End User Access
In the step, you will take on the persona of an end-user. Upon login as the end user, you will be directed to the End User focused Developer Hub in the Rafay Platform.
The Developer Hub is an extremely limited portal which allows the user to only see what they’re supposed to see that they cannot see what else their colleagues are doing.
Upon login, the end user's experience in the Developer Hub will look similar to the image below.
Create Workspace¶
Workspaces allow users to organize and manage resources. In our exercise, we will create two workspaces: "dev" and "qa". When we create instances and notebooks, we will launch them in one of these workspaces.
Workspaces are also access controlled and therefore aid with user collaboration. For example, the end user can invite collaborator "Sally" to the dev workspace and "John" to the qa workspace. They can also invite these collaborators with read/write privileges or limit their access to read only privileges. Let's create our workspaces now
- In the Workspaces menu, click on New Workspace
- Provide a unique name (e.g. dev) and Save
In a similar manner, create the "qa" workspace as well. Once both are created, you should see something like the image below.
Compute with GPUs¶
Now, you are ready to launch a compute instance with GPUs in it.
Create Compute Instance¶
In the Compute Instances menu, click on New Compute Instance. The user will be presented with the "small" SKU that was published by the PaaS Admin in the prior step.
Now, click on Select. This will now require the user to enter the following details:
- Unique name for the instance
- Description (optional)
- Select the workspace where you would like to deploy this instance (e.g. qa)
In the example below, we have selected the "small" SKU which will give us access to 1 GPU and 1 CPU, 1GB Memory etc.
The deployment of a compute instance can take a few minutes. Once it is deployed, it will look like the following:
Note
All the compute instances will be deployed into the compute instance onto the shared, multi-tenant host Kubernetes cluster that
Access Compute¶
The compute instance launched via self service by the end user is operating in a remote data center. The Rafay Platform allows users to access the compute instance remotely without requiring the use of a VPN or bastion hosts using the platform's zero trust kubectl capability.
Launch Jupyter Notebook¶
Jupyter Notebook is widely used because it provides an interactive environment where data scientists can easily combine code, text, visualizations, and other media within a single document, making it ideal for data exploration, prototyping, analysis, and presenting findings in a clear and organized way, especially in data science and machine learning projects.
We will now launch a Jupyter Notebook into the compute instance we created above.
In the Notebooks menu, click on New Notebook. The user will be presented with the SKU for Jupyter Notebook that was published by the PaaS Admin in the prior step.
Now, click on Select. This will now require the user to enter the following details:
- Provide a unique name for the notebook
- Select the compute instance where this will be deployed (from the above step)
- Select the workspace where you would like to deploy this instance (e.g. qa)
- Select the profile for the notebook from the available list (minimal, data science etc)
- Click on Deploy
In a few minutes, the notebook will be deployed to the selected compute instance. Click on the provided URL to access the notebook using your web browser.
Note
The notebook will be pre-configured with all the required libraries and frameworks as per the selected profile.
Invite Collaborators to Workspace¶
It is extremely common to data scientists to work in a collaborative environment. In the steps below, we will experience how the end user can invite collaborators to specific workspaces and limit access.
Info
You will need a user account in the Rafay Org with a 2nd email address to try out the scenario below.
In the steps below, you will invite collaborator to the "qa" workspace. You will then login as the collaborator to validate you can actually see the workspace.
Navigate to Workspaces and click on the "qa" workspace.
Click on the "Add Collaborator" button. You will be presented with a list of users in the Rafay Org to select for access to the workspace. Select a user from the list that you would like to invite as a collaborator and Save.
Logout and login into the Rafay Platform as the collaborator. Notice that you will have access to the Developer Hub and can view the same compute instance and other resources in the "qa" workspace.