HP on Cloud Computing


Published on

This is a presentation Christian Verstraete, CTO of cloud computing for HP, gave to attendees at PLM Connection Europe. See http://www.siemens.com/plm/blog for related blog post and more event coverage.

Published in: Technology
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

HP on Cloud Computing

  1. 1. ©2010 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice1 Christian Verstraete Chief Technologist – EB Cloud Solutions, HP Siemens PLM Connection, 20th October 2010 Cloud Computing In the Product Development Space
  2. 2. 2 Footer Goes Here2 – Definition of Cloud Computing • Public Cloud • Private Cloud • Community Cloud – Cloud and PD&E • High Performance Computing & Cloud • PLM as a service – HP Strategy – Conclusions Agenda
  3. 3. 3 Footer Goes Here3 Definition of Cloud Computing
  4. 4. 4 Footer Goes Here4 Cloud Definition – US Federal Government Definition Key Characteristics – On-demand self-service – Ubiquitous network access – Resource pooling • Location independence • Homogeneity – Rapid elasticity – Measured service – Software as a Service, an environment where users can run predefined applications directly from their web browser SaaS PaaS – Platform as a Service, an environment in which the user is provided with a rich environment in which he can run his applications as long as they are programmed in one of the languages supported by the platform (Java, Python or .Net) LaaS – Infrastructure as a Service, an environment that provides the user with processing power, networking, storage and the other necessary resources allowing him to run his software and applications – Cloud computing is a model for enabling available, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model promotes availability and is comprised of five key characteristics, three delivery models, and four deployment models. Private Cloud Community Cloud – Sold to the public, mega-scale infrastructure Public Cloud – Composition of two or more clouds Hybrid Cloud – Enterprise owned or leased – Shared infrastructure for specific community Source: Peter Mell & Tim Grance – National Institute of Standards and Technology – Information Technology Lab
  5. 5. 5 Footer Goes Here5 Public Cloud The pro’s and con’s
  6. 6. 6 Footer Goes Here6 Differences Between Datacenter and Cloud – Known Location(s) – 100s –1000s of nodes – Hardware resiliency – Physical/Virtual resources – Clusters/grids – Static  flexible – Shared storage (NAS …) – Management of Complexity – Unknown Location(s) – 10,000+ standardized nodes – Software resiliency – Virtual resources – Cloud – Elastic – Replicated storage (SAN) – Management Standardization Datacenter scale Cloud scale
  7. 7. 7 Footer Goes Here7 However, Keep in Mind – Security – Privacy/Compliance – Licensing – The Cloud Supply Chain – Lack of Standardization – Internet Bandwidth – Service Level Agreements – Actual Cloud Pricing
  8. 8. 8 Footer Goes Here8 Private Cloud Evolution of the Data Center
  9. 9. 9 Footer Goes Here9 The Evolution of the Data Center Provisioning & Workload Scheduling Cloud Computing (2010 – 2020) Cloud based Infrastructure Utility/Grid based Environment Business Services Cloud – Network Centric Virtual Computing (2005 – 2015) Virtual Servers, Storage & Networking Virtual Environment Virtual IT Services Virtualized – Software Centric Modular Computing (2000 – 2010) Blade Servers, Storage & Networking Operating Environment Applications Physical – Hardware Centric Virtualization
  10. 10. 10 Footer Goes Here10 Integrating private & public cloud Cloud Bursting Workload Management User Access Portal IT Process Automation (HP OO) Virtual + Physical Server Provisioning (HP SA) Network and Storage provisioning (HP BSA Suite,…) Load Balancing Optimization Policy Enforcement Fault Tolerance Adaptive Resource Orchestration Resource Manager Plug-Ins (iLO, BareMetal, batch, etc.) Inventory/Monitoring Hardware Config., Health, Workload, etc. Administrator Portraits Dynamic Compute Environment Workload Orchestrator Public Cloud – Jump to Public Cloud when available capacity is insufficient – Manage “public” workload in line with established policies
  11. 11. 11 Footer Goes Here11 Community Cloud Collaboration across the ecosystem
  12. 12. 12 Footer Goes Here12 Community Cloud Source: Peter Mell & Tim Grance – National Institute of Standards and Technology – Information Technology Lab Cloud Partner 1 Partner 2 Partner 3 Partner 4 Partner 5 Partner 6 Partner 7 OEM The cloud infrastructure is shared by several organizations and supports a specific community that has shared concerns (e.g. mission, security requirements, policy, and compliance considerations). It may be managed by the organizations or a third party and may exist on premise or off premise.
  13. 13. 13 Footer Goes Here13 Recall functional view Example: Gs1 product recall – Recall notification form completed (including attachments) – Information standardized – Existing risk management procedures remain in place Manufacturer Initiates Recall Retailer Receives Recall Notification Immediately Complements & enhances existing product recall processes by providing users: – Accurate & Complete info – In a Standard format – In Real Time – From a Single Trusted Source – Developed by industry for industry – Recall notification received with standard information set – Serves as basis to drive the physical recall process
  14. 14. 14 Footer Goes Here14 Cloud For Product Development & Engineering
  15. 15. 15 Footer Goes Here15 Different Environments at Different Stages Gate 1 Stage 1 Gate 2 Stage 2 Gate 3 Stage 3 Gate 4 Stage 4 Gate 5 Stage 5 Scoping Build Business Case Development Testing & Validation Launch Idea Screening Second Screen Go To Development Go To Testing Go To Launch Ideation Tools Process & Portfolio Mgmt. Mechanical Design. Electronic Design. Software Design. Testing Tools. Product Lifecycle Management Simulation
  16. 16. 16 Footer Goes Here16 R&D: Multiple Aspects Need to be Addressed Grid Remote Graphics Large Compute Capacity Delivered to small amount of users Complex imagery to be displayed quickly to users Simulation, digital testing, animation, etc. Cloud Compute Capacity Delivered to large amount of users PLM, Software Development, Case Tools etc. CAD, CAM, CAE, Animation, etc.
  17. 17. 17 Footer Goes Here17 Put the data close to the user or limit what needs to be transmitted The Visualization Dilemma Replication Servers Remote Graphics – Remote Utilities in which copies of data files are maintained, while synchronization happens in the background – Requires remote locations for such servers – Only transmit the image pixels – Requires blade workstations in the cloud – May need more complex provisioning logic
  18. 18. 18 Footer Goes Here18 High Performance Computing In the Cloud?
  19. 19. 19 Footer Goes Here19 Cloud & High Performance Computing Linking Resources to form infrastructure, harnessing unused ones. Ability to scale up and down with little need to know anything about the underlying environment. Tries to retrieve the highest performance from a grid of computers to address a cumbersome problem Cloud Grid High Performance Computing
  20. 20. 20 Footer Goes Here20 Grid, schedulers, compilersserver, storage, OS, network IPS IPS+ – Server: x86-64 (single, dual, quad core), Itanium – Interconnect: GigE, Infiniband – Storage: HPC, EVA, MSA home nodes – OS: Redhat Linux, Suse Linux, Windows, HP-UX – Network: SSH, VPN – Grid/Scheduler: Platform, United Devices, Datasynapse, Altair, Digipede, Nice – Compiler: Intel FSI, CAE, Life Science, Oil & Gas, EDA APS – LSTC – Schlumberger – Codefarm – Axway – Kognitio Server & Storage Utility HP Flexible Computing Services Management Additional Services – Hands and Eyes – Data Transport Service – Customer Furnished Equipment – Software Depot – Implementation – Software Installation – Availability Monitoring – User Management – Server Management – Storage Management – Network Management Access – Time Synchronization (NTP)– Mail Transport Service (SMTP) – Distributed Name Lookup (DNS) Storage – Data Protection (Core Backup/Restore) – High Performance Storage – Data Protection (Backup Integration) – Networked Storage (SAN & NAS) – Networked Storage (Local Replication) – Networked Storage (Remote Replication) Server – High-Speed Interconnects (Infiniband, Myrinet) – Virtual Machines – Additional Home Nodes – Added Memory Network – Point-to-Point WAN Connectivity – Multiple Redundant/Independent Networks – Additional External IP Address – Additional Externally Accessible Nodes – Additional VPN Capacity
  21. 21. 21 Footer Goes Here21 A leading supplier of oil field services and application software to the oil & gas industry Schlumberger Challenge Solution Results – Major oil and gas exploration and production companies around the world, customers for the company’s Eclipse Reservoir Simulation software, have a problem with too few compute cycles to run critical simulations fast enough in pursuing new oil reserves – HP Flexible Computing Services enabled Schlumberger to provide a complete environment for their major oil and gas customers to perform reservoir modeling, as a utility service • XC stack, LSF, and Eclipse on DL585s with Infiniband – Schlumberger was able to provide instant capacity to their Eclipse customers who otherwise could not have run critical Eclipse simulations – Schlumberger’s customer was able to do the equivalent of 6 years work (doing it the old way), in under 3 months – Schlumberger is extending the service to more customers, creating a new business for this service, and new uses for their application software The HP difference – Rapid delivery of resources – Latest technology for faster runs – Easier to contract with
  22. 22. 22 Footer Goes Here22 Automated Platform Provisioning Dynamic Development Environment Bundles/Components OS/Software Images Red Hat AS3 W2K3 Ent. Server 1. Team lead requests development and QA environments (client & server) 2. Rapid ADM selects appropriate host servers and unused VMs MS Office MS Visual Studio .NET Adobe Studio BEA WebLogic SQL Server 2000 Oracle 9i Host Farm RADM 4. Team lead is notified via email with system names and login credentials Development Client J2EE Application Dev Server Oracle DB Server J2EE Application QA Server 3. Servers are provisioned with requested platforms 5. Application team starts using new Rapid ADM environments VMs
  23. 23. 23 Footer Goes Here23 PLM as a Service
  24. 24. 24 Footer Goes Here24 PLM as a Service PLM Environment – Standardized SPLM Solution – Validated to OEM Standards – Fast to Deploy IT Estate – Standardized Apps Set – Set integrations – Fast to Deploy Delivery – Service Orientated Delivery – Recognized SLA’s – Lowest TCO PLM as a Service – Cost effective, standardized Supplier on-boarding solution – Preconfigured, pretested/certified solution that can be tailored to corporate- specific requirements – Cost effective upfront pricing including service/support SLA’s – Standardized sales-delivery model between Siemens PLM, HP and OEM
  25. 25. 25 Footer Goes Here25 HP’s Strategy For Cloud Computing
  26. 26. 26 Footer Goes Here26 HP Strategy Enable cloud providers Enable IT as THE business service provider Deliver cloud services from HP
  27. 27. 27 Footer Goes Here27 Conclusion
  28. 28. 28 Footer Goes Here28 Enterprise-class IT services CHAS M DEDICATED STANDARDIZED 1:1 binding workloads: machines IT “islands” VIRTUALIZED OPTIMIZED Virtualized Infrastructure: shared infrastructure improves utilization CLOUD BURSTING SERVICE SOURCING Agility: ability to source services from anywhere AUTOMATED SOA/SOI SERVICE ENABLED Automation: lower mgmt cost; eliminate human error $X/10 Global –class Public Cloud Major Application Changes IT Transformation vs. Cloud Computing CHASM IT Transformation Minor Application Changes Private Cloud/Internal Cloud “Pooled” resources / shared infrastructure Increasing levels of Automation Public Cloud Services Technology evolution IT Capability
  29. 29. 29 Footer Goes Here29 Calculating the Cloud Determining the true cost of hosting servers in the cloud Decision Tree 1. Sensitive Data Does the server house, transmit or processes sensitive data such as: – Personally identifiable information – Protected health information – Institutional or Personal Financial Records – Academic Records – Sensitive Intellectual Property or Research Data 2. Mission Critical Apps Is the server running a mission-critical application or service? 3. Uploads Required Does the server require large and frequent uploads 4. OS not Supported Does the server require an operating system not supported by the selected cloud provider? 5. Licensing Issues Are there software licensing issues which could prevent a server from running in the Cloud? – USB Dongle – Licensing manager IP restrictions 6. Retention Policies Are there data retention policies which require onsite and/or long-term data archives for the server 8. Connection Required Does the server need to be connected to peripheral devices or a Storage Area Network 7. Access Required Is physical access to the server required? Yes Server Costs – Server Memory Requirements – Server Processing Requirements – In-house Standalone Server – In-house Virtual Servers – Cloud charges for server instance Storage Costs – Amount of Storage Needed – Cost of in-house vs. Cloud storage – Cost for out-of-band uploads to the cloud – Cost for In-house backups vs. In the Cloud Bandwidth Costs – Costs of bandwidth per GB (In-house and Cloud) – Server bandwidth usage external/on-campus – Network Latency, in-house and Cloud Risks – Cloud provider employees access your server – Limited auditing capabilities – Minimal network security controls – Limited control of address and namesplace – Internet accessible web-based console to control server instances No Cloud Metrics Reconsider Consider
  30. 30. 30 Footer Goes Here30 Key Takeaways Cloud covers a number of independent concepts PD&E can take advantage of cloud but may require specific cloud features and address licensing issues HP should be your business partner in your cloud ventures
  31. 31. 31 Footer Goes Here Outcomes that matter. Mail: christian.verstraete@hp.com Blog: www.hp.com/blogs/manufacturing-distribution Twitter: @christianve