Migrate from Azure Lab Services¶
Azure Lab Services retires June 28, 2027. RosettaHub has been serving universities since 2016.
Azure Lab Services Retirement¶
Microsoft announced on July 18, 2024 that Azure Lab Services will be retired on June 28, 2027. New customers can no longer sign up. Existing customers retain access until the retirement date, after which lab accounts, lab plans, and labs will no longer be available.
Microsoft recommends migrating to Azure Virtual Desktop, Windows 365, Azure DevTest Labs, or Microsoft Dev Box -- all of which are Azure-only services that do not provide the managed virtual lab experience that educators rely on.
RosettaHub is a purpose-built alternative that goes well beyond what Azure Lab Services offered.
Why RosettaHub¶
RosettaHub is a multi-cloud Supercloud platform that has been delivering virtual labs to universities since 2016. It supports virtual machines, containers, Kubernetes clusters, Big Data clusters, HPC clusters, and any infrastructure based on IaC templates (Terraform, CloudFormation) -- as well as sandboxed federated access to native cloud consoles and tools.
Migrating from Azure Lab Services to RosettaHub gives your institution:
- More flexibility -- any instance type, any image size, any cloud provider
- Lower costs -- per-second billing at actual cloud prices, not 2-5x markups
- Better user experience -- single portal, no publishing delays, stop/start VMs without data loss
- Multi-cloud freedom -- AWS, Azure, GCP, OVH, OCI, and OpenStack from one platform
Feature Comparison¶
Platform Scope¶
| Capability | Azure Lab Services | RosettaHub |
|---|---|---|
| Description | Managed Azure-only service for provisioning templated VMs for hands-on labs | Multi-cloud SaaS platform for VMs, containers, Kubernetes, Big Data, HPC, IaC templates, and sandboxed cloud console access |
| Cloud support | Azure only | AWS, Azure, GCP, OVH, OCI, and OpenStack |
| IDE support | No | Yes -- all IDEs |
| Notebook support | No | Yes -- Jupyter, RStudio, Zeppelin, and more |
| Container-based labs | No | Yes -- start from container images with Jupyter, RStudio, VS Code pre-installed |
| Custom container images | No | Yes -- build and share custom Docker images |
| Direct sandboxed access to cloud services | No | Yes -- federated console access with sandbox enforcement |
| Real-time collaboration | No | Yes -- share formations, images, and storage with users, groups, and organizations |
Flexibility¶
| Capability | Azure Lab Services | RosettaHub |
|---|---|---|
| Instance types | Limited to 8 fixed configurations | Any instance type from any supported cloud provider |
| Image size limit | 128 GB | No limit |
| Stop and restart VMs | No -- stopping a lab terminates the VM and loses all disk data. Microsoft recommends removing the shutdown button from the Windows Start menu. | Yes -- VMs can be stopped and restarted without data loss |
| Spot/preemptible instances | No | Yes -- up to 90% cheaper than on-demand, with automatic spot recovery |
| Cloud storage integration | Manual configuration only | Auto-mount dedicated object or file storage per student in the lab setup |
| Run labs from shared images | No -- must create a new image first | Yes -- launch directly from shared formations without publishing |
Cost and Billing¶
| Capability | Azure Lab Services | RosettaHub |
|---|---|---|
| Billing granularity | Per hour. If you stop after 20 minutes and restart for 30 minutes, you are billed for 2 full hours. | Per second. You pay for the exact time your VM runs. |
| Infrastructure cost | Labs are priced at 2-5x the underlying Azure infrastructure cost | Labs run at actual cloud provider prices -- no markup on infrastructure |
| License model | Pay-as-you-go -- license costs scale with lab usage hours | Fixed annual per-user subscription -- predictable costs that don't increase with usage |
| Real-time cost tracking | No -- costs appear with ~24 hour delay from Azure billing | Yes -- costs are tracked in real time per user, per cloud account |
| Hard budget limits | No -- managers can set maximum lab hours, but no hard spend cap | Yes -- hard budget limits enforced in real time; launches blocked when budgets are exhausted |
Cost Examples¶
The table below compares Azure Lab Services hourly rates with equivalent AWS on-demand instances managed through RosettaHub. RosettaHub charges no markup on infrastructure -- you pay the cloud provider directly.
| Configuration | Azure Lab Services | RosettaHub (AWS) | Saving |
|---|---|---|---|
| Small -- 2 cores, 3.5 GB RAM | $0.20/hr | $0.046/hr (t3.medium) | 77% |
| Medium -- 4 cores, 7 GB RAM | $0.42/hr | $0.092/hr (t3.large) | 78% |
| Medium (nested) -- 4 cores, 16 GB RAM | $0.55/hr | $0.184/hr (t3.xlarge) | 67% |
| Large -- 8 cores, 16 GB RAM | $0.70/hr | $0.232/hr (m7a.xlarge) | 67% |
| Large (nested) -- 8 cores, 32 GB RAM | $0.84/hr | $0.333/hr (t3.2xlarge) | 60% |
| Small GPU (compute) -- 6 cores, 56 GB RAM | $1.39/hr | $0.79/hr (g3s.xlarge) | 43% |
| Small GPU (visualization) -- 6 cores, 56 GB RAM | $1.60/hr | $1.00/hr (g5.xlarge) | 38% |
| Medium GPU (visualization) -- 12 cores, 112 GB RAM | $4.08/hr | $1.81/hr (g5.4xlarge) | 56% |
Spot instances for even greater savings
With RosettaHub's spot instance support, these costs can be reduced by a further 60-90%. A $0.092/hr t3.large instance costs as little as $0.009/hr on spot pricing.
Cost Controls¶
| Capability | Azure Lab Services | RosettaHub |
|---|---|---|
| Automated shutdown on idle | Yes | Yes |
| Scheduled shutdowns | Yes | Yes |
| Stop idle VMs | Yes (terminates, loses data) | Yes (stops, preserves data -- VM can be restarted) |
| Spot instances | No | Yes -- with three automatic recovery modes: snapshot disk, hibernate, or launch replacement VM |
| Per-student hard budgets | No | Yes -- enforced in real time at every level of the organization hierarchy |
User Experience¶
| Capability | Azure Lab Services | RosettaHub |
|---|---|---|
| Single management portal | No -- administrators switch between the Azure portal (lab plans) and the Azure Lab portal (labs) | Yes -- one portal for everything |
| Time to first lab | ~15 minutes -- labs must be "published" (image creation) before students can use them | Immediate -- labs launch directly, no publishing step required |
| Quota management | Labs share a resource group quota; educators often hit limits when scaling | Each user gets a dedicated cloud account with independent quotas |
| User/group/organization management | No organizational mapping -- users managed per-lab via Active Directory | Full organization hierarchy -- departments, courses, groups with delegated management |
| Share lab templates | No | Yes -- share formations with users, groups, or entire organizations in a few clicks |
| Share labs between organizations | No | Yes -- cross-organizational sharing via URLs or marketplaces |
| Invite users to labs | Yes | Yes |
| Lab schedules | Yes | Yes |
Spot Instance Recovery¶
RosettaHub's spot instance support is particularly relevant for education, where labs are often fault-tolerant and the 60-90% cost savings can dramatically stretch a department's budget. When a cloud provider reclaims a spot instance, RosettaHub automatically handles the interruption with three recovery modes:
- Snapshot on Termination -- automatically creates an image of the VM's disk before it is reclaimed, preserving all student work
- Hibernate -- saves the full machine state (memory + disk) so the VM resumes exactly where the student left off when capacity returns
- Auto-replace -- automatically launches a replacement VM with similar characteristics
Migration Path¶
Step 1: Set Up Your RosettaHub Organization¶
- Register your organization on RosettaHub
- Connect one or more cloud accounts (AWS, Azure, GCP, or others)
- Configure your organization hierarchy to match your institutional structure (departments, courses, groups)
- Set budgets at each level
Step 2: Recreate Your Labs as Formations¶
For each Azure Lab Services lab plan:
- Create a formation that matches the lab's software stack. Choose the formation type that fits:
- Docker Formation for container-based labs (Jupyter, RStudio, VS Code)
- Cloud Formation for custom VM images
- Virtual Lab Formation for full desktop environments (Linux or Windows)
- Test the formation by launching it yourself
- Snapshot any customizations as reusable images
Start from existing formations
RosettaHub's marketplace includes pre-built formations for common lab types. Clone one and customize it rather than building from scratch. See the workspace and lab tutorials for step-by-step guides.
Step 3: Onboard Students¶
- Configure SSO integration with your institutional identity provider
- Register students individually or in batch via Excel upload
- Share formations with the appropriate course organizations
- Students log in and launch labs with a single click -- no cloud accounts or credentials needed
Step 4: Decommission Azure Lab Services¶
Once your labs are running on RosettaHub:
- Export any data or images from Azure Lab Services that you need to preserve
- Stop and delete Azure lab plans
- Cancel Azure Lab Services subscriptions before the June 28, 2027 retirement date
Next Steps¶
- For Education -- full overview of RosettaLabs capabilities
- Quick Start -- set up your RosettaHub account
- Formations -- learn how to create lab templates
- Tutorials -- step-by-step walkthroughs for workspaces, labs, and more
Need help migrating?
Contact us for a guided migration from Azure Lab Services. We can help you design your formation templates, configure your organization hierarchy, and onboard your first course.