use of LRMS (a “batch system software”) to handle VM scheduling – There is no need to develop special (and possibly unscalable, inefficient) resource brokering systems – The LRMS is totally invisible to users for e.g. Cloud instantiations Not Use of “Cloud over Grid” (or “Grid over Cloud”)’s concept to handle resources – WNoDeS simply uses all resources and dynamically presents them to users as users want to see and access them 8E. Ronchieri, EGI TF, September 2011, Lyon, France

Key Advantages Use of a common pool of resources – There is no need to dedicate resources to “user interfaces”, “Grid computing”, “Cloud computing”, “local users” – All types of resources are taken from a common pool, resulting in overall better utilization of resources Integrated support of old and new use cases (local access, Grid computing, Cloud computing) Re-use of ten years of worldwide development, expertise and resources brought about by Grid Computing applied to the key areas of Authentication, Authorization, Accounting, Information Systems, Brokering – This will make it possible for example to inter-connect Clouds without starting from scratch in these areas again Flexibility and scalability – At the core of WNoDeS there is a standard batch system used for resource provisioning and policing a mature, stable piece of software found in any sizeable resource center – There is no need to rewrite this key and complex part 9E. Ronchieri, EGI TF, September 2011, Lyon, France

Requirements Collected on the base of the needs of INFN users and resource providers to virtualize systems and services 10 Published in the document ‘WNoDES: Requirements for Virtualized Services’ [7] – It is a continuous work in progress due to the dynamic nature of requirements and the constant evolution of WNoDeS E. Ronchieri, EGI TF, September 2011, Lyon, France

Req. #1: Jobs submitted by users of a traditional computing center Users log on to a frontend system of a resource center and submit jobs to a computing farm 11 WNoDeS can associate jobs belonging to a user (or to a set of users) to VMs specifically created for them and customized to match the job requirements [1,2] – This is completely transparent for users – There is no need to change anything from the users’ point of view – The jobs will be executed on dynamically-created VMs Non WNoDeS WNs WN Standard job WN WNoDeS enabled WNs Virtual Machin e E. Ronchieri, EGI TF, September 2011, Lyon, France

Req. #2: Jobs submitted by users of a Grid-based distributed infrastructure WNoDeS handles these jobs as an extension of the Req. #1 simply taking into account special Grid requirements if present [1,2] There are two possibilities here: – All jobs beloning to certain Virtual Organizations (VOs) can be directed to pre‐packaged VMs This is completely transparent for users – Grid users can insert a special request into their JDL script for example to specify which VM they want their jobs to run on This is done by using standard EMI (European Middleware Initiative) gLite job management tools 12E. Ronchieri, EGI TF, September 2011, Lyon, France

Req. #3: Self-allocation of systems by users of a traditional computing center Local users need direct and local access (or interactive access) to the batch system or the distributed infrastructure without requiring root access. WNoDeS provides a mode called Virtual Interactive Pool (VIP) [3] – Systems are provisioned so that users can log on with their local account – Users can specify characteristics of VM such as machine type, number of CPUs, amout of RAM, local filesystems to be mounted – This is a kind of cloud computing applied to a traditional computing center designed to efficiently offer new services, without incurring the overhead to dedicate resources for this purpose – A caching mechanism is available to reduce provisioning time 13E. Ronchieri, EGI TF, September 2011, Lyon, France

Req. #4: Self-allocation of compute resources by distributed users (Cloud Computing) Distribute users like to autonomously self- allocate compute resources defined by several parameters WNoDeS can deliver Cloud services via [6] – A standard API called the Open Cloud Computing Interface (OCCI), developed by the Open Grid Forum (OGF) Rarely employed directly by users – A Cloud Web portal A user-friendly way to self-allocate resources 14E. Ronchieri, EGI TF, September 2011, Lyon, France

Req. #5: Resource access by distributed users through multiple authentication methods Users need to be authenticated whatever is the users’ authentication method WNoDeS can use any authentication method as long as it can be mapped into a known and unique resource ownership [2] – Ownership can be shared across different resource interfaces such as Grid and Cloud – Non-Grid users will be authenticated as well using Kerberos or Shibboleth by the IGI Web Portal [6] 15E. Ronchieri, EGI TF, September 2011, Lyon, France

Status WNoDeS 1 is in production at the INFN Tier-1 and at other INFN sites in Italy – Focusing on providing virtualization services for local and Grid accesses WNoDeS 2 is planned for release in November 2011 – Introducing (among other things) the Cloud portal, the VIP interface and support for multiple batch systems WNoDeS 2 installation and configuration also works with Torque and Maui as Open Source Batch System (tested at INFN Bari) – Using properties for queue, jobs and WN WNoDeS 2 configuration also works with Lustre as Open Source Shared File System to serve data and distribute Virtual Node images (tested at INFN Bari) – Observed good performance and scalability 16E. Ronchieri, EGI TF, September 2011, Lyon, France

Developments Program focuses in particular on: – Dynamic network virtualization – VM image abstraction – Efficient access to large storage systems Gateway NFS/GPFS – Inter-cloud connectivity 17E. Ronchieri, EGI TF, September 2011, Lyon, France

Overview Problem Statement WNoDeS as solution New features and why WNoDeS and IGI Details 18E. Ronchieri, EGI TF, September 2011, Lyon, France

WNoDeS Cloud Web Interface 19E. Ronchieri, EGI TF, September 2011, Lyon, France

Process Flow for the WNoDeS Web-Based Cloud Management 20E. Ronchieri, EGI TF, September 2011, Lyon, France

Mixed Mode We are investigating the possibility to use the same physical WN to execute both real jobs and VM jobs – This will increase the flexibility of the farm and allow better optimization of the resource usage There is no need to “reserve” WN to WNoDeS 21E. Ronchieri, EGI TF, September 2011, Lyon, France

Mixed Mode Old WNoDeS New WNoDeS Batch Server Manager WN Virtual Machine Standard job WNoDeS jobs & Standard jobs WN WNoDeS enabled WNs Virtual Machine Non WNoDeS WNs Batch Server Manager WNoDeS jobs Standard jobs WN Standard job 22E. Ronchieri, EGI TF, September 2011, Lyon, France

Software WNoDeS has been ported into ETICS during summer 23 – To handle daily builds via local remote build in order to get repository RPMs – To automate its deployment via WNoDeS metapackages – To adopt ETICS quality checks – To automate internal test procedures E. Ronchieri, EGI TF, September 2011, Lyon, France

Overview Problem Statement WNoDeS as solution New features and why WNoDeS and IGI Details 24E. Ronchieri, EGI TF, September 2011, Lyon, France

WNoDeS and IGI IGI, the Italian Grid Infrastructure, plans to offer an integrated computing infrastructure providing – Grid computing services – Cloud computing services – Inter-connection of Cloud computing resource centres for the Italian and research community The WNoDeS framework is the production solution for integrating Grid and Cloud infrastructure offered by IGI 25E. Ronchieri, EGI TF, September 2011, Lyon, France

Overview Problem Statement WNoDeS as solution New features and why WNoDeS and IGI Details 26E. Ronchieri, EGI TF, September 2011, Lyon, France

Information For further information and questions – WNoDeS mailing list, The WNoDeS Web site – Acknoledgements for this talk – D.Salomoni (INFN CNAF), A. Italiano (INFN CNAF), G. Dalla Torre (INFN CNAF), D. Andreotti (INFN CNAF), G. Donvito (INFN Bari), V. Spinoso (INFN Bari) 27E. Ronchieri, EGI TF, September 2011, Lyon, France

References [1] A. Italiano, et al, WNoDeS, a Tool for Integrated Grid/Cloud Aaccess and Computing Farm Virtualization, CHEP 2010, Taipei [2] D. Salomoni, et al, An Authentication Gateway for Integrated Grid and Cloud Access, CHEP 2010, Taipei [3] C. Grandi, et al, Virtual Pools for Interactive Analysis and Software Development through an Integrated Cloud Environment, CHEP 2010, Taipei [4] D. Salomoni, A. K. Calabrese Melcarne, A. Chierici, G. Dalla Torre, A. Italiano, “Performance improvements in a large scale virtualization system”, PoS(ISGC 2011 & OGF 31)049 [5] D. Salomoni, D. Andreotti, L. Cestari, G. Potena, P. Solagna, “A Web-based portal to access and manage WNoDeS Virtualized Cloud resources”, PoS(ISGC 2011 & OGF 31)054 [6] M. Bencivenni, et al, “A portal for an easy access to the IGI grid infrastructure”, poster 9 at EGI TF, September 2011, Lyon, France [7] D. Salomoni, E. Ronchieri, “WNoDeS: Requirements for Virtualized Services”, 28E. Ronchieri, EGI TF, September 2011, Lyon, France

Thanks! 29E. Ronchieri, EGI TF, September 2011, Lyon, France

Backup Slides 30E. Ronchieri, EGI TF, September 2011, Lyon, France

Measurements Using Lustre shared file-system to distribute Virtual Node images determines: – Good performance Tested already at 3-5GByte/s on a farm with 700TB of storage – Scalability Already tested with thousands of client and hundreds of file servers 31E. Ronchieri, EGI TF, September 2011, Lyon, France