Migration Service
In the ZCF architecture, Migration Service is used to migrate the operating systems and data of source VMs to ZStack Cloud. ZStack Cloud provides the entry point for the service and the target compute, storage, and network resources. ZMigrate organizes source platforms, data gateways, and migration tasks so that resource preparation, data synchronization, isolated testing, and business cutover can proceed in stages.
Migration Service Architecture
In the ZStack Cloud-integrated model, Migration Service consists of ZStack Cloud, ZMigrate, a source environment, a data gateway, target cloud resources, and migration supporting images. ZStack Cloud provides the service entry point and target resources, ZMigrate controls and orchestrates migration tasks, and the data gateway connects source reads with target writes.
Architecture Elements
Migration Service contains two types of logical relationship: management interaction and migration data transfer. Management interaction discovers source resources, stores target settings, starts tasks, and reports progress. Migration data transfer synchronizes source disk data to the target. An accessible management entry point indicates only that the control connection is available; data-transfer capability also depends on the actual path, sustained throughput, and target write capability.
| Architecture Element | Primary Responsibility |
|---|---|
| ZStack Cloud management entry point | Provides access to ZMigrate, target-resource settings, and task status. |
| ZMigrate management service | Organizes source platforms, data gateways, migration tasks, and execution status. |
| Source environment | Provides VMs to be migrated, their settings, and disk data. |
| Data gateway | Connects the platforms and data-transfer stages required for migration. |
| Target cloud resources | Host migration instances, target disks, and the final business VMs. |
| Migration supporting images | Provide boot environments required by the migration management service and Linux and Windows migration. |
Architecture Relationships and Data Paths
Migration Service uses management and data paths together to control migration and process data, while the target hosts the migration result.
- Management path: Handles resource discovery, target settings, task control, and status reporting. An accessible management page indicates only that the control connection is available.
- Data path: Handles source reads, gateway processing, network transfer, and target writes. Whether the synchronization path can transfer data depends on the actual network path, sustained throughput, and target write capability.
- Target hosting: Hosts migration instances, target disks, and final business VMs. After data synchronization, system conversion adapts the target operating system to the runtime environment; final business VMs run on ZStack Cloud compute, primary storage, and Layer 3 networking.
- Mode-specific differences: The exact combination of source reads, target writes, and temporary instances depends on the migration mode and version.
Migration Mechanisms
Migration Service combines data replication, target-resource mapping, and operating system conversion to enable source VM systems and data to run in a ZStack Cloud target environment. Data readiness, target-side operation, and business availability are separate states.
Data Replication and Runtime Adaptation
Data replication transfers source disk data to target storage, while runtime adaptation converts source VM runtime settings into the target resource model.
- Replication modes: Full replication covers the complete data scope, while incremental replication transfers changes since the previous synchronization.
- Change scope: After initial synchronization is complete, the system identifies the scope of source-data changes through scanning and comparison.
- Resource mapping: The target resource model includes runtime location, primary storage, disk bus, CPU mode, Layer 3 network, and IP address assignment.
- System conversion: System conversion adapts the migrated operating system to the target environment. Driver adaptation, tool installation, and network results depend on the operating system version.
Runtime Isolation and Data States
Migration Service maintains separate target-side runtime environments and data states to distinguish migration data, validation instances, and business-hosting VMs.
- Isolated runtime environment: Test instances use independent compute, storage, and network conditions and are isolated from the production network to prevent network-identity conflicts involving the same IP address, MAC address, or application identity.
- Data-ready state: Completed data replication means that corresponding disk data is present at the target; it does not mean that the application workload is available at the target.
- Snapshot recovery scope: A target snapshot records only the target-side state at the time of creation; recovery does not include changes made after that point.
- Unidirectional data flow: Migration data flows from the source to the target. Target-side business data is not returned automatically to the source; concurrent writes at both source and target can cause data divergence.
Migration Dependencies and Performance Factors
Migration feasibility and sustained transfer performance are jointly determined by source-side, target-side, network, and business-runtime characteristics. VM count is only one factor in migration scale; operating systems, disks, networks, and application dependencies also affect migration implementation.
Source and Target Dependencies
| Factor | Relevant Information | Impact on Migration Implementation |
|---|---|---|
| Source environment | Source platform version, connection method, and account permissions. | Affects resource discovery and data access. |
| Operating system | Distribution, version, architecture, boot method, and key drivers. | Affects system conversion and target-runtime compatibility. |
| Disks | Count, provisioned capacity, actual data volume, and special disk settings. | Affects transfer workload and target storage requirements. |
| Network | IP address, MAC address, gateway, DNS, routes, access control, and bindings. | Affects address retention, network connectivity, and application access. |
| Application | Databases, domain services, external dependencies, write suspension, and startup order. | Affects business consistency and service recovery. |
| Runtime characteristics | Data-change rate, peak load, and available time window. | Affects data volume to synchronize and cutover duration. |
A discoverable platform or a VM that can be listed does not establish that its operating system or disk combination is supported. Actual support conditions are determined by the ZStack Cloud and ZMigrate versions in use.
Network and Transfer Performance
Migration involves four network uses: management access, data synchronization, migration testing, and production business traffic. The data path depends on routes, access control, name resolution, address allocation, and bidirectional communication. An accessible management page does not establish that the data-synchronization path is available. When retaining a source IP address or MAC address, the target network also requires the corresponding address and Layer 2 conditions to avoid concurrent source and target connections with the same identity.
Migration consumes the ZMigrate management service and data gateway, migration supporting images, target VM disks, test instances, snapshots, and temporary resources during synchronization. Storage consumption is determined jointly by provisioned capacity, actual data volume, and snapshot changes.
Initial synchronization duration can be estimated from the relationship between data volume to transfer and the end-to-end effective transfer rate. The effective rate is constrained by source reads, gateway processing, network transfer, and target writes. Data-synchronization duration does not equal business-cutover duration, which is also affected by system conversion, target startup, and application recovery.
Migration State and Consistency Boundaries
Migration task state and execution logs reflect migration phases and fault domains. Disk-data synchronization and business consistency describe migration results at the storage and application layers, respectively. Migration Service provides task execution information but does not replace the consistency mechanisms of business systems.
Task State and Diagnostic Information
Task state identifies migration phases such as data synchronization, testing, and business cutover. Execution logs record information from source access, target-instance preparation, disk reads, target writes, system conversion, and startup. Together, state and logs present migration results and fault locations.
Data and Business Consistency Boundaries
Disk-data synchronization provides the storage-layer data foundation for migration, while application consistency also depends on a business system's own write-suspension, flush, and recovery mechanisms. Migration results for databases, multi-node applications, domain services, and similar systems are affected by their internal dependencies, data-consistency model, and startup order.
A business system spanning multiple VMs can contain components that must suspend writes together, can migrate earlier, or must start later. Migration Service provides the target runtime environment and task state, while business consistency and final business results remain determined by application-level mechanisms.