Skip to content

Platform Support Guide

Yellowbrick has two editions — Community Edition and Enterprise Edition — and targets multiple supported platforms. Although most core database documentation is applicable to all platforms and editions, there are differences. This Platform Support Guide explains the differences and contains sections of documentation that may be pertinent to only one platform or edition.

When looking at shared documentation, at the top of each page of you'll find capsules listing the combinations of editions and platforms that the page in question applies to.

The current set of platforms and editions are documented below.

Yellowbrick Editions

Community Edition (CE)
A freely available version of Yellowbrick, for use on scale-up (single node) hardware instances, typically used for development or smaller data set sizes. Community Edition runs in both general container environments such as Docker as well as on optimized cloud instances. Some enterprise-class features are explicitly disabled. Community Edition is available on public clouds and Docker Hub. For more information, click here.
Enterprise Edition (EE)
The fully featured, scale-out version of Yellowbrick with enterprise support. Enterprise Edition runs on all Yellowbrick appliances, as well as all public and private cloud platforms. The database engine is identical between platforms, but the infrastructure and storage architecture are different, leading to some differences in system views, SQL utility statements and administration.

Yellowbrick Appliances

Tinman appliance (EE-APPLIANCE-TINMAN)
Tinman is Yellowbrick's first-generation appliance used in appliance platforms data centers. It is a blade server architecture using Intel Broadwell CPUs with integrated 56gbit active/passive dual port Inifiband backplane and integrated NVMe storage devices.
Andromeda appliance (EE-APPLIANCE-ANDROMEDA)
Andromeda is Yellowbrick's second-generation appliance used in appliance platforms data centers. It is also a blade server architecture using AMD EPYC CPUs with integrated 100gbit active/active dual port Infiniband backplane and removable NVMe storage devices.

Yellowbrick Cloud Platforms

Yellowbrick software supports numerous cloud platforms by running on top of Kubernetes. Currently supported cloud platforms are as folllows:

Amazon Web Services (EE-CLOUD-AWS)
Yellowbrick runs on top of EKS (Elastic Kubernetes Services), stores metadata and row-oriented data on EBS (Elastic Block Store) and columnar data on S3.
Microsoft Azure (EE-CLOUD-AZURE)
Yellowbrick runs on top of AKS (Azure Kubernetes Services), stores metadata and row-oriented data on Azure Disk Storage and columnar data on Azure Blob Storage.
Google Compute Platform (EE-CLOUD-GCP)
Yellowbrick runs on top of GKE (Google Kubernetes Engine), stores metadata and row-oriented data on persistent disks, and columnar data on GCS (Google Cloud Storage).

Differences Between Appliance and Cloud Platforms

Both appliance platforms and cloud platforms use the same underlying database engine, that processes data using SQL in exactly the same way. The differences are in the following areas:

AreaAppliance platformsCloud platforms
Installation Appliance hardware is installed into a data center by Yellowbrick staff and comes with all database and system software preinstalled. Cloud software is installed into a cloud account using the Yellowbrick deployer.
Storage Appliances have co-located storage and compute. Data is stored on SSDs within compute nodes and protected using parity encoding. Cloud software makes use of separate storage and compute. The software stores data on both block storage and object storage from the cloud provider.
Compute nodes The number of compute nodes available to Yellowbrick is dictated by how many blades are installed in the appliance. Compute nodes are allocated and de-allocated dynamically from the cloud using Kuberentes, with the number of compute nodes limited solely by the capacity of a given availability zone.
Elastic compute clusters Appliances have one 'default' compute cluster comprising all intalled compute blades. Multiple compute clusters with separate storage and compute can be created on the fly. Users are assigned to clusters and intensive workloads can be load-balanced across them.
User Interface The Yellowbrick Systems Management Console (SMC) is used for management of the appliance, along with YBCLI for management of the physical platform. Third party SQL authoring tools must be used. Yellowbrick Manager is used for management of the instance as well as authoring and issuing ad-hoc SQL. Third party SQL authoring tools are optional.
Authentication and authorization Database local authentication, OAuth2.0 authorization and LDAP synchronization are supported. Kerberos is also supported via editing configuration files. OAuth2.0 allows easy integration with federated identity management solutions. Database local authentication and OAuth2.0 authorization are supported.

Differences Between Cloud EE and CE Platforms

Both CE and EE cloud platforms use the same underlying database engine, that processes data using SQL in exactly the same way. The differences are in the following areas:

AreaEE Cloud platformsCE Cloud platforms
Installation The EE Cloud platform is installed into a cloud account using the Yellobrick deployer. Cloud CE software is installed into a cloud account using the Cloud provider's specific Yellowbrick CE deployer.
Backup, Restore and Replication Support The EE Cloud platform supports backup, restore and replication. The CE Cloud platform does not support backup, restore and replication.
Compute node size The EE Cloud platform allows scaling up and down of compute nodes, based on the license purchased. The CE Cloud platform only supports a single compute node, and allows a single compute node with a max of 16 cores.
Compute nodes The EE Cloud platform allows scaling up and down of compute nodes. The CE Cloud platform only supports a single compute node, and does not support scaling up and down of compute nodes.
Elastic compute clusters Multiple compute clusters with separate storage and compute can be created on the fly. Users are assigned to clusters and intensive workloads can be load-balanced across them. CE cloud platforms have one 'default' compute cluster.