April 4-7, 2016 | Silicon Valley S6385 - WHAT IS CLOUD AND WHAT CAN IT DO FOR YOUR DESKTOP WORKLOADS Matt Coppinger Director, Technical Marketing, End User Computing, VMware Stephane Asselin Senior EUC Architect, Technical Marketing, End User Computing, VMware Jeff Weiss GRID Solutions Architect Manager, NVIDIA
Why GPU in the Cloud? Key technologies for GPU in the Cloud Performance & Sizing Best Practices AGENDA Deployment Quick Start Summary Resources & Next Steps 2
Why GPU in the Cloud? Key technologies for GPU in the Cloud Performance & Sizing Best Practices AGENDA Deployment Quick Start Summary Resources & Next Steps 3
MAJOR MARKETPLACE DEMAND FOR CLOUD-BASED WORKSTATIONS Large datasets More distributed are harder to than ever share Data leakage & Contractors need corporate revocable access security 4 4
TRYGRID DEMO 5
DESKTOP, APP + GPU VIRTUALIZATION ENABLES CLOUD-BASED WORKSTATIONS IT Manages Centralized Virtual Desktops Security & Compliance + Anywhere Access Desktop Virtualization with Server-side GPU Business Agility Improves the Highly Efficient Delivery of Remote Protocol Graphics Mobile & BYOD Access OpEx Savings Users Remotely Access from Various Devices 6 6
WHO BENEFITS THE MOST FROM GPU? DESIGNER POWER USER KNOWLEDGE WORKER TASK WORKER Source: Gartner & NVIDIA 2012 Research 7 7
THE IMPORTANCE OF GPU HIGH-PERFORMANCE GRAPHICS APPLICATIONS PLM & Volume Design Engineering & Design Media 8 8
HIGH-PERFORMANCE GRAPHICS USE CASES Image Oil & Gas Quality Architecture Design Diagnostics Interactivity Engineering Cost/Seat Lightweight CAD Media & Graphics 2D / 3D 3D Viewers GIS WebGL Productivity Apps with Graphics Aero Light Users Heavy Users Fewer Applications Many Applications Task Knowledge Worker Power User Designer Worker 9 9
DELIVERING A COMPLETE PORTFOLIO, FOR EVERY WORKER, EVERY USE CASE Oil & Gas Architecture Design Diagnostics Engineering Lightweight CAD Media & Graphics Image vDGA 3D Viewers Quality GIS WebGL Interactivity Dedicated Cost/Seat Graphics Productivity Apps with Graphics 2D / 3D Acceleration Aero NVIDIA GRID VGPU Mediated Passthrough vSGA Shared Graphics Soft 3D Acceleration Software 3D Light Users Heavy Users Fewer Applications Many Applications Task Knowledge Worker Power User Designer Worker 10 1 0
VIRTUAL BATTLESPACE IN THE CLOUD 11
12
Why GPU in the Cloud? Key technologies for GPU in the Cloud Performance Best Practices and Results AGENDA Deployment Quick Start Review of Cloud Desktop Components Resources 13
KEY TECHNOLOGY: VMWARE HORIZON Private Cloud infrastructure for delivering graphics rich desktops in the GPU enabled datacenter 14
WHY CHOOSE VMWARE FOR CLOUD-BASED WORKSTATIONS? Market leading hypervisor Proven, industry leading platform, fully integrated end- to-end solution Supports huge range of use cases Deliver a superior experience with state of the art protocols Choice of GPU technologies — vGPU, vDGA and others 15 1 5
VMWARE NVIDIA CLOUD ARCHITECTURE Any Device Business Mobility: Applications | Devices | Content vGPU Any Traditional Modern, Cloud Applications Applications Application Cloud Management HYBRID CLOUD PRIVATE MANAGED PUBLIC One Cloud Your vCloud Air vCloud Air Data Center Network Virtualized Compute, Network, Storage Converged Hyper-Converged Build-Your-Own Infrastructure Infrastructure 16 1 6
VMWARE HORIZON: CLOUD-BASED WORKSTATION SOLUTION FOR ALL USE CASES Virtual desktop and Desktop and app Secure containerized app delivery delivery from public desktops cloud Horizon FLEX Horizon Air Horizon 7 Simplifying App Management APP Horizon App Management NEW NEW NEW NEW NEW DESIGNERS/ POINT OF SALE DEVELOPERS ENGINEERS MAIN OFFICE REMOTE OFFICE MOBILE NON-EMPLOYEE 17 1 7
PRIVATE CLOUD VIRTUAL DESKTOP INFRASTRUCTURE Physical server host(s) • • Hypervisor Software Virtual machine management software • • Virtual desktop management system • Virtual desktop connection broker Access and authentication systems • • Display protocol Application Delivery mechanisms • • User Persona/Profile management Clever “stuff” that simplifies it all for you…. • 18 1 8
HORIZON 7: FIRST-TO-MARKET INNOVATIONS TO MAKE THE CLOUD A REALITY Enterprise Scale & Security Deployment GPU Options Virtualization Blast Extreme App Lifecycle Experience Management Smart Policies 19 CONFIDENTIAL 19
INTRODUCING BLAST EXTREME - THE MOBILE CLOUD PROTOCOL Broad Client Support: GRID Optimized: Use Less Bandwidth: H264 available in most devices Deliver more 3D users Deliver better video to more users Network Friendly: Built for the Cloud: Better Battery Life: TCP or UDP works in more places Adapts better to lossy networks Optimized for mobile workers 20 2
VMWARE AND NVIDIA REFERENCE ARCHITECTURE iOS MacOS Thin Client PC Android Kiosk Horizon Clients HTTPS/PCoIP/Blast Extreme DMZ (HTTPS/PCoIP/Blast) PCoIP or Blast Extreme ESX, vCenter, Horizon, SQL, AD traffic Horizon Connection Servers VMware Access Point Shared Storage App Volumes Identity Manager File Print NVIDIA Server vCenter Lic. Svr Horizon 3D Virtual Desktops Horizon Virtual Desktops RADIUS Active CAD/CAx MS Server Directory “Server” SQL SSD SSD View 3D DESKTOP CLUSTER DESKTOP CLUSTER vCenter Operations Composer Manager for Horizon MANAGEMENT CLUSTER 21 Shared Storage (Optional)
KEY TECHNOLOGY: GPU VIRTUALISATION How NVIDIA GRID is bringing GPU power to cloud based workstations 22
HOW DOES NVIDIA GRID WORK? Virtual Virtual Virtual Virtual Virtual Virtual Desktop Desktop Desktop Desktop Desktop Desktop Virtualization Layer Hypervisor Hardware CPUs Server 23
HOW DOES NVIDIA GRID WORK? Virtual Virtual Virtual Virtual Virtual Virtual PC Workstation PC PC Workstation Workstation Virtualization Layer NVIDIA Graphics NVIDIA Graphics NVIDIA Graphics NVIDIA Quadro NVIDIA Quadro NVIDIA Quadro Driver Driver Driver Driver Driver Driver vGPU vGPU vGPU vGPU vGPU vGPU Hypervisor NVIDIA GRID vGPU manager Hardware NVIDIA NVIDIA CPUs Server GPU GPU H.264 Encode 24
NVIDIA GRID PLATFORM Centralized virtual graphics capabilities for the enterprise Support, Updates & Maintenance NVIDIA GRID Software NVIDIA Tesla (Datacenter GPU) 25
DESKTOPS & VIRTUAL APPS Apps NVIDIA GRID NVIDIA GRID NVIDIA GRID Virtual Applications Virtual PC Virtual Workstation 26
GRID vGPU Architecture 101 A quick recap … Hypervisor VM Apps GRID vGPU Manager NVIDIA driver Tesla GPU 29
GRID vGPU Architecture 101 GPU engines Graphics And Compute Hypervisor VM Apps Copy Engine GRID vGPU Manager NVIDIA driver Video Encode Video Decode Tesla GPU Framebuffer (FB) Framebuffer Graphics Video Video Copy Compute Encode Decode Engine 30
GRID vGPU Architecture 101 GPU access Base Address Register (BAR) Hypervisor VM Apps A “window” into the GPU GRID vGPU Manager NVIDIA driver Used to access control registers and framebuffer GPU Base Address Register Virtual BAR Tesla GPU All driver accesses to GPU resources go through a BAR We assign the VM a subset of the GPU’s BAR – a virtual Framebuffer BAR Graphics Video Video Copy Compute Encode Decode Engine 31
GRID vGPU Architecture 101 Access protection Access to BAR controlled by Hypervisor VM Page Table Apps CPU’s Memory Management GRID vGPU Manager NVIDIA driver Unit (MMU) CPU’s MMU Hypervisor creates MMU page Translation tables that allow VM to GPU Base Address Register Virtual BAR Tesla GPU access its Virtual BAR Blocks access to locations outside the Virtual BAR Framebuffer Graphics Video Video Copy Compute Encode Decode Engine 32
GRID vGPU Architecture 101 Work posting VMs post work to GPU via Hypervisor VM Apps channels GRID vGPU Manager NVIDIA driver Each channel is a queue of work from an app to the GPU engines GPU BAR Tesla GPU Each VM has dedicated Channels channels, exposed inside VM’s Virtual BAR Framebuffer Graphics Video Video Copy Compute Encode Decode Engine 33
vGPU Architecture 101 Scheduling GPU’s hardware scheduler Hypervisor VM Apps places work from active GRID vGPU Manager NVIDIA driver channels onto GPU engines Engines sequentially execute work, signal work GPU BAR Tesla GPU completions to NVIDIA Channels driver Scheduler immediately loads engine with next Scheduler work item Framebuffer Graphics Video Video Copy Compute Encode Decode Engine 34
GRID vGPU Architecture 101 Memory protection All GPU engines access Hypervisor VM Sysmem resource memory GRID vGPU Manager NVIDIA driver Each vGPU is assigned fixed range of FB for its exclusive use Tesla GPU GPU’s MMU manages all engine accesses to FB and system memory Memory Management Unit (MMU) Prevents engines accessing Framebuffer outside of vGPU’s FB, or Graphics Video Video Copy the guest VM’s system Guest VM’s FB Compute Encode Decode Engine memory 35
Recommend
More recommend