Cisco UCS and Intersight — IVI's Compute Foundation for AIM

Cisco UCS serves as IVI's compute platform of choice for AIM environments — not because it's the only option, but because its fabric-based architecture aligns with the operational requirements of disaggregated infrastructure.

UCS Fabric Interconnects handle I/O virtualization at the network edge. vNICs and storage adapters are defined in software via service profiles, not hardwired to specific hardware. When a blade fails, you replace the hardware, reassign the service profile, and the new node comes up with identical network and storage configuration. This consistency reduces operational variance in managed environments.

Cisco Intersight delivers cloud-based lifecycle management across the UCS domain — firmware compliance, health monitoring, automation, and proactive alerting. Aegis LM uses Intersight as the management plane for UCS firmware lifecycle. This level of centralized control is not achievable with comparable simplicity on commodity servers.

IVI has designed, deployed, and managed UCS environments across hundreds of enterprise deployments. This is not a new capability — it's the compute foundation of IVI's data center practice. UCS serves as the compute layer in FlashStack (Cisco + Pure + VMware/Nutanix) and multiple Nutanix validated design guides. These validated designs reduce deployment risk and provide structured TAC escalation paths.

UCS Manager Mode (UMM) is being phased out. Cisco has confirmed that the M7 generation is the last UCS hardware generation to support UMM. All new AIM deployments use Intersight Managed Mode (IMM). For existing UMM environments, IVI recommends IMM migration as part of modernization planning.

Modular chassis design with UCS X9508 chassis supporting up to 8 compute nodes per chassis. UCS X-Series blades (X210c M6, X210c M7, X215c M6) run latest Intel Xeon Scalable or AMD EPYC processors. These are compute-only nodes with minimal local storage and maximum compute density — purpose-built for AIM disaggregated architecture where compute runs AHV and storage lives on Pure.

GPU support is available for AI/ML workloads on the same chassis. X-Series supports Intersight Managed Mode only. Fabric Interconnects 6454 or 6536 are recommended with X-Series.

C220 M7 provides 1U dual-socket high-density compute, optimal for compute-only nodes. C240 M7 offers 2U dual-socket with storage expansion when local storage is required. C245 M7 delivers AMD EPYC option. These can be managed standalone via IMC/CIMC or through Intersight. For AIM environments, prefer Intersight-managed C-Series for consistent lifecycle management.

B200 M5 and M6 remain widely deployed and supported under current maintenance. B480 M5 provides four-socket capability for larger workloads. These can be managed via UCS Manager or IMM — M5/M6 support both modes, but use IMM going forward. B-Series is not receiving new hardware generations. New blade deployments should use X-Series modular chassis.

Cisco Intersight is a SaaS-delivered infrastructure management platform. It connects to on-premises UCS domains via an Intersight Device Connector that runs as a process on Fabric Interconnects or as a virtual appliance, providing centralized management, monitoring, and automation.

UCS Manager Mode (legacy approach) runs management locally on Fabric Interconnects with XML API for automation, firmware management via UCS Manager upgrade wizard, and single domain view only. M7 is the last hardware generation to support UMM.

Intersight Managed Mode (current standard) delivers management via cloud-delivered Intersight SaaS platform. Server Profiles replace UCS Manager Service Profiles with modern REST API. Terraform provider enables Infrastructure as Code automation. Cross-domain fleet management provides visibility across all UCS domains in one view. Intersight Assist enables air-gapped deployments for organizations with connectivity restrictions. Intersight Workload Optimizer provides automated remediation workflows.

Aegis LM operational model with Intersight uses firmware compliance policies to define required firmware versions for BIOS, CIMC, VIC adapters, and FI NX-OS — Intersight flags any deviation. Automated firmware upgrade staging executes during defined maintenance windows. Health advisories and proactive alerts integrate with Cisco TAC for proactive hardware health monitoring, including PSIRT alerts and field notices.

IVI configuration standard for AIM uses Intersight Managed Mode for all new builds. Server Profile templates are defined per workload tier — Nutanix AHV compute nodes have specific templates covering vNIC count, NVMe/TCP storage adapter configuration, and boot policy. Firmware compliance policy aligns to IVI's tested baseline, updated as new validated versions are released. Intersight Assist deploys for environments with restricted internet connectivity.

UCS provides compute resources (CPU, RAM) and the network I/O layer for Nutanix AHV. UCS Fabric Interconnects serve as the aggregation point for all VM network traffic and storage traffic (NVMe/TCP) before handing off to the Arista Ethernet fabric.

IVI's server profile design for AIM compute nodes includes dedicated vNIC configuration: Management vNIC for dedicated VLAN covering AHV host management, Prism Element, IPMI/CIMC access. VM data vNIC as trunk port carrying VM VLANs, connects to standard VM network on Arista leaf. Storage vNIC (NVMe/TCP) dedicated to storage VLANs, MTU 9000 (jumbo), QoS policy applied to prioritize storage traffic. AHV live migration vNIC on dedicated VLAN for AHV live migrations (some environments combine with VM data VLAN depending on scale).

Each vNIC has primary and secondary paths through FI-A and FI-B respectively for redundancy. Failover is handled by the Cisco VIC adapter — transparent to AHV.

Fabric Interconnect uplink design connects FI-A and FI-B each to separate Arista leaf switches (dual-homed). Port-channels (LACP) from each FI to its Arista leaf provide bandwidth aggregation. Storage VLAN trunks on FI uplinks with jumbo frames enabled end-to-end (FI, Arista port, Arista L3 interface if inter-VLAN routing is required).

IVI's design capabilities cover UCS domain sizing including FI selection, blade/rack count, and compute-per-node specifications. Service Profile and Server Profile template design aligns to AIM workloads. vNIC architecture maps to Arista fabric design with storage, VM, and management VLANs fully mapped end-to-end before deployment. Firmware baseline establishment uses IVI-tested versions validated against Nutanix and Pure compatibility matrices. Intersight configuration includes organizations, profiles, policies, and firmware policy.

Deployment execution includes UCS domain initial setup with FI discovery and firmware update to baseline. Server profile template deployment pushes to all compute nodes. IMM migration handles existing UMM environments. Integration testing validates vNIC connectivity, storage path validation, and Intersight connectivity and alarm delivery.

Aegis managed services provide continuous health monitoring via LogicMonitor using UCS Manager XML API or Intersight REST API. Firmware lifecycle management operates through Intersight compliance policies. Incident response covers hardware failures, FI alerts, and service profile issues. IVI manages and responds to Intersight alarms with operational ownership.

UCS is the best fit for AIM environments when organizations are already running UCS and can protect existing Cisco investment while maintaining Cisco TAC relationships through the VMware exit. Organizations evaluating compute platforms for new Nutanix AHV builds who want fabric-managed compute with centralized lifecycle management via Intersight will benefit from this approach. Organizations with compliance requirements benefit from Intersight's firmware compliance policies and audit-trail change history. Environments where Infrastructure as Code matters gain value from Intersight's Terraform provider and REST API that integrates with GitOps workflows.

This approach is not the primary fit for organizations with strong Dell or HPE commitments who want to consolidate on their existing hardware vendor. IVI can support Nutanix AHV on Dell or HPE, but the AIM practice and managed services are optimized for UCS. Very small environments (under 10 VMs, single host) where UCS overhead is not justified should consider alternative approaches.

If you are already running UCS, continue with UCS. IVI supports through the VMware exit and manages the environment post-modernization to protect the investment. If you are not running UCS, evaluate UCS vs. commodity servers based on operational model requirements, scale, budget, and Intersight value for lifecycle management.

If your operations team has UCS expertise, this approach builds on existing skills with IVI augmenting through Aegis. If your team lacks UCS expertise, IVI's Aegis managed service absorbs the UCS operational burden — you don't need deep UCS expertise when IVI is co-managing.

If Infrastructure as Code is important to your roadmap, Intersight REST API and Terraform provider are strong arguments for UCS over commodity servers. If IaC is not a priority, this becomes less of a differentiator.

For new compute deployments, choose X-Series for high-density compute, GPU workloads, maximum blade density per rack, and single chassis management. Choose C-Series for rack-mount simplicity, smaller deployments, and standalone management option.