Cisco CCNP Designing Cisco Data Center Infrastructure for Traditional and AI Workloads (300-610 DCID) v1.2

Data center teams are being asked to support AI/ML clusters alongside traditional enterprise workloads, and design mistakes like an unsupported single-APIC deployment, unprotected Layer 2 interconnects, or mismatched RTO/RPO targets can turn into outages and costly rework.

Role fit: This course is for data center engineers and architects with 3–5 years of experience who are moving from configuration into design work, including whiteboarding, specifying equipment, and justifying architecture decisions. It assumes DCCOR-level knowledge, so IT Directors should sequence it after a DC Core course or equivalent data center experience.

Time investment: Plan for about 49 hours across 58 skills per learner. The material covers networking, UCS compute, storage networking, and data center automation and management concepts, including Nexus Dashboard.

Change-management note: Many skills end with design challenges that make team members defend the reasoning behind choices like ACI vs. NDFC vs. HyperFabric. That builds a shared design vocabulary across the team before a greenfield build or migration.

Two closing review skills, one for networking and one for compute, storage, and automation, reinforce readiness for the 300-610 exam. Training Managers can use Team Reporting to track each engineer’s progress toward CCNP Data Center readiness.

IT teams complete this training to move from configuring individual devices to defending design decisions that hold up in production. Instructor Jeff Kish builds the course around realistic design challenges, so practitioners learn to ask the "why" questions before equipment is ordered. Those questions cover how architecture maps to business requirements, how the fabric will scale, where implementation risks sit, and what physical or technical constraints need to be resolved early.

• Fewer avoidable implementation issues: Engineers learn to account for migration and design details that can break an otherwise strong architecture, including SVI placement, port compatibility, rack space, optics, cable counts, link planning, and power connections.
• Production-appropriate architectures from day one: Teams learn to evaluate Cisco fabric options such as HyperFabric, ACI, and NDFC based on workload fit, scale, security needs, and automation maturity. For AI and ML workloads, they also learn when an on-prem or private data center design makes more sense than a public cloud default.
• Business requirements agreed on early: Architects learn to validate port counts, speeds, redundancy, licensing, and future growth against what the organization actually needs before selecting hardware or finalizing the design.
• Storage and compute designs that are planned honestly: Teams learn to evaluate storage requirements alongside the network design, including unified ports, Fibre Channel over Ethernet where it already exists, and whether the selected Nexus switches support the capabilities the environment requires.

The result is a team better prepared to whiteboard, justify, and document AI-ready and traditional data center designs with fewer surprises during implementation.

Knowledge

• How AI/ML training and inference differ, and why AI workloads drive requirements for lossless Ethernet, QoS, and high-performance fabrics
• How vPC, peer-switch, object tracking, and VXLAN data and control planes support resilient L2 and L3 data center designs
• Why STP BPDUs must be blocked and TCN domains contained across Layer 2 DCIs, and when overlay-based interconnects are a better fit
• How RPO and RTO differ, and how snapshot granularity and backup design support recovery objectives
• When ACI, NDFC, or HyperFabric fits a given scale, security need, and operational maturity
• How UCS X-Series, IMM, Intersight, and Nexus Dashboard centralize compute and fabric management

Ability

• Justify design choices to stakeholders by working through the "why" and "why not" behind each decision
• Design vPC domains and redundancy, including peer-switch placement on the aggregation layer and object tracking for failover
• Evaluate SAN isolation trade-offs across Fibre Channel, iSCSI, and NVMe-oF, and flag red flags in proposed iSCSI builds
• Plan UCS network pathing, QoS, service profile policies, and templates, including removing single points of failure such as a lone VIC
• Plan centralized policy workflows so changes reach the whole fabric consistently
• Write backup and DR recommendations that align admin teams on agreed RPO and RTO targets