developer.nvidia.com Open in urlscan Pro
152.199.20.126  Public Scan

Form analysis
2 forms found in the DOM

GET https://developer.nvidia.com/blog/

<form class="gss form-search content-search" action="https://developer.nvidia.com/blog/" method="GET" id="nvidia-site-search-form" accept-charset="UTF-8" __bizdiag="1629530700" __biza="WJ__">
  <div>
    <div class="input-group">
      <input placeholder="Search" class="form-control form-text" type="text" id="edit-term" name="search_posts_filter" value="" size="15" maxlength="128">
      <span class="input-group-btn">
        <button type="submit" class="btn btn-default">
          <i class="fa fa-search" aria-hidden="true"></i>
        </button>
      </span>
    </div>
    <button class="btn element-invisible btn-primary form-submit" type="submit" id="edit-submit" name="op" value="Search">Search</button>
  </div>
  <input type="hidden" name="xsSDAqvzdtWrci" value="FDkgxvR5"><input type="hidden" name="FtnARZm_" value="LqAyvSwPhEi"><input type="hidden" name="kfYsaPoWwZxN" value="goqVT6QX">
</form>

GET https://developer.nvidia.com/blog/

<form class="gss form-search content-search" action="https://developer.nvidia.com/blog/" method="GET" id="nvidia-site-search-form" accept-charset="UTF-8" __bizdiag="1629530700" __biza="WJ__">
  <div>
    <div class="input-group">
      <input placeholder="Search" class="form-control form-text" type="text" id="edit-term" name="search_posts_filter" value="" size="15" maxlength="128">
      <span class="input-group-btn">
        <button type="submit" class="btn btn-default">
          <i class="fa fa-search" aria-hidden="true"></i>
        </button>
      </span>
    </div>
    <button class="btn element-invisible btn-primary form-submit" type="submit" id="edit-submit" name="op" value="Search">Search</button>
  </div>
  <input type="hidden" name="xsSDAqvzdtWrci" value="FDkgxvR5"><input type="hidden" name="FtnARZm_" value="LqAyvSwPhEi"><input type="hidden" name="kfYsaPoWwZxN" value="goqVT6QX">
</form>

Text Content

DEVELOPER
 * Home
 * Blog
 * Forums
 * Docs
 * Downloads
 * Training

 * 
   Search
 * Join
 * 


TECHNICAL BLOG

Subscribe
Technical Walkthrough Oct 03, 2022


EVOLVING FROM NETWORK SIMULATION TO DATA CENTER DIGITAL TWIN

By Eric Pulvino
Discuss (0)
+5
Like
Tags: Data Center, Digital Twin & Metaverse, featured, Technical Walkthrough



Digital twins are attracting increasing attention across industries. While the
concept is relatively new to many, digital twins are not new to IT, which has
for some time recognized the benefits. One such benefit is the value of
simulating a network environment. Network operators have been chasing network
simulators for years.

Cisco’s Packet Tracer was an early industry network simulator that was quite
popular. This simple tool provided a first exposure to network simulation for
countless classically trained network admins. Packet Tracer only offered the
capability to simulate a handful of generic network devices with a limited list
of supported features. Even then, it was easy to see the value network
simulation offered to operators.


THE RISE OF DATA CENTER INFRASTRUCTURE SIMULATION

The capabilities of network simulators have grown immensely over the years,
aided by the move to the cloud. Many infrastructure appliances were
re-envisioned as cloud-native offerings and run as VMs and containers in the
public cloud. They were also ideally suited for data center infrastructure
simulation. 

Armed with a plethora of new simulated device images, the value that can be
extracted from simulations has increased. What started as network simulation has
grown into a new category of holistic data center infrastructure simulation.
This increasingly complex environment is also increasingly driven by automation.
The adoption of automation is another key driver for the use of network
simulation.

Business leaders are realizing that critical business applications sit directly
on top of these brittle sets of interacting software systems, and the value of
simulation is becoming prominent to business productivity.


THE VALUE OF DATA CENTER SIMULATION

The benefits of data center simulation are evident in the data center lifecycle,
from planning to building and maintaining (Figure 1).

Figure 1. The data center deployment lifecycle and various use cases for a
digital twin at each phase 


PLANNING

On Day 0 of the data center deployment lifecycle, you can get ahead of supply
chain challenges and model your environment while your hardware is on order. In
the time which would normally be spent waiting for equipment to arrive, you can
accomplish many preliminary tasks, including:

 * Define cabling architecture
 * Create initial configurations
 * Automate and deploy the entire virtual data center

At this stage, simulation can help you build confidence that your solution is
going to work as intended and model the interaction surfaces between your
different software systems. For example, you can model your DCIM, your
automation platform and the applications themselves, as well as other tools. 

You can also get ahead of multi-vendor issues by verifying interoperability in
your virtual Proof of Concept (vPoC). You can train staff on your new solution
and build familiarity with your specific deployment even before the first
devices have arrived on the loading dock. To learn more, see Close Knowledge
Gaps and Elevate Training with Digital Twin NVIDIA Air.


BUILDING 

On Day 1 of deployment, you directly benefit from the lessons learned from using
simulation during planning. The resulting configuration, automation and topology
information generated in the digital twin can be leveraged to accelerate the
deployment of your physical data center. For larger deployments, you can use
technology such as Prescriptive Topology Manager to validate the cable plans in
physical deployments against the topology built ahead of time in your digital
twin to spot cabling issues. 

In environments with many thousands of cables, just answering the question “is
it plugged in” can be a monumental task. With the digital twin, cable validation
for the whole data center can be done in seconds. Aside from layer one, the
digital twin can be used as a reference for the physical deployment to verify
the initial state of the control plane in layers one through four.


MAINTAINING

With the data center deployed, the operational phase of the lifecycle begins.
During this phase, you can use the digital twin to model the changes in the
environment prior to deployment. You can attach the digital twin to your CI/CD
pipeline to automatically validate any configuration or topology change prior to
deployment and any resulting impact on connectivity of your applications. 

Additional benefits of the digital twin during this phase are focused around
operators, specifically the ability to troubleshoot the virtual environment in
ways which would never be allowed in production. This kind of deep
troubleshooting and chaos engineering can get ahead of numerous problems which
may be hiding in your architecture. It can also rapidly accelerate the
onboarding of new personnel giving them a risk-free learning environment that
identically matches production.


DATA CENTER DIGITAL TWINS

With current simulation technology, it is now possible to simulate thousands of
routers, switches, and data center infrastructure devices fully loaded with
configurations. 

The most important aspects of the future of data center digital twins include:

 * Connecting the digital twin with the physical twin and synchronizing the two
 * Increasing the accuracy of simulation such that all relevant behaviors of the
   data center and network can be simulated completely 

Increased accuracy and synchronization are what separate network simulation from
a true digital twin. And every IT administrator should be striving to achieve a
true digital twin.

Check out NVIDIA Air to start building your own data center digital twin.
Enhance your data center operations with the open network operating system
NVIDIA Cumulus Linux.


ABOUT THE AUTHORS

About Eric Pulvino
Eric is the product manager working on NVIDIA Air Data Center Infrastructure
Simulation Platform and the user experience across Ethernet switching products.
He spent his early time at Cumulus Networks (now part of NVIDIA), building
simulation and automation infrastructure in what eventually evolved to become
the foundation for NVIDIA Air. Eric is very customer focused, having worked for
much of his career in professional services and customer enablement roles
deploying, automating, and simulating customer networks.
View all posts by Eric Pulvino


COMMENTS


START THE DISCUSSION AT FORUMS.DEVELOPER.NVIDIA.COM



TOPICS

.NET
3D Pose Estimation
3D printing
5G
A100
A30
accelerated computing
Accelerating IO
advanced API performance
Aerial SDK
Aerospace
Agriculture
AI
AI / Deep Learning
AI Enterprise
AIoT
Alea GPU
Algorithms
AMA
AmgX
Annotation
Ansel
Apache Spark
AR / VR
AR/VR
Architecture / Engineering / Construction
Arm
ArrayFire
Artificial Intelligence
ASAP2
Astronomy & Astrophysics
Atomics
Audio Effects
Auditing
Automatic Mixed Precision (AMP)
Automatic Speech Recognition (ASR)
Automotive / Transportation
autonomous vehicles
autoscaling
avatar
AWS
Azure
Bayesian optimization
Beginner
benchmarks
BERT
best practices
Big Data & Data Mining
bioinformatics & genomics
biopharma
BLAS
blue noise
BlueField
Business Intelligence & Analytics
C++
Caffe
Clara
Climate/Weather/Ocean Modeling
Cloud
cloudRAN
CloudXR
CLX
community spotlight
Competition
Compilation
Computational Chemistry
Computational Fluid Dynamics (CFD)
Computational Photography
Computer Aided Design (CAD)
Computer Graphics & Visualization
Computer Vision / Video Analytics
Concurrency
ConnectX
consumer internet
Conversational AI / NLP
Cooperative Groups
Cosmology
COVID
COVID-19
criteo terabyte dataset
cross-process synchronization
CSI
CUBLAS
cuCIM
CUDA
CUDA C++
CUDA C/C++
CUDA Fortran
CUDA graph
CUDA Graphs
CUDA Refresher
CUDA-Pointpillars
CUDACasts
cuDF
cuDNN
cuFFT
cuML
Cumulus
cuQuantum
cuRAND
cuSOLVER
cuSPARSE
cuStateVec
cuStreamz
cuTensorNet
Cybersecurity / Fraud Detection
DALI
DARTS
DASK
Data Center
Data Center / Cloud
Data Center Network Architecture
data management
data preprocessing
Data Science
Database
DataFrame
DCGM
Debugging
decision trees
Deep Learning
Deep Learning News
deep neural networks
DeepStream
DesignWorks
DetectNet
developer
developer preview
Developer Spotlight
developer tools
Developer Top 5
Developer Voices
Development Tools & Libraries
Device queries
DGL
DGX
digital finger
digital fingerprinting
Digital Twin & Metaverse
DIGITS
Direct3D
Directives
DirectX
DirectX 12
DLI
DLI Teaching Kit
DLRM
DLSS
DOCA
Docker
DP4A
DPDK
DPU
DRIVE
DRIVE AGX
DRIVE Labs
DRIVE OS
DRIVE PX
DRIVE release
DriveWorks
drug discovery
DXR
Dynamic Parallelism
Edge Computing
Education & Training
EGX
Embedded
Emerging Chapters
Energy
Error handling
esports
Ethernet
ETL
Events
EVPN
Explainer
F#
featured
Federated Learning
ffmpeg
Financial Services
Finite Difference
firewall
Fleet Command
floating point
Fortran
foveated rendering
FP16
FP8
Game Development
GameWorks
GAN
GANs
Gaussian process
Gaze Handler
GDC
GeForce
geolocation
Geometry
Geoscience
Geospatial
global illumination
GLSL
GNMT
GOAI
Google
Government / National Labs
GOY
GPF
GPU
GPU computing
GPUDirect
Gradient Boosting
Graph Algorithms
Graph Analytics
Graph Coloring
Graph Cuts
graph neural networks
Graphics
green computing
GRID
GROMACS
Group Normalization
GstInterpipe
GStreamer
GTC
GTC Insider’s Guide
GtsD
GXF
H.264
H100
Hackathon
Hands-on
Hardware / Semiconductor
HashiCorp
HBAO+
Healthcare / Life Sciences
Helm
Helm Chart
Hemi
heterogeneous compute
HEVC
HGX
Higher Education / Academia
Highlights
histogram
HMD
Holoscan
homomorphic encryption
Hopper
Horovod
hospitality industry
HPC / Supercomputing
HPC SDK
HugeCTR
HuggingFace
hyperparameters
Hyperscale
hypervisor
IBM
ICYMI
idealo.de
ILSVRC
Image Processing
Image Recognition
Image Segmentation
ImageNet
In-situ
IndeX
Ineroperability
Inference
InfiniBand
information retrieval
Infrastructure
Innova-2
Installation
INT8
Intel
interactive
Internet of Things
Internet of Things (IoT)
Internet/Communications
Interoperability
Interview
IoU
Isaac ROS GEMs
Isaac SDK
Isaac Sim
IVA
Jasper
Java
JetBot
JetPack
Jetson
Jetson AGX Orin
Jetson Orin
JIT
JPEG2000
Julia
Jupyter Notebooks
Kaggle
Kaldi
Kepler
Keras
KFServing
Khronos
KITTI
Kubeflow
Kubernetes
Lambda
LAMP
lane keeping
Languages
Large Hadron Collider
large language models
latency
LaunchPad
leaf-spine
level of detail
libnvvm
Libraries
lidar
lighting
Linear Algebra
Linux
Linux on Windows
Linux Switch
LLVM
load balancing
LOD
Longstaff Schwartz
lossless
LSDalton
LSTM
LULESH
LZ4
machine learning
Machine Learning & Artificial Intelligence
Machine Learning and AI
Machine Translation
Magnum IO
Maintenance
maker
malware
Manufacturing
Mask R-CNN
Materials Science
MATLAB
Matrix Factorization
Mavenir
Maxine
Maxwell
MD.ai
MEC
Media & Entertainment
Media Server
medical devices
medical imaging
Megatron
Mellanox
MemN2N
MemNN
Memory
mesh shaders
meshlets
Mesos
Metropolis
micro-segmentation
microservice architecture
Microsoft
Microsoft DirectX Raytracing
Microsoft DXR
MIG
Minecraft
Minigo
MIPI
mitigate
Mixed Precision
ML
ML Flow
mlops
MLP
MLPerf
mNews
Mobile
Model Analyzer
model parallelism
model serving
model training
Modulus
Molecular Dynamics
MONAI
Monte Carlo
Morpheus
motion vectors
MovieLens
MPI
multi-camera
Multi-GPU
Multi-grid
multi-view rendering
multihoming
multinode
MVR
MxNet
n-body
N-series VMs
NAMD
nanotechnology
NanoVDB
NAS
NCCL
NCCL 2.4
NeMo
NetQ
network functions virtualization
network OS
Networking
neural networks
Neuroscience
News
NFS
NFV
NGC
NICs
NLP
NMT
Nomad
NRD
Nsight
nsys
Numba
Numerical Algorithms
NV
NV News
NVAIL
nvcc
nvComp
NVDEC
nVector
NVENC
nvGRAPH
NVIDIA Air
NVIDIA Ampere
NVIDIA Clara Imaging
NVIDIA DOCA
NVIDIA DRIVE
NVIDIA Flare
NVIDIA Fleet
NVIDIA Fleet Command
NVIDIA FleX
NVIDIA Highlights
NVIDIA Inception
NVIDIA KVM
NVIDIA Merlin
NVIDIA Pro Services
NVIDIA Research
NVIDIA RTX
NVIDIA-certified
nvidia-smi
nvJPEG
NVLink
NVML
nvprof
NVRHI
NVRTC
NvSci
NVSHMEM
NVSwitch
NVTabular
NVTAGS
NVTX
NVUE
NVVL
NVVP
O&G
Occupancy
ODTK
Omniverse
Omniverse ACE
ONNX
ONNX-Graphsurgeon
open graphics
open source
Open vSwitch
open-source software
OpenACC
OpenCV
OpenGL
OpenMP
OpenPower
openseq2seq
OpenShift
OpenVDB
Operations
optical flow
Optimization
OptiX
orchestration
Orin SoC
Other
OvS
P100
P4
PackNet
Pandas
Parabricks
Parallel Programming
Pascal
path tracing
PCI Express (PCIe)
PCL
perception
Performance
performance events
PF Scheduling
PGI
phishing
photography
physics
PhysX
PICA
PilotNet
pipeline state objects
pipelines
Pixvana
PLAsTiCC
plugin
point cloud
point clouds
Portability
Porting
Power
Precision Time Protocol
precompiled
Predictive
Pretrained Models
Prevent
Pro Tip
Profiling
Profiling Tool
profiling tools
programming languages
Programming Languages and Compilers
Project Aurora
Prototyping
PSO
Public Sector
Pure SONiC
PyData
PyProf
Python
PyTorch
Q2VKPT
QAT Toolkit
QODA
Quadro
Quake II
Quantum Chemistry
quantum computing
QuartzNet
question answering
quick link
R
RAN intelligent controllers
Random Forest
ranking algorithms
RAPIDS
Rasa
rasterization
ray tracing
RDMA
Real-Time Communication
Recommender Systems
Recommenders / Personalization
Red Hat
redaction
Reduction
reflex
Reinforcement Learning
release notes
Rendering
research
resnet-50
resolution
Resources
Retail
RetinaNet
reverse time migration
RHEL
Riva
Rivermax
RLE
RNN
Robotics
RoCE
Root cause analysis
ROS
roundup
RT Core
RTGII
RTX
Runtime Compilation
Scala
scene generation
scheduling
scientific computing
SDN
SDS
Security
SEED
Segmentation
SegNet
Seismic Processing
semantic segmentation
sensor processing
Server-side
shaders
Shap
Shape Sensing
Shared Memory
SIGGRAPH
Signal/Audio Processing
SigOpt
Sim2Real
Simulation / Modeling / Design
singularity
SLURM
Smart Cities / Spaces
smart hospitals
SmartNIC
Social Impact
software defined network
Software Defined Radio
Software Engineering
Software Tools and Libraries
SONiC
SPARQL
Sparsity
SPDK
spectroscopy
Spectrum
Spectrum-3
Speech & Audio Processing
SPIN Studio
SPIR-V
Sport Analytics
Spotlight
SQL
SSD
STAC
stacked memory
standard parallelism
Stereo
Storage
streaming
streaming AI
Streams
Summit
Super Resolution
Supercomputing / Cluster
SuperPOD
surround video
synthetic data
System Management
System on a Chip
System Performance
Tacotron
Tacotron2
TAO
TAO Toolkit
Technical
Technical Walkthrough
Tegra
telco
Telecommunications
telecoms
telemetry
Templates
temporal noise reduction
Tensor Core
Tensor Cores
TensorFlow
TensorFlow 2
TensorRT
Tensors
Teradici
Teraproc
Tesla P100
Tesla Platform
Tesla T4
Tesla V100
Text Processing
Text-to-Speech (TTS) / Speech Synthesis
texture space shading
Textures
TF32
The PhoenIX
Theano
thread groups
Thrust
tips
TITAN
TITAN RTX
TITAN V
TLAS
TLT
Tools
Top Stories
TOP500
Torch
training
transcoding
transcription
Transfer Learning
Transfer Learning Tookit
Transfer Learning Toolkit
Transformer
Transformers
trends
TREx
trie
Triton
TTS
TU102
Turing
Turing advanced shaders
tutorial
TX2
UDF
UE4.22
UFM
Ultrasound
Unified Fabric Manager
Unified Memory
Unity
Unity3D
Universal Scene Description
Unreal
Unreal Engine
User Group
V100
variable rate shading
vGPU
video
Video Codec SDK
Video Effects
video encoding
video processing
Video Streaming / Conferencing
virtual assistant
Virtual Background
Virtual GPU
virtual machines
virtual memory
virtual network functions
Virtual PC
Virtualization
Vision Programming Interface
Visual Profiler
Visualization
Vive Pro Eye
VKRay
VMware
VNF
voice assistant
Volta
vPC
VPI
vRAN
VRS
VRS Helper
VRWorks
Vulkan
Vulkan Ray Tracing
VXLAN
walkthrough
Warehouse
Warp 360
WaveGlow
WaveNet
webcam
Webinar
What Just Happened
Windows
Windows Subsystem for Linux
Windows Virtual Desktop
WinML
WJH
Workstation
WSL
WSL2
XDP
XGBoost
XLA
zero-trust
ZeroLight
ZTP

+ AR + Computer Vision + Conversational AI + Cybersecurity + Data Center + Data
Science + Networking + Recommenders + Rendering + Robotics + Simulation +
Automatic Speech Recognition (ASR) + Automotive + CUDA + Digital Twin &
Metaverse + Game Development + Healthcare + HPC + News + NVIDIA Research +
Omniverse + Technical Walkthrough

 * 



FEATURED

Simulation / Modeling / Design Anyone Can Build Metaverse Applications With New
Beta Release of NVIDIA Omniverse
Computer Vision / Video Analytics New on NGC: SDKs for Large Language Models,
Digital Twins, Digital Biology, and More
Computer Vision / Video Analytics Getting to Know Autonomous Vehicles
Data Center / Cloud Open-Source Fleet Management Tools for Autonomous Mobile
Robots
Simulation / Modeling / Design New Courses for Building Metaverse Tools on
NVIDIA Omniverse

RELATED POSTS

Rendering Accelerating Load Times for DirectX Games and Apps with GDeflate for
DirectStorage
Conversational AI / NLP Deploying a 1.3B GPT-3 Model with NVIDIA NeMo Megatron
Data Center / Cloud Enabling Enterprise AI Transformations for Telcos with
NVIDIA and VMware
Computer Vision / Video Analytics An AIoT Solution for Visual Blockage Detection
at Culverts
Networking Choosing NVIDIA Spectrum for Microsoft Azure SONiC
SIGN UP FOR NVIDIA DEVELOPER NEWS
Subscribe
Follow NVIDIA Developer
Facebook Twitter LinkedIn Instagram YouTube

Copyright © 2022 NVIDIA Corporation
 * Legal Information
 * Privacy Policy
 * Cookie Policy
 * Contact


 * 
   Search
 * Join
 * 

 * Home
 * Blog
 * Forums
 * Docs
 * Downloads
 * Training

More topics

Cookie Settings

NVIDIA websites use cookies to deliver and improve the visitor experience. Learn
more about the cookies we use on our Cookie Policy page.


More information
Allow All

Manage Consent Preferences

Required Cookies

Always Active
Required Cookies

These cookies are required for the site to function and cannot be turned off.

Performance Cookies

Performance Cookies

Counts website visits and clicks to understand where people most engage with
links to make the experience better.

Advertising Cookies

Advertising Cookies

Set by our advertising partners, these cookies are used to build a profile of
your interests and show you relevant ads on other sites. They do not store
personal information, but are based on uniquely identifying your browser and
internet device.

Save and Accept

Back Button

Back

Performance Cookies



Vendor Search Search Icon Filter Icon


Clear Filters

Information storage and access
Apply
Consent Leg.Interest

All Consent Allowed

Select All Vendors
Select All Vendors
Select All Hosts

Save and Accept


NVIDIA uses cookies to deliver and improve the website experience. See ourCookie
Policy to learn more.



Cookies Settings Accept All Cookies