A serverless runtime environment is an advanced serverless deployment solution that doesn't require downloading, installing, configuring, and maintaining a Secure Agent or Secure Agent group. You can use a serverless runtime environment in the same way that you use a runtime environment when you configure a connection or some types of tasks in Data Integration.
Compared to the multi-tenant model on the Hosted Agent, a serverless runtime environment uses an isolated, single-tenant model. The model provides a dedicated server with virtual machine resources to run tasks for your organization. The serverless runtime environment auto-scales with the size of the workload while your data remains in your cloud environment.
A serverless runtime environment is hosted in Informatica's Amazon Virtual Private Cloud (VPC) on the AWS cloud platform. The serverless runtime environment creates a cross-account elastic network interface (ENI) to connect to your cloud environment.
Note: To use a serverless runtime environment, your cloud environment must be on the AWS cloud platform and your VPC must have default tenancy. A serverless runtime environment cannot connect to a VPC with dedicated instance tenancy.
A serverless runtime environment supports local regions in each geo-location. For example, an AWS cloud platform in the United States (US) supports all US regions and an AWS cloud platform in Asia-Pacific (APAC) supports all APAC regions.
To use a serverless runtime environment, your organization must have the appropriate licenses. Additionally, users that require access to the Serverless Environments configuration page must have the PRIVILEGES.VIEW_AGENT_GROUP permission.
Using a serverless runtime environment for Data Integration Elastic
When you use a serverless runtime environment to run elastic jobs, the advanced serverless deployment of Data Integration Elastic is configured with the prerequisites to create an elastic cluster and to run the jobs on the cluster.
The serverless runtime environment manages the elastic cluster while the cluster adapts to workload changes by provisioning and deprovisioning resources.
Worker nodes in the elastic cluster are highly available. High availability mitigates job failures and maintains job performance when a worker node crashes.