Cross Architecture
Approach to AI
Dani Quiroga
Global Enterprise Solutions Engineer
AIHUB-1009
#CiscoLive
Cisco Webex App
https://ciscolive.ciscoevents.com/
ciscolivebot/#AIHUB-1009
Questions?
Use Cisco Webex App to chat
with the speaker after the session
How
1 Find this session in the Cisco Live Mobile App
2 Click “Join the Discussion”
3 Install the Webex App or go directly to the Webex space
4 Enter messages/questions in the Webex space
Webex spaces will be moderated Enter your personal notes here
by the speaker until June 7, 2024.
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 2
• Looking Back: the Journey to
the current AI Era
• Cross-Architecture
Agenda Considerations
• AI Readiness
• Cisco for AI
AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 3
Learning Objectives
By the end of this session, you’ll be able to:
• Approach your AI strategy with key
considerations in mind
• Describe cross architecture AI
Readiness
• Understand Cisco’s strategy for AI
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 4
The AI Era
The AI Era – How did we get here?
1940s 1950 1965 1997 2017 2022+
Math, neurophysiology and Turing Test Moore’s Law IBM Deep Blue defeats “Attention is all you need” Consumer-facing
computing research Kasparov in chess Transformer architecture, LLMs
Cambrian Explosion of LLMs
AI
Machine
Learning
IBM 701
computer
plays checkers Deep Learning
Growth in
availability of data
Artificial Neural Networks Gen AI
Computational Probability Domain-specific Models
models models Natural Language Processing BloombergGPT
Claude Code Generation
Image recognition ChatGPT
Statistical analysis
techniques Google Translate Foundational Models Image Generation
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 6
What really changed?
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 7
Emergence and availability of consumer and
enterprise grade models
Scientific research realm Consumer and business realm
Challenges for Organizations
Data Governance Scalable Systems Risk Management
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 8
Cross
Architecture
Considerations
Differentiate Features from AI vs Systems for AI
Features from AI Systems for AI
Background noise removal Deep Neural Network model design
Self-service tools Back-end and front-end networks
Predictive analytics in t-shoot Data storage
AI assistants Deep Learning frameworks and
SDKs
Typically delivered via SaaS On-prem, Cloud
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 10
What does a system for AI entail?
Training Inferencing
Cat or Dog?
Dataset New input
Untrained Neural
Network Model ?
Optimized
Trained Model
Retrieval
Augmented
Generation
Dog ✓ (RAG)
✓ 𝙭
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 11
What does infrastructure for AI entail?
Storage Compute GPU
• High throughput, lossless Frontend Network
fabric for training
10G | 25G | 50G | 100G | 400G | 800G
• Can require extensive
power and cooling
• Training and Inference can
differ in compute density
and utilization
Backend Network
Lossless | High-Throughput | Low Jitter | Low-Latency
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 12
It’s all about the data
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 13
It’s a cross-architecture endeavor…
People, places Access Network Cloud connectivity Apps
and things networks services infrastructure (data center, cloud)
Industrial
IOT
IOT Direct Cloud
Wireless
gateway connect providers
BYOD
Wireless Cloud
Seattle, WA network proxy
Regional
Corp devices data centers
ISP transit SaaS
Local ISP SWG
providers onramp Application
segmentation
DC colocation
Paris, France DNS
Io
IoT Firewall
T load
balancers
Mobile
VD networks VPN
VDI
I
400G/800G
Cameras
Cameras Edge data center
and infrastructure
and sensors Chicago,
Chicago, IL IL data
sensors
centers
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 14
Use Cases Across Industries
Public
Finance Healthcare MFG Automotive Retail
sector
Large Small
Content Creation | AI-Enabled Customer Service
Drug Discovery | Virtual Try-Ons | Automated Robotics | Fraud Detection
Cisco Deployment Templates
Cisco Validated
“Journey to AI” Guide Industry Partnerships
Designs
Training and Inferencing and
fine-tuning Retrieval Augmented
Orchestration, Observability, Policy Enforcement Tools
Generation (RAG)
Models and Frameworks | Software for AI
Pre-Trained Foundational Models | Domain-specific Models | Model Repositories
Accelerators | AI Infrastructure | Scalable Systems
AI-ready infrastructure from the data center to the edge General
AI specific
purpose
Cisco® UCS Nexus® switches
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 15
How can I start?
Initial Questions to Define Cross Architecture
Requirements
• What is the use case?
• Will I need to Train? Inference? Augment?
• What is my model size?
• How will I manage my compute resource
utilization?
• What security policies will I need to
implement?
• Will I need to train my staff?
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 17
Considerations for AI-Readiness
From the Cisco “Journey to AI” guide
• Form an AI committee and establish
clear ownership
• Project the required increase in GPUs
• Ensure hygiene, quality, and reliability
of in-house and external data sources
• Implement mechanisms to detect bias
and checks on AI algorithms
• Establish proficiency for responsible
usage of AI
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 18
Cisco for AI
The Cisco AI Stack
AI Models & Frameworks NGC Developer Cloud
ThousandEyes | Observability Platform | Motific
Secure Firewall | Secure Workload | Panoptica
AI Management Tools
Observability
Virtualization and Kubernetes
Security
Nexus
Infrastructure Management
Dashboard Intersight
AI infrastructure
Nexus UCS
Data center Edge Colocation Public cloud
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 20
Cisco’s AI approach
Using AI to drive operational Enabling Your infrastructure to
efficiency for better experiences support adoption of AI applications
AI-Native IT operations AI-Native infrastructure
Networking | Assurance | Collaboration Data Center Networking | Compute |
Security | Responsible AI Storage | Security | Trusted AI
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 21
What should I do
next?
Next Steps
• Determine your AI readiness across
your environment
• Continue building your AI literacy
• Develop a cross architecture roadmap
for AI
#CiscoLive AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 23
• Visit the Cisco Showcase
for related demos
• Book your one-on-one
Meet the Engineer meeting
Continue • Attend the interactive education
with DevNet, Capture the Flag,
your education and Walk-in Labs
• Visit the On-Demand Library
for more sessions at
www.CiscoLive.com/on-demand
AIHUB-1009 © 2024 Cisco and/or its affiliates. All rights reserved. Cisco Public 24
Thank you
#CiscoLive
