UberCloud: From Experiment to Marketplace
1 like938 views
UberCloud provides a platform for companies to build, deploy, and consume compute-intensive applications in the cloud, leveraging containerization to enhance usability and flexibility. The initiative, started in 2012, has involved numerous experiments with various cloud providers and experts, addressing challenges such as security and access to cloud resources. With a focus on transforming high-TCO software into accessible pay-per-use services, UberCloud aims to remove bottlenecks and improve the cloud experience for users.
UberCloud: From Experiment to Marketplace
- 1. Wolfgang Gentzsch & Burak Yenier The UberCloud Inc. The UberCloud 2012 -‐ 2015 From Experiment to Marketplace Our ‘last mile’: application containerization Recorded for the RichReport, January 2015
- 2. Introduction Ê The UberCloud enables companies to build, deploy, and consume compute and data intensive applications as a service: better, faster, and cheaper Ê Experiment started July 2012 to learn HPC Cloud challenges: now 166 experiments, 54 cloud providers, 80+ ISVs, 100s of experts, 2500 participants Ê Computing scenario: workstations, servers, and clouds Ê Pain: workstations limited, servers expensive, clouds new Ê Solution: Application containerization for UberCloud Marketplace
- 3. Sponsors , partners , customers
- 4. HPC in the Cloud: benefits Continue using your workstation (and your server) for your daily design, and use Cloud resources with additional benefits: Ê An HPC system at your finger tip, on demand, as you need it Ê Pay per use: no CAPital EXpenditure Ê Business flexibility: Scaling resources up and down Ê Low risk: working with multiple cloud providers Ê Latest technology: cloud providers upgrade continuously
- 5. HPC in the Cloud: challenges Ê It’s a new business and working paradigm Ê Cloud expertise needed Ê Access to clouds is cumbersome Ê Which cloud is right for me? Ê Loosing control over your jobs and data Ê Security, privacy, trust in service provider Ê Intellectual property Ê Software Licensing Ê Heavy data transfers
- 6. Solution: UberCloud SaaS Factory UberCloud SaaS Factory Software SaaS Commercial ISV Open Source In-‐house Public Cloud Private Cloud On-‐Premise $$ $ We turn high-‐TCO, complex software into user-‐friendly pay-‐per-‐use services: Development Services (Docker) -‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐ Deployment & Management Services -‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐ Marketplace Providers Consumers $ UberCloud SaaS Factory
- 7. Builder Launcher Controller ISV Data Tools Stackable units with tools (ex: encryp?on), ISV applica?on codes (ex: OpenFOAM). Just add your own codes and data. Run anywhere with UberCloud Run Time. Scale up or down the compute power as needed. Collect granular usage data, logs. Monitor, alert, report. On-‐Premise Private Cloud Public Cloud Run Time Run Time Run Time Removing the Bottlenecks with Containers: Build once, run anywhere
- 8. Builder Launcher Controller ISV Data Tools Stackable units with tools (ex: encryp?on), ISV applica?on codes (ex: OpenFOAM). Just add your own codes and data. Run anywhere with UberCloud Run Time. Scale up or down the compute power as needed. Collect granular usage data, logs. Monitor, alert, report. On-‐Premise Private Cloud Public Cloud Run Time Run Time Run Time Containers: Build once, run anywhere
- 9. Builder Launcher Controller ISV Data Tools Stackable units with tools (ex: encryp?on), ISV applica?on codes (ex: OpenFOAM). Just add your own codes and data. Run anywhere with UberCloud Run Time. Scale up or down the compute power as needed. Collect granular usage data, logs. Monitor, alert, report. On-‐Premise Private Cloud Public Cloud Run Time Run Time Run Time Containers: Build once, run anywhere
- 10. Builder Launcher Controller ISV Data Tools Stackable units with tools (ex: encryp?on), ISV applica?on codes (ex: OpenFOAM). Just add your own codes and data. Run anywhere with UberCloud Run Time. Scale up or down the compute power as needed. Collect granular usage data, logs. Monitor, alert, report. On-‐Premise Private Cloud Public Cloud Run Time Run Time Run Time Containers: Build once, run anywhere
- 11. UberCloud Containers: Reducing / Removing Cloud Challenges http://www.theubercloud.com/workstations-‐servers-‐clouds-‐comparing-‐apples-‐apples/
- 13. ANSYS Container Demo Ê ANSYS multi-‐physics, Mechanical , Fluent, CFX, Workbench, Icepak, EM, LS-‐DYNA and LS-‐PrePost Ê All running inside one Docker container, in the cloud Ê Pre-‐ and post-‐processing in the cloud work the same way that it would work on the engineer’s desktop Ê User experience in the cloud matches that on the desktop Ê GPU support for high-‐resolution, fast, remote visualization Ê Automated monitoring keeps you connected, under control Ê DEMO: http://youtu.be/-‐5HV69Bjm20 (4 minutes)