Migration, Hybrid, and Multicloud

Google Cloud PCA — Section 11

Migration, Hybrid, and Multicloud

Goal: recognize how an organization should assess, plan, migrate, modernize, or operate workloads across on-premises, Google Cloud, VMware, edge, and other cloud environments. PCA questions usually test the migration path and tradeoff, not command syntax.

PCA lens: start with business goals and constraints: downtime tolerance, dependency mapping, licensing, data size, networking, compliance, operational skills, modernization appetite, and total cost. Then choose the least disruptive service that still meets the long-term architecture goal.

Fast Decision Matrix

Service and Concept Deep Dives

Google Cloud Migration Center

Why an org uses it: Migration Center gives organizations a planning and assessment layer before they start moving workloads. It helps inventory assets, assess current environments, estimate cloud spend, and build migration plans. This is the tool to use when the org is still discovering what exists and deciding the right migration path.

PCA trigger: Pick Migration Center when the scenario says assess an on-prem estate, discover servers/databases, estimate migration cost, plan migration waves, or build a migration business case.

Common trap: Do not jump straight to Migrate to VMs, GCVE, or GKE if the question says the organization does not yet know its inventory, dependencies, utilization, or cost profile.

Storage Transfer Service

Why an org uses it: Storage Transfer Service is for moving large datasets over the network into Cloud Storage from sources like other clouds, HTTP/S endpoints, or on-premises file systems. It is useful when data can be transferred online and the org wants scheduled, repeatable, managed transfers.

PCA trigger: Pick it for online data migration, recurring transfers, object storage movement, or cloud-to-cloud/on-prem-to-Cloud Storage transfers.

Common trap: Do not use it as the answer when the data set is too large for the available network window. That is where Transfer Appliance becomes the better PCA answer.

Transfer Appliance

Why an org uses it: Transfer Appliance is a physical data transfer option. An organization uses it when network upload would take too long, bandwidth is limited, or data movement needs to be performed offline for practical reasons.

PCA trigger: Pick it when the scenario mentions petabyte-scale data, limited bandwidth, transfer windows measured in months, or shipping a device being faster than uploading.

Common trap: Do not overuse it for normal cloud-to-cloud or ongoing transfers. It is for very large bulk migration, not continuous synchronization.

Database Migration Service / DMS

Why an org uses it: DMS gives organizations a managed path to migrate supported relational databases with lower downtime than dump-and-load approaches. It is best for database modernization into managed Google Cloud databases like Cloud SQL or AlloyDB.

PCA trigger: Pick it when the requirement is migration of MySQL, PostgreSQL, SQL Server, or similar supported relational databases to managed Google Cloud database services.

Common trap: Do not confuse DMS with Datastream. DMS is a database migration service. Datastream is CDC/replication for streaming change data into pipelines or analytics destinations.

Datastream

Why an org uses it: Datastream provides change data capture so organizations can replicate database changes continuously. It is often used for analytics replication into BigQuery, data pipelines, or keeping downstream systems updated while a source database continues changing.

PCA trigger: Pick Datastream when the scenario says CDC, continuous replication, streaming database changes, or real-time sync into analytics.

Common trap: Do not pick Datastream just because the word database appears. If the ask is a managed database cutover to Cloud SQL or AlloyDB, DMS is usually the cleaner answer.

Migrate to Virtual Machines

Why an org uses it: Migrate to Virtual Machines helps organizations move VM-based workloads into Compute Engine without fully rewriting the application. It is a rehost or lift-and-shift option when speed and lower application change are more important than modernization.

PCA trigger: Pick it when the org has VMware, AWS, Azure, or other VM workloads and wants to move them to Compute Engine quickly with limited refactoring.

Common trap: Do not pick it when the requirement says the org wants to modernize applications into containers or managed serverless. That points toward Migrate to Containers, GKE, or Cloud Run depending on the app.

Migrate to Containers

Why an org uses it: Migrate to Containers helps transform suitable VM workloads into containers that run on GKE. Organizations use it when they want modernization benefits without manually rebuilding every application from scratch.

PCA trigger: Pick it when the question says convert legacy VM workloads into containers, modernize to GKE, or reduce VM management while keeping application changes controlled.

Common trap: Do not assume every VM should be containerized. Stateful, tightly coupled, licensed, or hardware-dependent workloads may be bad candidates.

Google Cloud VMware Engine / GCVE

Why an org uses it: GCVE lets organizations run VMware environments natively on Google Cloud. It is valuable when a company has deep VMware investments, wants rapid data center exit, or needs vCenter/vSphere operational continuity while gaining cloud proximity.

PCA trigger: Pick GCVE when the scenario says VMware lift-and-shift, keep vCenter operations, avoid refactoring, data center evacuation, or migrate VMware workloads quickly.

Common trap: Do not pick GCVE as the most modern or cost-optimized answer when the app can move directly to managed services, GKE, Cloud Run, or Compute Engine.

Bare Metal Solution

Why an org uses it: Bare Metal Solution provides dedicated physical servers adjacent to Google Cloud regions for specialized workloads. Organizations use it for workloads that cannot be virtualized easily, require direct hardware access, or have strict licensing requirements such as some Oracle patterns.

PCA trigger: Pick it when the scenario says specialized legacy workload, direct low-level server access, strict Oracle/licensing constraint, or cannot run on standard virtualized cloud infrastructure.

Common trap: Be careful: this is specialized and not a default migration target. Google documentation also notes Bare Metal Solution availability is moving toward a specialized allowlist model, so treat it as an edge-case PCA answer.

Google Distributed Cloud / GDC

Why an org uses it: GDC extends Google Cloud infrastructure and Kubernetes-style operations into data centers and edge locations. Organizations use it when workloads must run locally due to latency, regulatory, sovereignty, disconnected, or operational requirements.

PCA trigger: Pick GDC when the scenario says on-premises cloud, edge deployment, local data processing, disconnected/air-gapped style requirements, or using Google Cloud capabilities outside Google regions.

Common trap: Do not confuse GDC with GCVE. GCVE is for VMware private clouds on Google Cloud. GDC is for extending Google Cloud-style infrastructure to on-premises or edge environments.

GKE Enterprise, Fleets, Config Sync, and Policy Controller

Why an org uses it: These services help organizations govern Kubernetes at enterprise scale. GKE Enterprise provides the broader enterprise Kubernetes capabilities, fleets group clusters, Config Sync keeps config consistent from source control, and Policy Controller enforces policy across clusters.

PCA trigger: Pick these when the scenario says many clusters, hybrid/multicloud Kubernetes, centralized policy, GitOps configuration, fleet-level governance, or consistent security controls across environments.

Common trap: Do not treat these as simple workload migration tools. They are governance and operations tools after the organization has chosen Kubernetes as the platform.

Hybrid and Multicloud Connectivity

Why an org uses it: Migration and hybrid architectures often depend on private, reliable connectivity between on-premises, Google Cloud, and other clouds. Organizations use VPN for encrypted internet-based tunnels, Interconnect for dedicated/private high-throughput links, Cloud Router for dynamic BGP, and NCC for hub-and-spoke network management.

PCA trigger: Pick hybrid networking concepts when the question mentions private connectivity, routing, DNS, firewall dependencies, latency, data transfer, or workloads that will span environments during migration.

Common trap: Do not design a migration without validating IP ranges, DNS, routing, firewall rules, bandwidth, and latency. PCA questions often hide the real issue in connectivity and dependencies.

Workload Disposition and Migration Planning

Why an org uses it: Organizations rarely migrate everything the same way. They decide whether each workload should be rehosted, replatformed, refactored, repurchased, retained, or retired. This avoids wasting effort moving unused systems and helps choose the right technical path.

PCA trigger: Pick this concept when the question asks for migration strategy, migration waves, prioritization, business value, or balancing speed against modernization.

Common trap: Do not assume the most cloud-native option is always correct. If the business requires speed, low change, or data center exit, rehost or GCVE may be better short-term answers.

Dependency Mapping, Network Planning, and Workload Testing

Why an org uses it: Before migration, organizations need to understand dependencies among apps, databases, identity systems, DNS, batch jobs, firewall rules, and network paths. They also need testing plans for cutover, rollback, performance, security, and user impact.

PCA trigger: Pick these concepts when the scenario asks what to do first, how to reduce migration risk, how to plan cutover, or why a migrated app failed after moving.

Common trap: Do not skip directly to production migration. The safer PCA answer is often assess, map dependencies, pilot, test, migrate in waves, then optimize.

Licensing Implications and Financial Impact / TCO

Why an org uses it: Commercial software licensing can drive architecture choices. VMware, Oracle, Microsoft, and other enterprise licenses may favor sole-tenant nodes, GCVE, Bare Metal Solution, or different migration timing. TCO also includes operations, support, network egress, storage growth, and refactoring cost.

PCA trigger: Pick this concept when the scenario mentions BYOL, Oracle RAC, VMware, compliance audits, budget justification, right-sizing, cost estimation, or business-case approval.

Common trap: Do not compare services on compute price alone. PCA questions often expect full migration economics: licensing, operations, downtime, networking, training, and modernization value.

High-Value Comparison Patterns

Common PCA Architecture Patterns

PCA Traps to Watch For

• If the org has not assessed its environment, the first answer is often Migration Center, inventory, dependency mapping, or a pilot, not the final migration tool.

• If downtime must be minimized for a database migration, think DMS/replication patterns instead of manual export/import.

• If the organization wants to keep VMware operations, choose GCVE. If they want to leave VMware behind, choose Compute Engine, GKE, Cloud Run, or managed services.

• If bandwidth is the blocker, do the math. Transfer Appliance can be the answer even though Storage Transfer Service sounds more cloud-native.

• If the workload requires edge/on-prem locality, GDC is more relevant than simply building everything in a Google Cloud region.

• If licensing is the constraint, the most modern service may be wrong. License terms can force sole-tenant, GCVE, Bare Metal Solution, or a phased plan.

• If the scenario says multicloud Kubernetes governance, think GKE Enterprise/fleets/policy/config, not just networking.

Mini Scenario Drill

Cram Summary

• Assessment first: Migration Center before picking target platforms when inventory/dependencies are unknown.

• Rehost: Migrate to VMs. Modernize to GKE: Migrate to Containers. Keep VMware: GCVE.

• Specialized physical legacy workloads: Bare Metal Solution. On-prem/edge Google Cloud-style infrastructure: GDC.

• Online data movement: Storage Transfer Service. Offline huge data movement: Transfer Appliance.

• Database migration: DMS. Continuous change capture/analytics replication: Datastream.

• Enterprise Kubernetes governance: GKE Enterprise + fleets + Config Sync + Policy Controller.

• Migration planning must include dependencies, networking, DNS, security, testing, rollback, licensing, and TCO.