Task 4.3: Design migrations

• AThe Azure Migrate appliance with agentless dependency analysis

  Correct

  The appliance discovers the VMware VMs and collects connection data without agents, so you can group servers by dependency.

• BThe Azure Site Recovery Deployment Planner run against vCenter

  Incorrect

  The planner estimates replication bandwidth and storage for disaster recovery; it doesn't map server dependencies.

• CAgent-based dependency analysis with agents on every VM

  Incorrect

  Agent-based analysis requires installing agents on the servers, which isn't allowed.

• DA spreadsheet built from the CMDB and interviews with app owners

  Incorrect

  Manual inventories miss undocumented connections and don't feed the assessments.

Azure Migrate discovers VMware VMs through an appliance and offers agentless dependency analysis, which captures TCP connection data from the servers so you can visualize dependencies and create groups for assessment and migration waves.

• AAn offline migration with Azure Database Migration Service

  Incorrect

  Offline migration takes the app down for the whole copy, and there's no readable copy in Azure beforehand.

• BA BACPAC export from SQL Server imported into Azure

  Incorrect

  A BACPAC is a one-time copy that is stale as soon as it's imported, and large imports take hours.

• CActive geo-replication from SQL Server to the managed instance

  Incorrect

  Active geo-replication works between Azure SQL databases, not from an on-premises SQL Server.

• DThe Managed Instance link, followed by a planned failover to Azure

  Correct

  The link replicates in near real time to a readable replica in SQL Managed Instance and supports a planned failover for cutover.

The Managed Instance link uses distributed availability groups to replicate a SQL Server database to SQL Managed Instance in near real time. The replica can serve read-only workloads, and a planned failover migrates with minimal downtime.

• ARehost

  Correct

  Rehosting moves the workload as it is to Azure IaaS, which fits a hard deadline and no code changes.

• BRefactor

  Incorrect

  Refactoring changes code to use cloud services, which the deadline and the no-change rule rule out for now.

• CRearchitect

  Incorrect

  Rearchitecting redesigns the app's architecture, the most effort of the choices here.

• DRetain

  Incorrect

  Retaining keeps the app where it is, but the datacenter is closing.

The Cloud Adoption Framework lists strategies such as retire, retain, rehost, replatform, refactor, rearchitect, rebuild and replace. Rehost moves a workload with no code changes and suits deadline-driven moves such as a datacenter exit.

• AAzure Data Box devices shipped to the lab

  Correct

  Data Box moves large data offline on devices of up to 120 TB or 525 TB usable each, avoiding the slow shared link.

• BAzCopy over the existing internet link

  Incorrect

  At 200 Mbps, 350 TB would take about five months at full speed, well past the deadline.

• CAzure File Sync with cloud tiering

  Incorrect

  File Sync still uploads every byte over the network link, which is too slow here.

• DAzure Storage Mover over the internet link

  Incorrect

  Storage Mover copies over the network too, so the link remains the bottleneck.

Microsoft's data migration guidance chooses offline transfer when data size and available bandwidth make an online copy too slow. Data Box devices hold up to 120 TB or 525 TB usable; later changes can follow online.

• AAs-is on-premises sizing

  Incorrect

  As-is sizing copies the on-premises configuration, so the oversized VMs stay oversized.

• BPerformance-based sizing with a one-month performance history

  Correct

  Performance-based sizing uses collected utilization data, with a chosen history and percentile, to right-size targets.

• CSizing set to the largest VM size in each series

  Incorrect

  Picking the largest size ignores utilization and raises cost.

• DSizing based on the reserved instance term

  Incorrect

  Reservation terms affect pricing, not which size fits the workload.

Azure Migrate assessments offer as-is on-premises sizing and performance-based sizing. Performance-based sizing calculates compute and storage recommendations from collected utilization data, using the performance history and percentile utilization you set.

• AA backup and restore with the native PostgreSQL client tools during a weekend outage

  Incorrect

  Exporting and restoring 2 TB means a long outage, not a few minutes.

• BAzure Data Box to ship a backup of the database files

  Incorrect

  An offline copy is stale on arrival and still needs a long catch-up.

• CAzure Site Recovery replication of the database server

  Incorrect

  Site Recovery would move the server as a VM, not into the flexible server PaaS target.

• DThe migration service in Azure Database for PostgreSQL in online mode

  Correct

  The built-in migration service supports online migrations that keep changes flowing until a short cutover.

Azure Database for PostgreSQL flexible server includes a migration service that moves databases from on-premises, VMs and other clouds, with offline and online modes; online mode keeps syncing changes so downtime is limited to the cutover.

• AAzure Storage Mover

  Correct

  Storage Mover is a managed migration service with source-side agents and centrally managed, monitored migration jobs.

• BAzCopy run from cron jobs on several Linux servers

  Incorrect

  Scripts on servers can copy the data, but there's no central job management or reporting.

• CAzure File Sync

  Incorrect

  File Sync syncs Windows file servers with Azure file shares; it doesn't move NFS shares into Blob containers.

• DAzure Data Box devices

  Incorrect

  With a 10 Gbps link and three months, an online, managed copy is practical and avoids shipping devices.

Azure Storage Mover is a hybrid migration service for moving file shares to Azure Storage. Agents run near the source, and migration projects and jobs are defined and monitored centrally in Azure.

• AAzure Site Recovery replication of the 40 IIS servers

  Incorrect

  Site Recovery moves the servers as VMs, which keeps IIS to manage rather than moving apps to App Service.

• BRewriting every app as Azure Functions

  Incorrect

  Rewriting 150 apps is a rebuild, far more work than moving them to App Service.

• CAzure Migrate web app discovery and assessment, then bulk migration

  Correct

  Azure Migrate discovers and assesses ASP.NET apps on IIS at scale without agents and supports bulk migration to App Service.

• DCopying each site's files to Blob Storage with static website hosting

  Incorrect

  Static website hosting can't run server-side ASP.NET code.

Azure Migrate offers at-scale, agentless discovery and assessment of ASP.NET web apps on IIS and bulk migration of multiple apps to App Service.