Cloud Computing and Virtualization with Globus Oakland, May 2008 Kate Keahey (keahey@mcs. anl. gov) Tim Freeman (tfreeman@mcs. anl. gov) University of Chicago Argonne National Laboratory
Cloud Computing Tutorial Hands-on l To participate in the hands-on part of the tutorial, send your PKI X 509 subject line to nimbus@mcs. anl. gov l The first 10 requests will be given access to the nimbus cloud l Hurry! 05/14/08 Virtual Workspaces: http//workspace. globus. org
Overview l l l l l 05/14/08 Motivation The Workspace Ecosystem: Abstractions and Background The Workspace Deployment Tools Managing Resources with Virtual Workspaces Appliance management and contextualization Virtual Cluster Management with Workspace Tools Application Example: the STAR experiment Cloud Computing Run on the cloud: hands-on tutorial Virtual Workspaces: http//workspace. globus. org
Motivation
A Good Workspace is Hard to Find ? 05/14/08 1) Configuration: finding environment tailored to my application 2) Leasing: negotiating a resource allocation tailored to my needs Virtual Workspaces: http//workspace. globus. org
Consumer’s Perspective: Quality of Life l Real life applications are complex u l … and require complex, customized environments u l Rely heavily on the right combination of compiler versions and available libraries Environment validation u 05/14/08 STAR example: Developed over more than 10 years, by more than 100 scientists, comprises ~2 M lines of C++ and Fortran code To ensure reproducibility and result uniformity across environments Virtual Workspaces: http//workspace. globus. org
Consumer’s Perspective: Quality of Service l There is life beyond submitting batch jobs u l Control of resources u l Explicit SLA: different sites offer different quality of service Satisfying peak demand u 05/14/08 Resource leases rather than job submission Experiment season, paper deadlines, etc. Virtual Workspaces: http//workspace. globus. org
Provider’s Perspective l Providing resources is easy, providing environments is hard u l Fine-tuning environments for different communities is expensive u u u l 05/14/08 User comment: “I have 512 nodes I cannot use” ; -) Evaluating, installing and maintaining software packages etc. Reconciling conflicts Coordinating update schedules for different communities is a nightmare It may be hard to justify configuring/dedicating resources if they are only needed 1% of the time - even if the 1% is very important for one of your users Virtual Workspaces: http//workspace. globus. org
The Workspace Ecosystem: Abstractions and Background
Virtual Workspaces l A dynamically provisioned environment u u u l Appliances/virtual appliances u u 05/14/08 A complete environment: a complete (software) environment as required by our community or applications provisioned on demand. Resource allocation: provision the resources the workspace needs (CPUs, memory, disk, bandwidth, availability), allowing for dynamic renegotiation to reflect changing requirements and conditions. Deployment point of view A complete environment that can be packaged in various formats Packaging point of view Virtual Workspaces: http//workspace. globus. org
Workspace Implementations Traditional tools l u u Base environment (discovery) Virtual machines u u Typically long deployment time u Isolation Complete environment l Automated configuration l u Contextualization Short deployment time Very good isolation Runtime performance impact Performance isolation Runtime environment Paper: “Virtual Workspaces: Achieving Quality of Service and Quality of Life in the Grid” 05/14/08 Virtual Workspaces: http//workspace. globus. org
The Virtues of Virtualization App App App Guest OS (Linux) Guest OS (Net. BSD) Guest OS (Windows) VM VM VM Parallels Xen VMWare Virtual Machine Monitor (VMM) / Hypervisor UML Hardware KVM etc. l l l 05/14/08 Bring your environment with you Excellent enforcement and isolation Fast to deploy, enables short-term leasing Have a performance impact but it is acceptable for most modern hypervisors Suspend/resume, migration Virtual Workspaces: http//workspace. globus. org
Creating a Virtual Cluster that Works Create a functioning virtual ensemble Put the VMs in context Contextualization layer Deploy VMs onto the resource VM VM Obtain a lease on a raw resource Resource Deploy virtual machines 05/14/08 VM Virtual Workspaces: http//workspace. globus. org
The Workspace Ecosystem Appliance Deployment: Mapping environments onto leased computing resources Coordinating creation of virtual resources A mix of open source software and proprietary tools communicating via common protocols Resource Providers: Grid providers: Tera. Grid, OSG, etc. Commercial providers: EC 2, Sun, etc. Appliance Providers: off-the-shelf environment bundles certified/endorsed for safety leverage appliance software commercial and open “marketplaces” 05/14/08 Virtual Workspaces: http//workspace. globus. org
Roles and Responsibilities l Division of labor u u Virtual organizations provide appliances u l Resource providers provide resources Middleware that maps appliances onto resources Appliance management software u u l Appliance creation, maintenance, validation, etc. Not an appliance provider Shifting the work around u 05/14/08 Into the hands of the parties most motivated and qualified to do it Virtual Workspaces: http//workspace. globus. org
Workspace Deployment Tools
Virtual Workspaces: Vital Stats l Virtual Workspace software allows an authorized client to dynamically deploy and manage workspaces u l Currently implements workspaces as Xen VMs u l l l KVM coming this summer Also, contextualization layer Globus incubator project Started ~2003, first release in September 2005 Current release 1. 3. 1 (March ‘ 08) Download it from: u 05/14/08 Virtual Workspace Service (VWS), workspace control, Context Broker http: //workspace. globus. org Virtual Workspaces: http//workspace. globus. org
Using Workspaces (Deployment) Pool node VWS Service Pool node Pool node Pool node Workspace -Workspace metadata -Pointer to the image -Logistics information -Deployment request -CPU, memory, node count, etc. 05/14/08 Virtual Workspaces: http//workspace. globus. org
Using Workspaces (Interaction) The workspace service publishes information on each workspace as standard WSRF Resource Properties. Pool node Users can interact directly with their workspaces the same way the would with a physical machine. 05/14/08 Pool node Pool node Users can query those properties to find out information about their workspace (e. g. what IP the workspace was bound to) Pool node VWS Service Pool node Trusted Computing Base (TCB) Virtual Workspaces: http//workspace. globus. org
Workspace Service (what sits inside) Workspace WSRF front-end that allows clients to deploy and manage virtual workspaces VWS Service Pool node Pool node Pool node Workspace back-end: Resource manager for a pool of physical nodes Deploys and manages Workspaces on the nodes Each node must have a VMM (Xen) installed, as well as the workspace control program that manages individual nodes Contextualization creates a common context for a virtual cluster 05/14/08 Trusted Computing Base (TCB) Virtual Workspaces: http//workspace. globus. org
Workspace Service Components l GT 4 WSRF front-end u u Leverages GT core and services, notifications, security, etc. Roughly follows the OGF WS-Agreement provisioning model l Workspace Service back-end u u l Works with multiple Resource Managers Workspace Control for on the node functions Contextualization u 05/14/08 Lease descriptions Publishes available lease terms Put the virtual appliance in its deployment context Virtual Workspaces: http//workspace. globus. org
Managing Resources with Virtual Workspaces
Workspace Back-Ends l Default resource manager (basic slot fitting) u u l “datacenter technology” equivalent Used for OSG Edge Services project Challenge: finding Xen-enabled resources u Amazon Elastic Compute Cloud (EC 2) l u u u l Solution: develop a back-end to EC 2 Grid credential admission -> EC 2 charging model Address contextualization needs Challenge: integrating VMs into current provisioning models u 05/14/08 Software similar to Workspace Service (no virtual clusters, contextualization, fine-grain allocations, etc. ) Solution: gliding in VMs with the Workspace Pilot Virtual Workspaces: http//workspace. globus. org
The Workspace Pilot l l Challenge: how can I provide a “cloud” using virtualization without disrupting the current operation of my cluster? Flying Low: the Workspace Pilot u u l Deployment u u u 05/14/08 Integrates with popular LRMs (such as PBS) Implements “best effort” leases Glidein approach: submits a “pilot” program that claims a resource slot Includes administrator tools Testing @ U of Victoria (Atlas), Ian Gable and collaborators Adapting for the use of the Atlas experiment @ CERN, Omer Khalid Tera. Port (small partition) Virtual Workspaces: http//workspace. globus. org
Workspace Pilot in Action Level 2: provision VMs VWS Level 1: provision raw resources VM VM Xen dom 0 VM VM LRM/PBS Xen dom 0 05/14/08 Virtual Workspaces: http//workspace. globus. org
The Pilot Program l l Uses Xen balloon driver to reduce/restore domain 0 memory so that guest domains (VMs) can be deployed Secure VM deployment u u l The pilot requires sudo privilege and thus can be used only with site administrator’s approval The workspace service provides fine-grained authorization for all requests Signal handling u SIGTERM: pilot exceeded its allotted time l l Default policy: one VM per physical node Available for download u u 05/14/08 Notifies VWS, allows it to clean up After a configurable time period takes things into its hands. Workspace Release 1. 3. 1: http: //workspace. globus. org/downloads/index. html Virtual Workspaces: http//workspace. globus. org
Workspace Control l VM control u l Integrating a VM into the network u u u l l 05/14/08 Assigning MAC addresses and IP addresses DHCP delivery tool Building up a trusted (non-spoofable) networking layer VM image propagation Image management and reconstruction u l Starting, stopping, pausing, etc. creating blank partitions, sharing partitions Contextualization information management Talks to the workspace service via ssh Can be used as a standalone component Virtual Workspaces: http//workspace. globus. org
Workspace Back-Ends l Default resource manager (basic slot fitting) u u l “datacenter technology” equivalent Used for OSG Edge Services project Challenge: finding Xen-enabled resources u Amazon Elastic Compute Cloud (EC 2) l u u u l Solution: gliding in VMs with the Workspace Pilot Long-term solutions Leasing model with explicit terms u u 05/14/08 Solution: develop a back-end to EC 2 Grid credential admission -> EC 2 charging model Address contextualization needs Challenge: integrating VMs into current provisioning models u l Software similar to Workspace Service (no virtual clusters, contextualization, finegrain allocations, etc. ) Semantically rich leases: advance reservations, urgent leases, renegotiable leases, etc. Cost-effective lease semantics Virtual Workspaces: http//workspace. globus. org
Appliance Management and Contextualization
Where Do Appliances Come From? Marketplaces (VMWare, EC 2, Workspace …) Appliance Provider (a user, a VO, a Grid…) appliance description Good… but: maintenance? ease of use? formats? 05/14/08 Virtual Workspaces: http//workspace. globus. org
Where Do Appliances Come From? Marketplaces Appliance Management Software (VMWare, EC 2, Workspace …) (OSFarm, r. Path, …)) Xen appliance description VMware CDROM Appliance Provider (a user, a VO, a Grid…) Better 05/14/08 Virtual Workspaces: http//workspace. globus. org
Deploying Appliances l Appliances need to be “portable” u l Making the appliance contextaware: u u u l So that they can be reused in many contexts VM VM Other appliances Site-specific information (e. g. a DNS server) User/group/VO/Grid-specific information (e. g. public keys, host certs, gridmapfiles, etc. ) VM VM site Security issues u u 05/14/08 Who do I trust to provide legitimate context information? How do I make sure that appliances adhere to my site policies? Virtual Organization Virtual Workspaces: http//workspace. globus. org
Where Do Appliances Come From? Marketplaces (VMWare, EC 2, Workspace …) Appliance Management Software (OSFarm, r. Path, Cohesive. FT…)) appliance description Xen VMware CDROM appliance assertions appliance contextualization Appliance Provider (a user, a VO, a Grid…) 05/14/08 Virtual Workspaces: http//workspace. globus. org
Creating Virtual Clusters with Workspace Tools
Make Me a Working Cluster l You got some VMs and you’ve deployed them… Now What? u l What network are they connected to? Do they actually represent something useful? (like a ready -to-use OSG cluster? ) Do the VMs know about each other? Can they share some disk? How do they integrate into the site storage/account system? Do they have host certificates? And a gridmapfile? And all the other things that will integrate them into my VO? Challenge: what is a virtual cluster? u A more complex virtual machine l u u l Networking, shared storage, etc. Available at the same time and sharing a common context Example: an OSG cluster Solutions u u u Ensemble management Exporting and sharing common context Sophisticated networking configurations. Paper: “Virtual Clusters for Grid Communities”, CCGrid 2006 05/14/08 Virtual Workspaces: http//workspace. globus. org
Contextualization Challenge: Putting a VM in the deployment context of the Grid, site, and other VMs l u Assigning and sharing IP addresses, name resolution, application-level configuration, etc. Solution: Management of Common Context l u contextualization agent Common Context IP hostname pk Configuration-dependent l u u provides&requires Common understanding between the image “vendor” and deployer Mechanisms for securely delivering the required information to images across different implementations Paper: “A Scalable Approach To Deploying And Managing Appliances”, Tera. Grid conference 2007 05/14/08 Virtual Workspaces: http//workspace. globus. org
Contextualizing Appliances Context Broker Appliance Provider appliance context Appliance context agent Appliance context template appliance context Appliance Deployer generic context disk image application-specific context agents appliance content Resource Provider 05/14/08 Virtual Workspaces: http//workspace. globus. org
Application Example: Virtualization with the STAR experiment
Virtual Workspaces for STAR l STAR image configuration u l Using the workspace service over EC 2 to provision resources u u 05/14/08 A virtual cluster composed of one OSG headnode and multiple STAR worker nodes Allocations of up to 100 nodes Dynamically contextualized for out-of-the-box cluster Virtual Workspaces: http//workspace. globus. org
Virtual Workspaces for STAR l Deployment stages: u u Deploy the virtual machines u l Create an “ensemble” defining the virtual cluster Contextualize to provide an out-of the-box cluster Contextualization: u u l Cluster applications: NFS & PBS Grid information: gridmapfile and host certificates Runs u u 05/14/08 Using VWS on the nimbus cloud for small node allocations (VWS + default + Context Broker) Using VWS with EC 2 backend for allocations of ~100 nodes (VWS + EC 2 backend + Context Broker) Virtual Workspaces: http//workspace. globus. org
with thanks to Jerome Lauret and Doug Olson of the STAR project Running jobs : : : 94 Running jobs 142 Running jobs 124 109 42 73 0 Running jobs : 150 230 VWS/EC 2 Running jobs : : : 76 Running jobs 282 Running jobs 243 140 195 221 0 Running jobs : 300 PDSF WSU Running jobs : : : 96 Running jobs 195 Running jobs 183 136 152 37 54 0 Running jobs : 200 150 Fermi Job Completion : 05/14/08 BNL File Recovery : Virtual Workspaces: http//workspace. globus. org Running jobs : : 50 Running jobs 39 15 21 27 34 0 9 Running jobs : 42 50
with thanks to Jerome Lauret and Doug Olson of STAR project thanks to Jerome Lauret and Doug Olson of the STAR project Nersc PDSF EC 2 (via Workspace Service) WSU 05/14/08 Accelerated display of a workflow job state Y = job number, X = job state Virtual Workspaces: http//workspace. globus. org
Cloud Computing
The Workspace Cloud Client l We took the workspace client and made it easy to use u u l Allows scientists to lease VMs roughly following Amazon’s EC 2 model (simplified) u l u 05/14/08 PKI X 509 credentials and quotas instead of payment The goal is to restore/evolve this functionality as user requests come in u l Narrowing down the functionality Wrapper on top of the workspace client Saving VMs, network configurations In the future: richer leases, etc. “Cloudkit” coming out in next release, due soon Virtual Workspaces: http//workspace. globus. org
Nimbus @ University of Chicago l Objectives u u u l Vital Stats u u u l Deployed on 16 nodes of Tera. Port cluster @ UC Powered by the workspace set of tools Image management handled via grid. FTP Made available mid-March ‘ 08 http: //workspace. globus. org/clouds/ To obtain access mail nimbus@mcs. anl. gov u 05/14/08 Make it easy for scientific community to experiment with this mode of resource provisioning Learn about the requirements of scientific projects and evolve the infrastructure Features, SLAs, security and sharing concerns, etc. Available to scientific, educational projects, open source testing, etc. Virtual Workspaces: http//workspace. globus. org
Science Clouds l A group of clouds making resources available “on the nimbus model” u u l Nimbus, Stratus@UFL (Mauricio Tsugawa), FZK in Germany (almost done, Lizhe Wang), others expressed interest EC 2 Some differences in setup, policies u UFL requires private networks (using Open. VPN) l u l EC 2 requires payment Cloud federation u Moving an app from a hardware platform to a cloud is relatively hard l u u 05/14/08 Currently you’d use the same credential for the cloud and for the virtual private network Need image, learn new paradigm, etc. Moving between clouds is relatively easy … if you have “rough consensus” on interfaces, image formats, etc. Virtual Workspaces: http//workspace. globus. org
Who runs on the clouds and what do they do? 05/14/08 Virtual Workspaces: http//workspace. globus. org
Related Projects l Portal development (Josh Boverhof, LBNL) l Workspace KVM backend (Michael Fenn, Clemson University) l Integration with the Nebula project (University of Madrid) 05/14/08 Virtual Workspaces: http//workspace. globus. org
Let’s get on the cloud!
Parting Thoughts 05/14/08 Virtual Workspaces: http//workspace. globus. org
Parting Thoughts l Come and run on science clouds l Not just cloud computing u l A bunch of technologies have to come together to make cloud computing widespread The way we do computing is changing u u 05/14/08 Today we build horseless carriages Tomorrow we might do things differently Virtual Workspaces: http//workspace. globus. org
Credits l Workspace team: u l Guest appearances u l Ian Foster, Frank Siebenlist With thanks to many collaborators: u 05/14/08 Kate&Tim Jerome Lauret (STAR, BNL), Doug Olson (STAR, LBNL), Marty Wesley (r. Path), Stu Gott (r. Path), Ken Van Dine (r. Path), Predrag Buncic (Alice, CERN), Haavard Bjerke (CERN), Rick Bradshaw (Bcfg 2, ANL), Narayan Desai (Bcfg 2, ANL), Duncan Penfold-Brown (Atlas, uvic), Ian Gable (Atlas, uvic), David Grundy (Atlas, uvic), Ti Legget (University of Chicago), Greg Cross (University of Chicago), Lizhe Wang (FZK), Marcel Kunze (FZK), Mauricio Tsugawa (UFL), Jose Fortes (UFL), Renato Figueiredo (UFL), Omer Khalid (CERN), Artem Harutyunyan (CERN), Mike Fenn (U of Clemson), Sebastien Goasguen (U of Clemson), Josh Boverhof (LBNL), Leve Hajdu (STAR, BNL), Lidia Didenko (STAR, BNL), David Bartle (Atlas, uvic), Lee Liming (ANL), Frank Wuerthwein (OSG, SDSC), Abhishek Rana (OSG, SDSC), Jeff Chase (Duke), and many others. Virtual Workspaces: http//workspace. globus. org
Sponsors l NSF SDCI “Missing Links” l NSF CSR “Virtual Playgrounds” l Tera. Grid l DOE Sci. DAC CEDPS 05/14/08 Virtual Workspaces: http//workspace. globus. org
Скачать презентацию Cloud Computing and Virtualization with Globus Oakland May
09c6c14c56fd8db44c535070ca7d6869.ppt