Applies to: 
Kyvos Enterprise Kyvos Cloud (SaaS on AWS) 
Kyvos AWS Marketplace
Kyvos Azure Marketplace Kyvos GCP Marketplace Kyvos Single Node Installation (Kyvos SNI)
Previous versions of Kyvos relied on external compute clusters, such as EMR, Databricks, or Dataproc in cloud environments, and Hadoop, CDH, HDP clusters on-premises.
Kyvos has introduced Modern Architecture in which you can select the Kyvos Native compute engine for processing semantic models. This advancement has removed the need for traditional Spark jobs and external compute engines like Databricks, Dataproc, or EMR when processing semantic models and constructing aggregates. Additionally, this dependency is eliminated related to the external compute cluster, such as operating costs and security concerns.
Following are the supported Native compute options:
Shared Query Engine: In this mode, the Query Engines serve queries as well as handle the semantic model processing. This dual role is named SHARED as the same process performs both activities.
Dedicated Compute Server: In this mode, the semantic model is processed via dedicated service.
For cloud-based deployments, the semantic model is processed using Kubernetes (K8S) cluster-based nodes.
For On Premises environments, semantic models are processed on dedicated nodes.
Deployment Method | Compute Type | AWS | AZURE | GCP | ON PREM |
Automated | External Compute | Yes (EMR / Databricks) | Yes (Databricks) | Yes (Dataproc) | Not Applicable |
Only Kyvos Native Compute (SHARED_QE) | Yes | Yes | Yes | ||
Only Kyvos Native Compute (K8S) | Yes | Yes | Yes | ||
Wizard | External Compute | Yes | Yes | Yes | Yes |
Only Kyvos Native Compute (SHARED_QE) | Not Available | Not Available | Not Available | Yes | |
Only Kyvos Native Compute (K8S) | Not Available | Not Available | Not Available | Not Applicable | |
Only Kyvos Native Compute (Dedicated Compute) | Not Applicable | Not Applicable | Not Applicable | Yes |
Choose the installation guide according to your environment.