Request for Review and Guidance on Data Center Migration Network Design

We are currently planning a data center migration and would appreciate your expert advice on the proposed network architecture and migration approach. Existing Environment We currently host three separate customer environments. Each customer environment operates on its own IP address range and is fully segregated from the others. Each customer has a dedicated firewall. The customer environments are being migrated from the existing data center to a new facility. New Data Center Requirements We own a public IPv4 /24 block that can be advertised through our Internet Service Provider(s). The new data center will utilize this single /24 public IP block as the primary address space available for customer connectivity. During the migration, the existing customer IP ranges will not be retained. Instead, each customer environment will need to be allocated address space from the new /24 block. We need to maintain complete segregation between customer environments while sharing common data center infrastructure where appropriate. We prefer to use static routing within our environment and do not intend to run dynamic routing protocols internally. The ISP may use dynamic routing protocols on their side for Internet redundancy and route advertisement. Each customer will continue to use its own dedicated firewall. Firewalls are currently deployed as standalone units (not HA pairs). We have a redundant pair of core switches. Access switches may be shared between customers or dedicated, depending on the recommended architecture. High availability, operational simplicity, and future scalability are important design goals. Design Options Under Consideration Option 1 – Layer 2 Core Core switches operate primarily as Layer 2 aggregation devices. Routing is performed by the customer firewalls. VLANs are extended through the core layer. Public IP subnets from the /24 allocation are presented directly to customer firewalls. Option 2 – Layer 3 Core Core switches provide Layer 3 gateway functionality. Customer VLANs are routed at the core layer. Static routing is used between the core switches and customer firewalls. Public IP allocations are segmented and routed to the respective customer environments. Areas Where We Would Appreciate Guidance Which architecture would you recommend for this migration: Layer 2 core or Layer 3 core? What are the operational and scalability implications of each design? What is the recommended method for allocating and managing multiple customer networks from a single public /24 block? Would you recommend assigning public IP subnets directly to customer firewalls or using routed networks behind the firewalls? Is a shared access-switch infrastructure with VLAN segregation considered a best practice for this type of environment? What is the recommended approach for Internet redundancy while maintaining static routing internally? Are there any concerns regarding spanning-tree dependencies, failure domains, asymmetric routing, or future growth? What migration strategy would you recommend when moving existing customer environments from their current IP ranges to the new shared /24 address pool? Existing Compute and Storage Environment In addition to the network migration, we are also planning to migrate the existing compute and storage infrastructure to the new data center. Current environment details: Approximately 9–10 physical servers. Infrastructure includes VMware ESXi hosts, storage systems, QNAP appliances, and supporting hardware. Approximately 16 virtual machines are currently running across the environment. We currently support 4–5 customers hosted within the data center. Each customer has its own dedicated LAN environment and security boundaries. Customer workloads may consist of application servers, database servers, file services, and other business-critical systems. The migration will involve relocating both the underlying infrastructure and the associated workloads to the new data center. Additional Guidance Requested From a server and workload migration perspective, we would appreciate recommendations on the following: What is the recommended approach for migrating the physical servers, storage systems, and virtual workloads to the new data center? Should the migration be performed as a physical relocation of equipment, or would a rebuild/redeployment approach be more appropriate? What dependencies should be assessed between the server, storage, firewall, and network migrations? What is the recommended migration sequence for network, storage, virtualization hosts, and customer workloads? How should customer downtime be minimized during the migration process? What validation and testing activities should be completed before, during, and after the migration? Are there any specific considerations for VMware ESXi hosts, shared storage platforms, or QNAP systems during a data center migration? What rollback strategy would you recommend in the event of migration issues? What are the key risks and common pitfalls to avoid during the migration of customer workloads and supporting infrastructure? Based on the environment described, would you recommend a phased migration by customer, by application, or by infrastructure component? Migration Scope Summary The migration scope currently includes: Dual core switching infrastructure. Internet connectivity and public IP allocation redesign. Customer firewall connectivity and segregation. Multiple customer LAN environments. Approximately 9–10 physical servers. VMware ESXi infrastructure. Storage systems including QNAP appliances. Any recommendations, sample topologies, or best-practice guidance would be greatly appreciated.