Post on 26-May-2018
© 2014 VMware Inc. All rights reserved.
Delivering Transformational User Experience with Blast Extreme Adaptive Transport and NVIDIA GRID.Kiran Rao – Director, Product Management at VMwareLuke Wignall – Sr. Manager, Performance Engineering at NVIDIA
Challenges for Virtual GraphicsProfessional graphics workloads require great user experience in both LAN & WAN environments.
UXRequire “snappy”
experience
Rely on heavy
encoding and
decoding
User density is
limited by CPU
bottleneck
Server
vSphere
Quadro Driver
VMware Horizon Gets Even Better with NVIDIA GRID
NVIDIA GRID
Virtual PC
NV
IDIA
GR
ID m
anagem
ent to
ols
NV
SM
I, N
VM
L
NVIDIA Tesla GPU
H.264 Encode
NVIDIA GRID
Virtual Workstation
NVIDIA
Graphics DriverNVIDIA Quadro
Driver
NVIDIA GRID
vGPU manager
vGPU vGPU
Scheduling – 3D, CE, NVENC, NVDEC
Blast Extreme: Unified Protocol for All VMware Products
• A new VMware controlled protocol for a richer app & desktop experience
• Protocol optimized for mobile and overall lower client TCO
• Horizon remote experience features work with Blast Extreme and updated Horizon clients
• Performance on par or exceeding all competitive protocols
• Rapid client proliferation from strong Horizon Client ecosystem
5
2013 2015 2016 2017
BEAT
Horizon Clients / Broadest Support For Every Use Case
Feature-Rich User Experience
Blast Extreme Designed for All Use Cases
66
Hosted Apps & RDS
Desktops
Printing Scanning & Imaging Devices
USB Audio In/OutSmart Card
Client DriveRedirection
Windows Media Redirection
File TypeAssociation
Unified Communi-
cations
Webcams
Session Enhancement SDK
RDP VC Bridge SDK
Mac iOSLinuxWindows Android Chrome OS HTML (Blast Only)
SDKsWindows VDI, RDSH Apps/Desktop & Linux VDI
Thin Clients
Blast Extreme – A Year Of Progress
7
Use Less Bandwidth:Up to 50% reduction
Broad Client Support: 125+ thin/zero clients certified
GRID Optimized:Superior 3D App
Experience
Network Friendly: Blast Extreme Adaptive Transport
Monitoring:More insight and control
WAN Optimized:Faster access across WAN
Tesla Lineup for GRID
The most powerful data center GPUs targeted at graphics virtualization
M10 M6 M60
GPU Quad Mid-level Maxwell Single High-end Maxwell Dual High-end Maxwell
CUDA Cores 2560 (640 per GPU) 1536 4096 (2048 per GPU)
Memory Size 32 GB GDDR5 (8 GB per GPU) 8 GB GDDR5 16 GB GDDR5 (8GB per GPU)
H.264 1080p30 streams 28 18 36
Max vGPU instances 64 16 32
Form Factor PCIe 3.0 Dual Slot (rack servers) MXM (blade servers) PCIe 3.0 Dual Slot (rack servers)
Power 225W 100W (75W opt) 240W / 300W (225W opt)
Thermal passive bare board active / passive
USER DENSITYOptimized
BLADEOptimized
PERFORMANCEOptimized
NVIDIA Blast Extreme Acceleration
• Reduces overall latency
• Offloads CPU workload to GPU
• Increases scalability
• Improves user experience
• Lowers N/W bandwidth demand
• Supported with single and multiple
monitor use case
GRID GPU
3D
HW Encoder
Framebuffer
AppsApps
Apps
Graphics
commands
Context/Display Capture
Render Target
Front Buffer
H.264
streams
Remote Client
Blast Extreme Adaptive Transport: BEAT
CONFIDENTIAL11
Maintain a Great User Experience
Across a Wide Variety of Network Types:Dynamically Adjust to:
Corporate LAN
Public Wi-Fi
Mobile networks
Varying speeds
Severe packet loss
• Excellent
– TCP Only
– Ideal for Corporate LAN
• Typical (Default)
– UDP for protocol transport, TCP for control and broker communications
– Falls back to TCP if UDP connection is blocked
– Ideal for most situations. Dynamically adjusts for packet loss and jitter
• Poor
– UDP Only
– For protocol transport, control and broker communications
– Requires VMware Unified Access Gateway 2.9 or later
– Necessary when network conditions are so poor can’t make a broker connection typically greater than 20% packet loss
12
Blast Extreme Adaptive Protocol Settings
Blast Extreme Adaptive Transport – by the Numbers
CONFIDENTIAL 13
Improvements
Up to 50% bandwidth
reduction out-of-box
Over 4x faster file transfers for
cross-continental connections~100 ms with slight packet loss (1%)
over previous
Over 6x faster file transfers for
trans-continental connections~200 ms with slight packet loss (1%)
over previous versions
Delivers with Challenging Networks
Over 13x higher average frame rates
under extremely poor network conditionsLow bandwidth, high latency, significant
packet loss (1.5 Mbps, 200 ms, 20%)
Over 70% higher average frame rates
under poor network conditionsLow bandwidth, high latency, and medium
packet loss (1.5 Mbps, 200 ms, 5%)
Over 57% higher average frame rates
with high latency and slight packet lossLow bandwidth, high latency, slight
packet loss (10 Mbps, 200 ms, 1%)
Over 2x faster file transfers under
extreme network conditionsLow bandwidth, high latency, significant
packet loss (1.5 Mbps, 200 ms, 20%)
Windows 10 requires more resources for improvement User Experience
Windows 10 requires more GPU frame bufferWindows 10 requires more CPU cycles
0
100
200
300
400
Windows 7(single
1920x1080)
Windows 10(single
1920x1080)
Windows 10(single
2560x1600)
Windows 10(dual
1920x1080)
0
10
20
30
40
50
60
70
80
90
100
CP
U h
ost
utiliz
ation %
Time
Windows 7 Windows 10
64 x Tesla M10-1B VMs on a host running LoginVSI knowledge worker workload
15% more CPU utilization
Host CPU offloadingBlast Extreme decreases CPU utilization on the host, up to 42%
Lower is
better
63 x Tesla M10-1B VMs on a host running LoginVSI knowledge worker workload and 1 additional VM measuring latency
0
10
20
30
40
50
60
70
80
90
100
NOGPU-PCoIP GPU-PCoIP
NoGPU-JPEG GPU-JPEG
NOGPU-Blast-H.264 CPU GPU-BLAST-H.264CPU
GPU-BLAST-NVENC
0
15000
30000
45000
60000
75000
90000
Total sum
CPU only vs. NVIDIA GRID
0.0
1.0
2.0
3.0
4.0
5.0
Horizon 7 with PCoIP - No GPU Horizon 7 with Blast Extreme and H.264 HW
GPU with NVENC provide an average positive increase to UX of 34%
Higher is
better
Testing ran on two identical systems, CPU system was loaded up to 60-80% utilization, the GPU system ran the same workload
User Experience Scale
1 Unacceptable, unusable -
fire someone in IT!
2 Barely useable, borderline,
but I’ll get tired of this soon
3 Tolerable, I guess I can
make do
4 Pretty good for a virtual
desktop
5 Outstanding - as good (or
almost) as physical
+20% +5%+19% +65%
+6% +21%+55% +26%
+9%+13%+13% +30% +68%+133%
CLICK TO PHOTON SIMPLIFIEDCLICK-TO-PHOTON CAPTURES THE OVERALL LATENCY
Mouse button
released
Mouse click
processed
Packetized and
encoded
Packet Received Packed Decoded
Frame displayedPacket
transmitted
Network Latency on the WAN
(i.e. 50ms)
Network Latency on the WAN
(i.e. 50ms)
Access Device
Server
CLICK-TO-PHOTON LATENCY
Network Latency
Packet ReceivedMouse click
processed
New Frame
renderedFrame Captured via
NVIDIA NVFBC
Frame Encoded via
NVIDIA NVENC
Frame
transmitted
Packet Decoded Application
CLICK to PHOTON Latency
65
185
155 165125
107
250
170
240
160
110
0
50
100
150
200
250
300
Local PC withIntegrated
GPU
Blast ExtremeNo GPU -
JPEG/PNG
Blast ExtremeM10-1B -
JPEG/PNG
Blast ExtremeNo GPU -
H.264Software
Blast ExtremeM10-1B -
H.264Software
Blast ExtremeM10-1B -
H.264Hardware
Idle, 1 VM
Scale, 64VMs
Comparing latency of single VM and at scale(80%) at <1ms network latency
Lower is
better
63 x Tesla M10-1B VMs on a host running LoginVSI knowledge worker workload and 1 additional VM measuring latency
ms
Blast Extreme Adaptive Transport User Experience: FPS in LAN & WAN Environments
0
5
10
15
20
25
30
1%PL , 5%PL 20%PL
FPS:10Mbps, 200ms RTT
BEAT PCoIP
0
5
10
15
20
25
30
1%PL , 5%PL 10%PL 20%PL
FPS:1.5Mbps, 200ms RTT
BEAT PCoIP
0
5
10
15
20
25
30
100Mbps 200Mbps 300Mbps
FPS: LAN, 0% PL, 0ms RTT
BEAT PcoIP
Instant Clone Support for NVIDIA GRID vGPU Desktops
Overview
• Option to provision vGPU desktops using Instant Clones
• Admin to select profile during pool creation. Note: only 1 profile supported per ESX cluster
• Compatible with NVIDIA GRID M Series
Benefits
• Broader set of use cases – no longer have to choose between better provisioning and better 3D graphics
• 2x faster provision over View Composer
23
High Availability with vSphere 6.0 U3/6.5
• vSphere HA enables high availability of VMs in case of server HW failure, by immediately detecting the failure (and thus VM crash) and restarting the VM on another host in the cluster
• vSphere HA is now extended to NVIDIA GRID vGPU-backed VMs with vSphere 6.5/6.0 U3
• If a vGPU server fails with vSphere HA, the VMs will automatically be started on another vGPUenabled host on the same cluster
24
App
Volumes
Design First for User ExperienceEnsure the right level of performance and then use that determine density
User experience must be the equivalent of
they are used to today
Determine how many users you can put on a host based
on requirements of users
One Size Does Not Fit AllDifferent users have different requirements
• Quadro features
• 512MB-1GB
– Office user
• 1-2GB
– Power user
• 2-4GB
– Designer/
engineer
• 8GB
– CUDA
• At least 2-8 vCPUs
• Recommend at
least 4 vCPUs for
Power Users
• 4GB
– Power User
• 8GB
– Mid Eng./Video
• 16GB
– Advanced Eng.
• 32GB
– CAD/CAM
• 64GB
– Digital Mock Up
• vGPU
– 64 bit Windows
– 64 bit Linux
• Install
– VM Tools
– View Agent
– NVIDIA driver
(vGPU)
vGPUvCPU vRAM OS
Performance
User
requirements
Mobility
Security
• CAD users can demand more than 1TB of
storage per desktop
• SIZE and PERFORMANCE matter now
• Virtual SAN – SSD based local storage
• Understand local storage capacity
• All Flash based SAN
Sizing Rich Graphics User: StorageStorage is still a critical factor in performance
Things to think about
with storage
VM VM VM VM
VMware vSphere
Virtual SAN
…
Virtual SAN
Clustered Datastore
Determine Business Drivers
1-2+ weeks
Determine Proof of Concept Scope and
Success Criteria
1-2 weeks
Order GRID enabled Servers per Scope
2-3 weeks
Work with Infrastructure team in parallel on
network address and server infrastructure
2-3 weeks
Rack Servers
1-2 days
Install vSphere 6 and NVIDIA GRID drivers
4 hours
Install vCenter 6 and VMware Horizon 6.2
4 hours
Create GRID enabled vCenter cluster
30 minutes
Create GM images perscope definition
1 day
Create desktop pools
1 day
Invite Pilot Users to Start Testing
Run Pilot and Measure Results
4-8 weeks
Present Results and Move to Next Phase
2-4 weeks
Getting a True Proof Of Concept Up and RunningDetermine business drivers, scope, and success criteria early to ensure successful pilot phase
• Start with the work they actually do today
• Understand their full workflow
• Leverage real files, data and models
• Compare experience on current system to
virtual system
• Score overall experience on a scale of 1-5
Leverage Real User Testing Define acceptance criteria and get feedback
Define User
Acceptance Criteria Monitor
ConfigureRun
Monitor for Proactive ManagementLeverage granular insights to track trends and optimize environment
How do different projects impact usage?
Are there unused cycles during night hours?
Will environment updates impact users?
SCALABILITY COLLABORATION EFFICIENCY SECURITY
50% 20,000+
PRODUCTIVITY
14,700 5,000 $10M+improvement in
workflow cycle time
for an engineering
firm’s remote
CATIA users
engineers accessing
a single, centralized
desktop image for
their virtualized
Siemens NX
workstations
km between an
engineering
firm’s Revit teams
collaborating from
offices in Holland
and Australia
employees at a
global transportation
company receiving
remote video
training instead of
traveling
product development
savings for an
automaker through
intellectual property
protection and real-
time supplier
negotiations
Imagine the Possibilities…Amazing benefits of deploying VMware Horizon with NVIDIA GRID
NVIDIA Blast Extreme Acceleration
User Experience
• 6-13% higher average FPS
• 27-51ms less latency
• 48-89% bandwidth reduction
with TCP and 19% with UDP
You don’t have to chose, you get all at the same time
Source: NVIDIA GRID Performance Engineering Lab
Improves User Experience AND Reduces TCO!
Scalability
• 18% higher user density for
ESRI ArcGIS Pro
VMware & NVIDIA GRID Test DriveTry graphics accelerated virtual desktops for free
• Instant access to the latest NVIDIA GRID
with VMware Horizon 7
• Experience workstation applications like
AutoCAD, SOLIDWORKS, and ArcGIS Pro
in a remote desktop
• Prove to yourself in less than an hour
virtual workstations are here today
www.nvidia.com/trygrid
Free Cloud Trial
Resources
• NVIDIA GRID & VMware Horizon Deployment Guide
– http://www.nvidia.com/object/grid-vgpu-deployment-guide.html
• VMware Horizon with View on Virtual SAN Reference Architecture
– http://www.vmware.com/files/pdf/techpaper/vmware-horizon-view-virtual-san-reference-architecture.pdf
• NVIDIA GRID Certified Servers
– http://www.nvidia.com/object/enterprise-virtualization-where-to-buy.html
• ESRI ArcGIS Pro App Guide with VMware Horizon and NVIDIA GRID vGPU
– http://www.nvidia.com/esriappguide
• Autodesk Revit App Guide with VMware Horizon and NVIDIA GRID vGPU
– http://www.nvidia.com/revitappguide