Corso DCAIE – AI Solutions on Cisco Infrastructure

Obiettivi | Certificazione | Contenuti | Tipologia | Prerequisiti | Durata e Frequenza | Docenti | Modalità di Iscrizione | Calendario

Il Corso DCAIE – AI Solutions on Cisco Infrastructure è parte del percorso Cisco CCNP Data Center. Il corso DCAIE introduce i Partecipanti ai concetti, alle architetture e alle tecnologie fondamentali per distribuire, migrare e gestire soluzioni Artificial Intelligence su infrastrutture Cisco Data Center. Il corso è pensato per fornire una visione tecnica delle esigenze specifiche dei workload AI/ML, evidenziando l’impatto che applicazioni di Machine Learning, Deep Learning e Generative AI hanno su network, compute, storage, orchestration e operations. Durante il corso vengono trattati i principali elementi necessari per supportare ambienti AI su infrastruttura Cisco, tra cui AI/ML clusters, Jupyter Notebook, Python, Generative AI, RAG, open source GPT models, workload placement, interoperability, AI policies, sustainability, security e governance. Il programma approfondisce inoltre le architetture di rete per workload AI, includendo optical and copper technologies, connectivity models, Layer 2 e Layer 3 protocols, fog computing, RDMA, RoCE, high-throughput converged fabrics, lossless fabrics, ECN, PFC e congestion visibility. Il percorso affronta anche le componenti infrastrutturali necessarie per l’esecuzione di soluzioni AI/ML, con focus su AI-enabling hardware, compute resources, virtual resources, storage resources, software-defined storage, data preparation, AI/ML workload data performance e configurazione di fabric ottimizzate tramite NDFC. Viene inoltre introdotto l’utilizzo di modelli GPT locali con Retrieval-Augmented Generation (RAG) per attività di network engineering e supporto operativo. Il Corso contribuisce alla preparazione dell’esame di Certificazione CCNP Data Center DCAI (Esame 300-640).

Obiettivi del corso

Di seguito una sintesi degli obiettivi principali del Corso DCAIE – AI Solutions on Cisco Infrastructure:

• Comprendere i concetti fondamentali di Artificial Intelligence, Machine Learning, Deep Learning e Generative AI applicati al Data Center.
• Analizzare l’impatto dei workload AI/ML su network, compute, storage, orchestration, performance e operations.
• Valutare architetture e tecnologie per AI infrastructure, incluse AI/ML clusters, workload placement, interoperability, governance e sustainability.
• Approfondire tecnologie di rete per ambienti AI, tra cui RDMA, RoCE, ECN, PFC, high-throughput fabrics e lossless fabrics.

Certificazione del corso

Esame 300-640 DCAI Cisco Certified Specialist – Data Center AI Infrastructure;
Esame Parte della Certificazione CCNP Data Center. Questo esame valuta le competenze del candidato nella progettazione, implementazione, gestione e ottimizzazione di infrastrutture Data Center dedicate a workload AI/ML. Il superamento dell’esame consente di ottenere la certificazione Cisco Certified Specialist – Data Center AI Infrastructure e soddisfa il requisito concentration per il percorso Cisco CCNP Data Center. L’esame verifica la capacità dell’esaminato di comprendere i fondamenti di Artificial Intelligence, Machine Learning, Deep Learning e Generative AI, con attenzione all’impatto delle applicazioni AI sulle architetture Data Center. Sono testate competenze relative ad AI/ML clusters, modelli pre-trained, fine-tuning, optimization, RAG, open source GPT models e utilizzo di strumenti come Jupyter Notebook per attività tecniche e operative. Una parte centrale riguarda la progettazione dell’infrastruttura per workload AI, includendo compute resources, AI-enabling hardware, virtual resources, storage resources, workload placement, interoperability, data preparation e data performance. L’esame copre inoltre tecnologie di networking ad alte prestazioni, tra cui RDMA, RoCE, high-performance Ethernet fabrics, lossless fabrics, ECN, PFC, connectivity models, Layer 2 e Layer 3 protocols applicati ad ambienti AI/ML. Il candidato deve dimostrare competenze anche su monitoring, operations e troubleshooting di infrastrutture AI-enabled, con focus su Cisco Nexus Dashboard Insights, NDFC, telemetry, congestion visibility, log correlation, analisi dei flussi AI/ML e risoluzione dei problemi comuni nelle fabric AI/ML.

Contenuti del corso

Fundamentals of AI

• Core concepts of Artificial Intelligence, Machine Learning, and Deep Learning
• Differences between traditional AI approaches and modern AI systems
• Main AI techniques and their practical applications
• Role of AI in automation, analytics, and decision support
• Impact of AI adoption on Cisco Data Center infrastructure

Generative AI

• Key concepts of Generative AI and foundation models
• Differences between traditional AI and generative AI methodologies
• Main challenges of Generative AI, including accuracy, bias, and hallucination
• Use cases of Generative AI in IT operations and infrastructure management
• Future trends of Generative AI in enterprise environments

AI Use Cases

• AI use cases for network management and Data Center operations
• Intelligent automation for infrastructure and operational workflows
• Predictive analytics for performance, capacity, and fault prevention
• Anomaly detection for network monitoring and security
• AI-driven support for troubleshooting and operational optimization

AI-ML Clusters and Models

• Architecture and key components of AI/ML clusters
• Basic management principles for AI/ML cluster environments
• Use of pre-trained Machine Learning models
• Concepts of model acquisition, fine-tuning, optimization, and usage
• Operational considerations for AI/ML models in Data Center environments

AI Toolset Mastery – Jupyter Notebook

• Use of Jupyter Notebook for AI and infrastructure-related activities
• Execution of Python-based workflows for AI-assisted operations
• Use of Generative AI to support network operations and automation
• Development of scripts and technical workflows in notebook environments
• Productivity enhancement through AI models and interactive toolsets

AI Infrastructure

• Essential components of modern AI infrastructure
• Relationship between AI workloads and Data Center architecture
• Infrastructure requirements for supporting AI/ML applications
• Design considerations for compute, network, storage, and orchestration
• Operational impact of AI workloads on Cisco Data Center environments

AI Workload Placements and Interoperability

• Strategies for effective AI workload placement
• Interoperability requirements across AI infrastructure components
• Evaluation of compute, storage, and network dependencies
• Optimization of workload distribution for performance and efficiency
• Infrastructure planning for scalable AI/ML environments

AI Policies

• Governance frameworks for enterprise AI systems
• Compliance standards and policy considerations for AI deployments
• Security and operational policies for AI-enabled infrastructure
• Risk management related to AI usage and infrastructure exposure
• Alignment between AI policies, business requirements, and IT operations

AI Sustainability

• Principles of sustainable AI infrastructure
• Environmental and economic considerations for AI deployments
• Optimization of resource consumption in AI/ML environments
• Efficiency strategies for compute, storage, and networking resources
• Sustainability impact of AI workload design and infrastructure decisions

AI Infrastructure Design

• Design principles for Data Center infrastructure supporting AI workloads
• Evaluation of architecture options for AI/ML applications
• Cost, performance, efficiency, and scalability considerations
• Infrastructure design decisions for AI-enabled environments
• Alignment of Data Center design with AI workload requirements

Key Network Challenges and Requirements for AI Workloads

• Network requirements driven by AI/ML application behavior
• Performance challenges related to throughput, latency, and congestion
• Impact of distributed AI processing on Data Center networks
• Requirements for reliable and scalable AI workload connectivity
• Network design considerations for AI training and inference workloads

AI Transport

• Role of transport technologies in AI/ML Data Center workloads
• Use of optical and copper technologies for high-performance connectivity
• Transport requirements for large-scale AI data movement
• Performance considerations for AI/ML traffic patterns
• Selection of transport options based on workload and infrastructure needs

Connectivity Models

• Network connectivity models for AI/ML infrastructure
• Design patterns for connecting compute, storage, and network resources
• Connectivity requirements for AI clusters and distributed processing
• Evaluation of scalability, resiliency, and performance in connectivity models
• Integration of connectivity design into Data Center AI architectures

AI Network

• Network design principles for AI/ML workload environments
• Role of Layer 2 and Layer 3 protocols in AI infrastructure
• Network considerations for fog computing and distributed AI processing
• Optimization of network behavior for AI training and inference
• Design of dedicated networks for AI/ML workloads

Architecture Migration to AI/ML Network

• Migration strategies toward dedicated AI/ML network architectures
• Assessment of existing Data Center infrastructure for AI readiness
• Planning of network changes to support AI workload requirements
• Transition from traditional network designs to AI-optimized designs
• Risk reduction during migration to AI-enabled infrastructure

Application-Level Protocols

• Role of application-level protocols in AI/ML infrastructure
• Protocol requirements for distributed AI workloads
• Interaction between applications, data flows, and network behavior
• Impact of application communication patterns on infrastructure design
• Considerations for performance, reliability, and scalability

High Throughput Converged Fabrics

• Architecture of high-throughput converged Ethernet fabrics
• Features required to support AI/ML traffic at scale
• Fabric design considerations for high-performance Data Centers
• Support for compute, storage, and AI workload convergence
• Operational considerations for converged AI infrastructure fabrics

Building Lossless Fabrics

• Principles of lossless fabric design for AI/ML workloads
• Use of RDMA and RoCE in high-performance Data Center networks
• Network mechanisms required to reduce packet loss and congestion
• QoS tools for building reliable AI-ready fabrics
• Design considerations for performance-sensitive AI applications

Congestive Visibility

• Use of congestion visibility tools in AI/ML fabrics
• Monitoring of traffic behavior and network congestion
• Role of ECN and PFC in congestion management
• Use of Cisco Nexus Dashboard Insights for congestion monitoring
• Analysis of how AI/ML application stages affect infrastructure performance

Data Preparation for AI

• Core steps of the data preparation process for AI workloads
• Challenges related to data quality, structure, and readiness
• Techniques for preparing data for AI/ML processing
• Impact of data preparation on AI model performance
• Relationship between data lifecycle and AI infrastructure requirements

AI/ML Workload Data Performance

• Analysis of data performance requirements for AI/ML workloads
• Monitoring of AI/ML traffic flows and workload behavior
• Use of Cisco Nexus Dashboard Insights for data flow visibility
• Identification of performance bottlenecks in AI-enabled environments
• Optimization of infrastructure for AI/ML data movement

AI-Enabling Hardware

• Role of specialized hardware in AI workload acceleration
• Hardware requirements for reducing AI training times
• Processing needs of AI, Machine Learning, and Deep Learning tasks
• Infrastructure impact of GPU and accelerator-based systems
• Evaluation of hardware choices for AI-enabled Data Centers

Compute Resources

• Compute requirements for running AI/ML solutions
• Role of servers, accelerators, and processing resources in AI infrastructure
• Resource planning for training, inference, and distributed AI workloads
• Performance considerations for compute-intensive AI tasks
• Operational management of compute resources in AI environments

Compute Resource Solutions

• Existing compute solutions for AI/ML infrastructure
• Evaluation of Cisco Data Center compute options for AI workloads
• Integration of compute platforms into AI-enabled architectures
• Scalability and performance considerations for compute resource planning
• Operational considerations for managing AI compute environments

Virtual Resources

• Virtual infrastructure options for AI/ML deployments
• Considerations for deploying AI workloads in virtualized environments
• Impact of virtualization on performance, flexibility, and operations
• Resource allocation and isolation for AI workloads
• Evaluation of virtual resources within Data Center AI architectures

Storage Resources

• Data storage strategies for AI/ML environments
• Storage protocols and their role in AI infrastructure
• Software-defined storage considerations for AI workloads
• Performance requirements for AI data access and movement
• Integration of storage resources into AI-enabled Data Center designs

Setting Up AI Cluster

• Steps for setting up an AI cluster environment
• Configuration of infrastructure components for AI/ML workloads
• Use of NDFC to configure fabrics optimized for AI/ML
• Validation of AI cluster readiness and connectivity
• Operational considerations for AI cluster deployment

Deploy and Use Open Source GPT Models for RAG

• Deployment of open source GPT models for technical use cases
• Use of Retrieval-Augmented Generation (RAG) for network engineering tasks
• Integration of local GPT models with infrastructure-related data
• Application of RAG to improve contextual accuracy and response relevance
• Operational considerations for using GPT models in Cisco Data Center environments

Attività Laboratoriali

• AI Toolset—Jupyter Notebook
• AI/ML Workload Data Performance
• Setting Up AI Cluster
• Deploy and Use Open Source GPT Models for RAG

Tipologia

Corso di Formazione con Docente

Docenti

I docenti sono Istruttori accreditati CISCO e certificati in altre tecnologie IT, con anni di esperienza pratica nel settore e nella Formazione.

Infrastruttura laboratoriale

Per tutte le tipologie di erogazione, il Corsista può accedere alle attrezzature e ai sistemi reali Cisco presenti nei Nostri laboratori o direttamente presso i data center Cisco in modalità remota. Ogni partecipante dispone di un accesso per implementare le varie configurazioni avendo così un riscontro pratico e immediato della teoria affrontata. Ecco di seguito alcune topologie di rete dei Laboratori Cisco Disponibili:

Dettagli del corso

Prerequisiti

Si consiglia la partecipazione al Corso Cisco DCCOR.

Durata del corso

• Durata Intensiva 4gg;

Frequenza

Varie tipologie di Frequenza Estensiva ed Intensiva.

Date del corso

• Corso Cisco DCAIE (Formula Intensiva) – Su Richiesta – 09:00 – 17:00

Le iscrizioni sono a numero chiuso per garantire ai tutti i partecipanti un servizio eccellente.
L’iscrizione avviene richiedendo di essere contattati dal seguente Link, o contattando la sede al numero verde 800-177596 o inviando una richiesta all’email [email protected].