Artigo

VMware: ESXi, vCenter, VCF, and VVF — 2026 technical guide

EnQ Digital·02 de setembro de 2026

This technical guide brings together, in a single view, the current-generation VMware platform: the ESXi hypervisor, vCenter Server as the control plane, and the two packaged offerings — VMware vSphere Foundation (VVF) and VMware Cloud Foundation (VCF) — covering per-core licensing, management, identity, automation, security, performance, and continuity.

This guide is intended for architects, engineers, infrastructure managers, pre-sales teams, and decision makers. The capacity examples are sizing models, not guarantees for any given workload — they must be recalculated against real measurements of CPU, memory, latency, IOPS, throughput, RPO, and RTO.

Five-layer map of the VMware platform: consumption, operations, control, infrastructure, and hardware, with the technologies of each layer
From certified hardware to infrastructure-as-a-service consumption — the platform’s five layers.

About this material: content consolidated from official Broadcom/VMware sources available as of September 2, 2026, including VMware Cloud Foundation 9.x and VMware vSphere Foundation in the current generation. Licensing, promotional metrics, versions, support, and availability may change — confirm the commercial proposal, the Product Guide, and the applicable contractual terms before purchasing or renewing.

Executive overview and platform map

The VMware stack starts with ESXi, the bare-metal hypervisor installed directly on servers. vCenter Server adds inventory, policies, clusters, and distributed services. VVF packages the workload and HCI platform; VCF expands the set into a full-stack private cloud with software-defined networking, automation, fleet management, Kubernetes, and advanced services.

LayerMain functionTypical technologies
ConsumptionPortal, catalog, API, governance, and costsVCF Automation, APIs, IaC
OperationsHealth, capacity, logs, alerts, and complianceVCF Operations
ControlInventory, clusters, policies, and lifecyclevCenter, vLCM
InfrastructureCompute, storage, and virtual networkingESXi, vSAN, NSX
HardwareCPU, RAM, NVMe, NIC, HBA, and acceleratorsHCL-certified servers

Decision in one sentence: choose VVF when the core need is managed virtualization and HCI; choose VCF when the goal is to operate an integrated, automated, multitenant, self-service-oriented private cloud.

Key capabilities

CapabilityWhat it deliversOperational value
vMotionLive migration of VMs between hostsMaintenance without planned downtime
HAAutomatic restart after a host failureFast recovery
DRSResource-oriented load balancingBetter utilization and fewer hotspots
vSANPolicy-based distributed storageHCI and horizontal scale
VKSKubernetes integrated into the control planeVMs and containers on the same platform
NSX / VPCSoftware-defined networking and isolationAgility and segmentation
VCF OperationsObservability, capacity, and diagnosticsLower time to resolution
VCF AutomationCatalog, workflows, policies, and APIsSelf-service IaaS

Three operational planes

  • Data plane: VMs, containers, networks, and volumes run on the hosts.
  • Control plane: vCenter, managers, and controllers maintain state, inventory, and policies.
  • Management plane: Operations, Automation, and Fleet Management monitor, provision, and update the platform.

ESXi: the virtualization layer

ESXi provides workload isolation and schedules vCPUs onto physical CPUs, managing memory, devices, drivers, datastores, and virtual switches. Its design reduces the attack surface compared with a general-purpose operating system, but it demands discipline around firmware, compatibility, hardening, and lifecycle.

CPU and NUMA

  • Size vCPU based on observed demand, not theoretical maximums. Oversized VMs may wait longer for scheduling.
  • Keep critical VMs within a single NUMA node whenever possible; very wide VMs may access remote memory with higher latency.
  • Analyze CPU Ready, Co-Stop, utilization, application latency, and effective frequency together.

Memory

  • Avoid sustained pressure that triggers ballooning, compression, and swapping; host swap is the last resort and degrades latency.
  • Reservations guarantee capacity, Limits impose a ceiling, and Shares define priority under contention.
  • Factor in hypervisor overhead, agents, vSAN, and infrastructure services when calculating cluster capacity.

Virtual networking

ElementUseRecommendation
Standard vSwitchIsolated host or simple environmentGood for small scale; per-host configuration
vSphere Distributed SwitchCentralized policy and advanced featuresPreferable in enterprise clusters
Port groupVLAN and logical policyStandardize names and IDs
vmkernelManagement, vMotion, vSAN, NFS, FTSeparate traffic and apply redundancy
NIC teamingAvailability and distributionValidate failover with the physical switches

Storage and lifecycle

ESXi can consume VMFS over FC/iSCSI SAN, NFS, and vSAN. The choice must balance latency, throughput, resilience, operations, backup integration, and cost. A VM uses configuration files, virtual disks, temporary snapshots, and logs — snapshots are not a substitute for backup. For lifecycle management, validate hardware, firmware, drivers, and the compatibility matrix; define the desired image in vSphere Lifecycle Manager; test on a pilot group; enter maintenance mode by evacuating workloads; and validate health, network, storage, alarms, and performance after the change.

A metric that matters: IOPS without block size, read/write ratio, queue depth, and latency is an incomplete number. Record the workload’s entire profile.

vCenter Server: central control

The vCenter Server Appliance centralizes hosts, VMs, templates, tags, permissions, clusters, and tasks. It is not in the I/O path of VMs: if vCenter stops, workloads continue running, but management functions and dependent services become limited. Its recovery must be planned and tested.

Inventory objectResponsibility
DatacenterLogical container for clusters, hosts, networks, and datastores
ClusterDomain for HA, DRS, EVC, and lifecycle
Resource poolHierarchical reservation, limit, and priority
FolderOrganization and permission scope
Tag / CategoryMetadata for policy, automation, and governance
Content LibraryTemplates, OVFs, ISOs, and content distribution

vCenter high availability

Protect the appliance with native file-based backup to an external destination, cluster HA, and a coherent copy of configurations. The strategy should include restoring the appliance, DNS, NTP, certificates, break-glass credentials, and identity dependencies. Avoid creating a circular dependency in which the only DNS, AD, or repository needed to recover vCenter is unavailable within the very environment affected.

Administration antipatterns

  • Running day-to-day operations with administrator@vsphere.local.
  • Granting a global role when the operator only works within a single folder or cluster.
  • Using the same technical account for backup, monitoring, and automation.
  • Keeping certificates, DNS, and NTP without validity and availability monitoring.
  • Relying solely on an appliance snapshot instead of a restorable native backup.

RBAC principle: assign corporate groups to minimum roles at the correct scope. Avoid granting permissions directly to users and do not use Administrator for daily tasks.

Licensing: VCF versus VVF

The current model is a per-physical-core subscription. For each physical CPU on an ESXi host, a minimum of 16 licensable cores applies, even if the processor has fewer. All physical cores running the software must be counted. Commercial terms, term length, support, and add-ons must be confirmed in the applicable proposal.

Side-by-side comparison of VMware vSphere Foundation (VVF) and VMware Cloud Foundation (VCF) by positioning, included vSAN, networking, automation, lifecycle, and fit
VVF organizes the base as a workload and HCI platform; VCF adds the private cloud operating model.
CriterionVMware vSphere Foundation (VVF)VMware Cloud Foundation (VCF)
PositioningWorkload / HCI platformFull private cloud platform
ComputevSphere Enterprise Plus + vCenter StandardvSphere integrated into the VCF stack
Included storage0.25 TiB vSAN per core1 TiB vSAN per core
Software-defined networkingNot the core of the packageIntegrated NSX
OperationsVCF Operations for infrastructureUnified full-stack operations
Automation / self-servicevSphere APIs and automation; smaller scopeVCF Automation, catalog, and governance
LifecyclevLCM and VVF InstallerFleet Management and full-stack lifecycle
KubernetesVKS includedVKS integrated into the private cloud
Best fitVirtualization, consolidation, HCIInternal IaaS, multitenancy, networking/security, and cloud operating model

Core count formula

Licensable core formula for VCF and VVF, with three calculation scenarios resulting in 96, 160, and 512 cores
Licenses = sum, per physical CPU, of max(physical cores, 16). SMT does not double the count.
ScenarioCalculationLicensable cores
3 hosts, 2 CPUs/host, 12 cores/CPU3 × 2 × max(12, 16)96
4 hosts, 2 CPUs/host, 20 cores/CPU4 × 2 × 20160
8 hosts, 2 CPUs/host, 32 cores/CPU8 × 2 × 32512

vSAN example

A cluster with 4 hosts, 2 CPUs of 20 cores per host: 160 cores. The included entitlement is 40 TiB under VVF and 160 TiB under VCF. If the cluster contributes 120 TiB raw to vSAN, VVF would require 80 TiB in addition; VCF would have coverage included, subject to the applicable terms. Licensing considers the raw physical capacity contributed, not the usable space shown after policies and overhead.

VCF 9: license management

  • A single license file and a consolidated view in VCF Operations, replacing multiple per-component keys.
  • Connected or disconnected / air-gapped mode.
  • Usage submission and updates within each 180-day window.
  • Without usage data, existing workloads continue running, but management may go disconnected and new workloads can no longer be created, per the official VCF 9 overview.

Commercial caution: do not estimate TCO simply by multiplying a “price per core” found online. Include subscription term, support, add-ons, excess vSAN, certified hardware, services, taxes, exchange rate, migration, and operations.

Decision tree

  • Do you just need to consolidate VMs, HA/DRS, vMotion, external storage, or simple HCI? Start by evaluating VVF.
  • Do you need integrated NSX, VPCs, self-service, catalog, governance, and full-stack lifecycle? Evaluate VCF.
  • Do you need high vSAN density? Compare VVF’s 0.25 TiB/core entitlement with VCF’s 1 TiB/core and the add-ons.
  • Do you need DR, advanced security, load balancing, or ransomware protection? Check the advanced services; not everything is part of the core.

Management dashboards and observability

The right dashboard depends on the question. vSphere Client is excellent for administering objects and tasks. VCF Operations correlates health, capacity, cost, alerts, and trends. VCF Automation delivers a catalog and consumption governance. Fleet Management organizes the stack’s lifecycle.

DashboardPrimary userQuestions answered
vSphere ClientVirtualization administratorWhich VM, host, network, or datastore needs action?
VCF OperationsNOC, SRE, capacity, FinOpsWhere is there risk, waste, anomaly, or saturation?
VCF AutomationDevOps, platform team, internal usersHow to request and govern infrastructure as a service?
Fleet ManagementPlatform teamWhich instances, versions, credentials, and updates make up the fleet?
Business Services ConsoleSubscription managementWhat license allocation and consumption exist?
BlockIndicators
AvailabilityDisconnected hosts, HA admission control, VMs without redundancy, hardware failures
CPUUsage, Ready, Co-Stop, contention, frequency, and NUMA
MemoryConsumed, Active, balloon, compression, swap, and reservations
StorageRead/write latency, IOPS, throughput, congestion, capacity, and resync
NetworkDrops, errors, saturation, failover, latency, and loss
CapacityDays remaining, reclaim, growth, N+1/N+2 headroom
ExperienceProvisioning time, workflow failures, incidents, and SLO

Avoid noise: a dashboard with 200 ownerless alerts is decoration. Consolidate correlated symptoms, define SLOs, and generate tickets only when there is a clear action — with symptom, probable cause, context, and action (runbook, owner, priority, and closure criteria).

Active Directory, Entra ID, and Azure

“Azure AD” has been renamed Microsoft Entra ID. vCenter can federate authentication with corporate providers, enabling SSO and modern policies. Traditional Active Directory remains relevant in on-premises environments. The architecture must distinguish authentication, authorization, and identity synchronization.

ModelWhen to usePoints of attention
AD over LDAPLocal AD and classic integrationTLS, certificates, bind accounts, and DC dependency
Integrated Windows AuthenticationSpecific legacy casesEvaluate support and product direction
AD FSFederation already standardized on AD FSFarm availability, claims, and certificates
Microsoft Entra IDCloud identity, MFA, and conditional accessEnterprise application, groups, claims, and emergency accounts
Okta / PingFederateAlternative corporate IdPGroup mapping and certificate cycle
  • Create and authorize the enterprise application according to the vCenter version’s documentation.
  • Configure the identity provider in vCenter and validate URLs, tenant, client ID, and certificates/secrets when applicable.
  • Map Entra ID groups to vCenter roles with least privilege.
  • Apply MFA and Conditional Access at the IdP, with controlled exceptions only for recovery.
  • Test login, logout, expiration, group removal, IdP unavailability, and the break-glass account.

Group example

Corporate groupSuggested roleScope
GG-VMW-Platform-AdminsPlatform administratorvCenter/Datacenter, restricted group
GG-VMW-VM-OpsCustom VM operatorProduction folders/resource pools
GG-VMW-AuditorsRead-only + eventsDatacenter
GG-VMW-BackupMinimum permissions for the backup productNecessary objects
GG-VMW-DevelopersCatalog consumerProjects/organizations in Automation

Important boundary: connecting vCenter to Entra ID does not “move VMware to Azure” — it is identity federation. Workload extension, DR, or operation in public cloud are separate architectural decisions. Keep at least two local recovery accounts protected in a vault that work even if DNS, AD, Entra ID, or external connectivity fail.

Orchestration, automation, and APIs

Automation executes tasks; orchestration coordinates multiple tasks, dependencies, approvals, policies, rollback, and state. In VCF, the goal is to turn infrastructure into a consumable service with guardrails, not just to write VM-creation scripts.

LayerToolsExample
API / SDKvSphere API, REST, Unified SDKCreate a VM and query inventory
CLIPowerCLI, govc, esxcliReports and batch operations
IaCTerraform providersDeclare networks, VMs, and policies
WorkflowVCF Automation OrchestratorApproval, IPAM, DNS, CMDB, and rollback
CatalogVCF AutomationBlueprint/templating and self-service
Guest configurationAnsible, scripts, cloud-initInstall middleware and apply baseline

Best practices

  • Idempotency: repeating an action must not create duplicates or corrupt state.
  • Secrets: use a vault, rotation, and service identities; never embed passwords in code.
  • Mandatory tags: owner, application, environment, data-class, backup-policy, cost-center, and expiry.
  • Git: version templates and workflows, review via pull request, and promote across environments.
  • Observability: log requester, parameters, steps, return, and correlation ID.
  • Rollback: distinguish reversible from destructive actions and require appropriate approval.

Minimum catalog governance

ControlImplementation
QuotaLimit per project, environment, and resource class
ApprovalOnly when risk, cost, or privilege justifies it
ExpirationDefault TTL for labs and review before renewal
EvidenceOwner, ticket, template version, and workflow result
DecommissioningFinal backup when needed, DNS/IP revocation, and CMDB write-off

Expected outcome: a standardized request can drop from days to minutes, but the correct KPI includes success rate, compliance, rework, cost, and recovery time — not just speed.

Performance, sizing, and benchmarks

Performance must be measured end to end. The hypervisor can be healthy while the database, network, or storage limits the application. Use baselines, percentiles, and controlled tests; averages hide spikes.

DomainIndicatorInvestigation signal
CPUCPU ReadySustained growth with application latency
CPUCo-StopWide SMP VM waiting for scheduling
MemoryBalloon / SwapReal pressure or inadequate reservations
StorageGuest/kernel/device latencyLocate the queue and the limiting layer
StorageIOPS + block size + R/WCharacterize the demand, not just the volume
NetworkDrop / error / throughputQueue, MTU, driver, uplink, or congestion
vSANCongestion / resyncImpact of rebuild, policy, or device

Example 1: general virtualization cluster

Four hosts, each with 2 CPUs of 24 cores and 512 GB of RAM. Gross total: 192 cores and 2,048 GB. Reserving one host for N+1 failure, the approximate operational capacity is 144 cores and 1,536 GB before overhead. With 120 VMs averaging 2 vCPU and 8 GB, there are 240 vCPU and 960 GB configured. The operational vCPU:pCore ratio is 1.67:1, and configured memory occupies 62.5% of N+1 capacity.

ItemValueReading
Gross CPU192 cores4 × 2 × 24
N+1 CPU144 cores3 hosts available
Configured vCPU240 vCPU120 × 2
N+1 vCPU:pCore ratio1.67:1Conservative for a mixed workload
N+1 RAM1,536 GBBefore overhead
VM RAM960 GBNominal headroom of 576 GB

Example 2: transactional storage

Peak demand of 80,000 IOPS, 8 KiB block, 70% read and 30% write. Approximate logical throughput: 80,000 × 8 KiB = 625 MiB/s. The architecture must sustain the pattern with compatible latency, device failures, and protection policy. Also test rebuild/resync, snapshots, and backup, since steady state alone does not represent real-world operation.

Example 3: vendor-published targets

Published metricReference valueContext
End-to-end migrationUp to 65% fasterParallel DRS vMotion processing
Source CPU in encrypted vMotionUp to 70% lowerIntel QAT hardware offload on compatible hardware
“Monster VM” scaleUp to 960 vCPU and 16 TB RAMCurrent VVF generation limit
Gain on 4-socket systemsUp to 68%Topology/NUMA-aware scheduling
vGPU vMotionUp to 6× fasterOffload and higher-bandwidth networks
TroubleshootingUp to 60% less effortFull-stack proactive diagnostics

Beware of “up to”: these values depend on hardware, version, and methodology and should not be added together to form a business case. Run a PoC with the real profile and defined acceptance criteria — set an SLO and hypothesis, capture a production baseline, reproduce CPU/memory/I/O/network and concurrency, test steady state, peak, maintenance, failure, and recovery, and compare percentiles, not just the average.

Networking, storage, availability, and DR

A resilient platform is designed starting from acceptable impact, not from a product list. Translate criticality into SLO, RPO, and RTO, and only then choose HA, replication, backup, and DR.

TrafficFunctionRecommended practice
ManagementHost and appliance managementRedundancy, ACLs, and controlled administrative access
vMotionMemory/state migrationHigh bandwidth, low latency, and isolation
vSANCluster I/O and resyncDedicated/logical network, compatibility, and headroom
Storage IPNFS / iSCSIMultipathing, queues, consistent MTU, and QoS when needed
WorkloadsApplications and tenantsSegmentation, policies, and observability
Backup / replicationCopy and DRAvoid competing with production during critical windows

Availability

  • N+1 protects against a single host failure; N+2 may be needed during maintenance or in critical clusters.
  • HA restarts VMs — it does not equal uninterrupted continuity. Fault Tolerance addresses specific cases with its own requirements.
  • Admission Control must reserve failover capacity; disabling it masks the risk.
  • Stretched clusters reduce RTO for site failure, but require compatible latency, witness, network, and dependency design.

Backup and DR

MechanismProtects againstDoes not replace
HAHost failureBackup or site-level DR
vMotionPlanned maintenanceReplication or backup
SnapshotShort-term changeIndependent backup
ReplicationSite/array loss depending on designImmutable copy against corruption/ransomware
Immutable backupDeletion, corruption, and ransomwareContinuity plan and restore capability
SRM / DR orchestrationFailover sequencing and testingInfrastructure and data at the destination

The real test: RPO and RTO are only reliable after a test that restores applications, identity, DNS, network, secrets, and integrations — not just powering on VMs. The test should cover simultaneous host and storage-path failure at peak, vCenter/DNS/NTP/IdP/KMS unavailability during recovery, site loss with partial isolation, isolated restore after ransomware, and return to the primary site.

Security, hardening, and operations

Security in VMware spans hardware, ESXi, vCenter, network, identities, VMs, automation, and the supply chain. The goal is to reduce the attack surface, limit lateral movement, detect drift, and recover quickly.

ControlImplementation
IdentityFederation, MFA, groups, least privilege, and break-glass
Management planeDedicated network, jump host / PAM, ACLs, and centralized logs
HostSecure Boot, TPM 2.0, lockdown when applicable, minimal services
EncryptionUp-to-date TLS, VM/vSAN encryption, and highly available KMS
NetworkSegmentation, distributed firewall / add-ons per license
LifecyclePatches, HCL, desired images, and remediation window
AuditingEvents, changes, login, tasks, configuration, and retention
RecoveryImmutable backups, tested restore, and credentials outside the failure domain

Day 2 routine

  • Daily: critical alarms, failover capacity, backups, hardware, and failed tasks.
  • Weekly: anomalies, old snapshots, orphaned VMs, datastore, resync, and urgent patches.
  • Monthly: capacity forecast, license entitlements, privileged accounts, certificates, and sample restores.
  • Quarterly: DR test, RBAC review, firmware/HCL, runbooks, and dependency risks.
  • Annually: architecture, contract, TCO, hardware cycle, SLOs, and modernization plan.

Reference architectures and deployment

Pattern A — VVF with external storage

ComponentDesign
Compute4 to 8 ESXi hosts, N+1, compatible CPUs, and RAM according to profile
ControlvCenter + VCF Operations
StorageRedundant FC/iSCSI SAN or NFS; multipathing
NetworkvDS, redundant uplinks, and segregated VLANs
UseVM consolidation, ERP, databases, middleware, and VDI

Pattern B — VVF HCI with vSAN

ComponentDesign
ClusterTechnical minimum per architecture; 4+ hosts is common for margin
StoragevSAN ESA/OSA depending on hardware and compatibility
ManagementvCenter + Operations + SPBM
ScaleAdd nodes or use storage cluster / disaggregated per support
UseSimplified operation of compute and storage in one cluster

Pattern C — VCF private cloud

ComponentDesign
Management domainControl, management, and lifecycle services
Workload domainsClusters separated by SLA, hardware, tenant, or purpose
NetworkNSX, VPCs, overlays, and gateways per architecture
ConsumptionAutomation, catalog, projects, quotas, and policies
OperationsOperations, Fleet Management, logs, and integrations
UsePrivate cloud, internal IaaS, modern apps, Kubernetes, and private AI

Design rule: do not start with the number of hosts. Start with the workload, failure domain, SLA, growth, compatibility, operations, and licensing — hardware is the consequence. The phases are assessment, HLD, LLD, build, validate, migrate, and operate.

Migration, costs, and checklist

StrategyAdvantageRisk / limitation
vMotion / cross-vCenterLow disruption when compatibleNetwork, latency, versions, EVC, and storage
HCXMobility at scale and network extensionAppliance planning and throughput
Backup / restoreIndependent and verifiableCopy window and time
ReplicationReduced RPO and controlled cutoverConsistency and bandwidth
RebuildModernizes OS and configurationMore application and testing effort

3-year TCO model

CategoryItems
SoftwareVCF/VVF, additional vSAN, advanced services, backup, observability
HardwareHosts, disks, switches, optics, SAN, support, and parts
FacilitiesRack, power, cooling, space, and connectivity
ServicesAssessment, deployment, migration, training, and support
OperationsStaff, on-call, patches, capacity, audit, and incidents
RiskDowntime, lock-in, delay, rework, and obsolescence

VVF × VCF decision checklist

  • Inventory of hosts, sockets, cores, and raw vSAN capacity completed.
  • Workloads classified by criticality, RPO, RTO, CPU, RAM, I/O, network, and growth.
  • Need for NSX, microsegmentation, VPCs, and multitenancy validated.
  • Need for catalog, self-service, quotas, chargeback, and APIs validated.
  • Advanced services separated from the core and confirmed in the proposal.
  • 3-to-5-year TCO including hardware, migration, operations, support, exchange rate, and taxes.
  • PoC measuring percentiles, failure, maintenance, backup, and restore.
  • Exit/mobility plan and data portability documented.

Essential glossary

TermShort definition
ESXiBare-metal hypervisor of the vSphere platform.
vCenterCentral plane for management, inventory, and policies.
VCFVMware Cloud Foundation; full-stack private cloud platform.
VVFVMware vSphere Foundation; workload / HCI platform.
vSANSoftware-defined storage integrated into vSphere.
NSXSoftware-defined networking and security.
VKSvSphere Kubernetes Service.
DRSDistributed Resource Scheduler.
HAHigh Availability; automatic restart after failure.
SPBMStorage Policy-Based Management.
RPOMaximum acceptable data loss.
RTOMaximum time to restore service.
HCLHardware/software compatibility list.
EVCCPU compatibility for mobility between hosts.
NUMANon-uniform memory architecture in multiprocessor servers.

Conclusion

ESXi and vCenter remain the robust foundation for enterprise virtualization. VVF organizes that foundation as a modern workload and HCI platform. VCF adds the private cloud operating model: integrated networking, full-stack management, fleet lifecycle, automation, self-service, and governance. The right choice is not “which has more features,” but which fits the organization’s operating model, risk, and TCO.

Practical recommendation: for a traditional VM environment with a central team and external storage, VVF tends to be the natural starting point. To offer internal IaaS, multiple tenants, Kubernetes, end-to-end automation, and software-defined networking, VCF tends to reduce manual integrations — provided the organization adopts platform processes and actually uses these capabilities.

Official sources and further reading

This material was prepared from official Broadcom/VMware sources, with a cutoff date of September 2, 2026. Always consult the applicable release notes, interoperability matrices, the VMware Compatibility Guide, the Product Guide, and the current commercial proposal.

  • Licensing for VMware Cloud Foundation 9.0
  • VMware Cloud Foundation 9.x — Data Sheet
  • VMware vSphere Foundation — Data Sheet
  • Feature Comparison & Upgrade Paths: VCF and VVF
  • Counting Cores for VCF/VVF and TiBs for vSAN — Broadcom KB 313548
  • VCF 9.0 Product Subscription and Licensing
  • vCenter identity federation with Microsoft Entra ID
  • Performance Best Practices for VMware vSphere 8
  • Operations in VMware Cloud Foundation 9.0
  • Five reasons to upgrade from VVF to VCF

Editorial note

Names, brands, and products belong to their respective owners. The numerical capacity and performance examples were organized for didactic purposes and do not constitute a quote, lab benchmark, or licensing advice — final sizing and economics depend on the actual hardware, SLA, and commercial contract of each project.

EnQ Digital: Cloud • Data Center • Baremetal • Storage • Support • Security