Choosing a Cloud Platform

Choosing a Cloud Platform#

Northwestern has institutional agreements with three major cloud providers:

  • Amazon Web Services (AWS) : The most widely adopted cloud platform, offering the broadest range of compute, storage, and machine learning services

  • Microsoft Azure : Microsoft’s cloud platform, well integrated with Windows-based tools and widely used for data analysis and collaboration

  • Google Cloud Platform (GCP) : Google’s cloud platform, known for its data analytics, genomics, and AI and machine learning capabilities

Each platform covers the core needs of most research projects, including virtual machine hosting, scalable storage, and machine learning pipelines, but differs in pricing, available tools, and areas of strength.

See Requesting an Account for the benefits of a Northwestern-affiliated account.

This page will help you decide which cloud provider to use for your research. See the following sections for advice on picking a platform.

Start With One Platform

Starting with a single provider is strongly recommended for teams new to cloud. Using multiple platforms increases complexity, extends the learning curve, and complicates cost management.

1. Check for Restrictions#

Not every researcher has a free choice of platform. Check whether either of the following applies to your project before working through the questions below:

  • Legal or contractual agreements: Research data is often subject to a variety of security requirements from funding agencies, data use agreements (DUAs), or other contractual requirements. These agreements may stipulate that you must use a specific cloud platform or that you cannot use cloud resources. Make sure that you understand these agreements before using cloud resources. If you’re not sure, request a Research Cloud Consultation .

  • Privacy concerns for human subjects: If your project involves personally identifiable information (PII) or protected health information (PHI) about human subjects, not all cloud platforms meet the security requirements necessary to protect this highly sensitive data. Northwestern maintains a Business Associate Agreement (BAA) with both AWS and Azure; see Working with Sensitive or Regulated Data for details. If you’re unsure if your data is appropriate to use in the cloud, request a Research Data Management Consultation .

Feinberg School of Medicine (FSM) Researchers

FSM maintains its own cloud and data storage policies. Before choosing a platform, contact FSM IT to determine which platform is appropriate for your project.

2. Use Previous Experience#

If you or your collaborators have already used AWS, Azure, or GCP, staying with that platform is usually the fastest path forward.

3. Go to Where the Tools Are#

If your research depends on a capability unique to one platform, including artificial intelligence and machine learning (AI/ML), that need should carry more weight than general familiarity.

For example:

  • AWS has the largest third-party ecosystem and the broadest overall service catalog, so niche or specialized tools are often available there first. For AI and machine learning, Amazon Bedrock provides access to Claude and other foundation models.

  • Azure includes Microsoft Foundry for access to OpenAI models, along with Azure Batch and Synapse Analytics for data processing and research toolkits for genomics and imaging.

  • GCP includes Vertex AI for access to Google Gemini models and BigQuery for large-scale data analysis, and is well suited for AI-heavy and data-intensive research that benefits from reproducibility.

Northwestern’s AI Guidance

Before adopting any AI tool or service, review Northwestern’s AI Tools and Use Guidance to confirm it meets University requirements for your data.

No Preference?#

For researchers approaching this decision with no previous cloud experience and are using a standard workload, such as:

  • running an analysis environment or application on a virtual machine

  • hosting a structured dataset in a database

  • storing and archiving research data

We recommend AWS because:

  • AWS has the largest catalog of services and most extensive documentation

  • Northwestern IT and Amazon’s enterprise architects provide the best level of support for users of the 3 platforms

  • Northwestern’s Cloud Community of Practice is largely composed of AWS users

The combination of diverse services, good documentation and local knowledge makes this our recommended option.

Still Not Sure?#

If the advice above doesn’t help you make a decision, request a Research Cloud Consult to talk through the tradeoffs with RCDS.

Choosing a platform now does not lock you in. If your project’s needs change later, you can request an account on a different platform.

Next Steps#

Once you’ve settled on a platform using the questions above, the next step is setting up your account. Requesting an Account covers the benefits of a Northwestern-affiliated account and walks through the request process, including estimating costs and submitting your request.