The DTI team will update this page when information about these resources becomes available and policies surrounding their use for the DTI are clear.
Details to come; please contact help@c3dti.ai for assistance.
Details to come; please contact help@c3dti.ai for assistance.
Azure
Microsoft provides a generous (but limited) annual gift of Azure credits to help support C3.ai DTI-sponsored research.
If you're interesting in using Azure resources for a current C3.ai DTI project, please contact the DTI team at help@c3dti.ai.
Microsoft Azure infrastructure can support your program of compliant stewardship of data in most scenarios. If your data requires special attention, please let us know when asking for Azure resources.
When you're logged in to the Azure portal, ensure that you've selected the appropriate Account and Directory for your Subscription. You can set these by clicking on your personal identifier in the upper right of the portal:
Azure Portal dashboards are called "Blades." Here are links to the Azure Portal and to some Blades that may help you manage your Azure use:
If you are one of your research group's Azure coordinators, you can grant fine-grained access to your Azure Subscription and Resources to other research group members via the Access control (IAM) options in the Subscriptions Blade for your subscription.
Your research group's Azure coordinator(s) can view DTI Assigned credit and Consumed credit for your group's subscription in the Education | Courses Blade.
For C3.ai DTI research groups with an Azure sponsorship, your team's designated Azure manager(s) can see your sponsorship's assigned credit and credit consumed via this link, once you are signed in to your Azure context:
https://portal.azure.com/#blade/Microsoft_Azure_Education/EducationMenuBlade/classrooms
The DTI team can can access (modestly) more-detailed daily reporting by Azure resource category; please contact help@c3dti.ai for assistance with spend management.
Details to come; please contact help@c3dti.ai for assistance.
Azure VM offerings are grouped into series, with each series optimized for distinct purposes – e.g., GPU, data protection, hyper-threading, compute, memory, and so on. Virtual Machine series descriptions are available here:
https://azure.microsoft.com/en-us/pricing/details/virtual-machines/series/
To browse more-detailed specifications of Azure Virtual Machines, see this page:
https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/
Note that both pricing and VM size availability vary by region.
Read about Data, OS and Temporary disks, including notes regarding data persistence, on this Azure documentation page:
https://docs.microsoft.com/en-us/azure/virtual-machines/managed-disks-overview#disk-roles
There are three distinct ways to manage (create, destroy, inspect, modify) your Azure resources:
You can manage Azure resources via the Azure Web portal:
The Web portal provides a point-and-click GUI that offers quick and easy access. Note that other resource management methods (whether used alone or in combination with Web portal use) can provide better reproducibility, flexibility, and higher efficiency, particularly for complex and larger-scale deployments.
There are several ways to engage Azure via the command line. Documentation for access, installation and use is available here:
https://docs.microsoft.com/cli/azure/
Note that you can install and run the Azure command line tool on a local machine or access it in a Web browser via Azure Cloud Shell:
https://docs.microsoft.com/azure/cloud-shell/
Cloud Shell also offers additional, complementary features; see documentation link immediately above.
You can develop an application to interact with Azure to manage resources on your behalf via HTTP operations with Azure API service endpoints. More information is available in this documentation:
https://docs.microsoft.com/rest/api/azure/
Note that applications require configuration of an authentication token that represents appropriate role-based permissions. See documentation above or contact the DTI team at help@c3dti.ai.