VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

Yesterday’s dispositions included a category that deserves its own post: the VMware estate that must move, datacenter exit, hardware refresh avoidance, licensing pressure, but cannot be replatformed on the timeline. Azure VMware Solution runs actual vSphere, vSAN, and NSX on dedicated Azure bare metal, and the strategic honesty about what it is for matters as much as the mechanics.

What You Get and What It Costs

A private cloud is a cluster of dedicated nodes (three minimum, scaling to large clusters) running the full VMware stack, managed by Microsoft at the infrastructure layer, hardware, hypervisor lifecycle, while your team keeps vCenter, the operational tooling, and the muscle memory intact, which is the entire value proposition: migration measured in weeks because nothing about the workloads changes, not the OS, not the IP addresses, not the runbooks. The economics are node based and the FinOps posts apply with force: reserved instances on the nodes are the difference between defensible and painful, the Azure Hybrid Benefit and extended security update angles for the Windows and SQL guests stack on top, and right sizing means cluster count, so the Migrate assessment from yesterday, with its AVS specific sizing against vSAN policies and compression, is the input that keeps you from buying a node too many.

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama

HCX and the Networking Reality

VMware HCX is the migration machinery: site pairing between on premises and the private cloud, then bulk migration in scheduled waves, live vMotion for the tender workloads, and, the feature that collapses timelines, layer 2 network extension, the on premises VLAN stretched into AVS so machines move without re IP while dependencies straddle both sides mid wave. Treat extension as scaffolding with a demolition date: stretched networks are a dependency on the link and a blast radius spanning sites, so each wave completes with cutting the extension over to native NSX segments. The Azure side wiring follows the series’ networking arc: the private cloud connects to your hub over ExpressRoute (with Global Reach or transit options linking on premises, AVS, and Azure VNets), NSX handles the east west micro segmentation inside, and the egress, DNS, and inspection patterns from the September posts apply to VMware guests exactly as to native VMs, one network estate, two hypervisors.

Integration and the Exit Question

The quiet advantage over the old datacenter is proximity: AVS guests sit milliseconds from the platform services this series covers, so databases replatform to the managed engines while app tiers stay put, backups land in Azure Backup and files on Azure NetApp Files (attached as datastores for the storage hungry clusters, often cheaper than adding nodes for capacity), monitoring flows through Arc enabled onboarding of the guests into the Update Manager, Defender, and Log Analytics machinery, and new capabilities get built cloud native beside the estate rather than inside it. That points at the strategic honesty: AVS is either a permanent operating model for a VMware committed organization, a legitimate choice, priced knowingly, or a decompression chamber, the estate moves fast, the datacenter closes, and then the Migrate dispositions resume at leisure, workloads replatforming out of the private cloud year by year until the cluster count steps down. Both are valid; the failure mode is not choosing, running AVS by inertia at native VMware cost plus cloud premium while calling it transformation. Write the intended trajectory into the business case, revisit it at the Well Architected cadence, and the datacenter that moved without changing either earns its keep or finishes its change.

Cheers
Osama