Choosing the right deployment #

Starburst Enterprise is available for on-premise usage or in private clouds, and can be run on bare metal servers, virtual machines and containers, all managed by you.

You can also run Starburst Enterprise on public cloud provider systems, and their virtual machine or container offerings, or for further simplicity and convenience use their marketplaces.

That’s a lot of choice! This guide helps you select the best solution for your organization.

The deployment you choose depends on a couple of key factors:

  • People and available skills in your organization
  • Variety and location of data sources
  • Location of your computing resources
  • Security and governance requirements

People #

As you decide the best Starburst product and deployment style for your organization, the skill sets available in your organization is the most important factor. You need the expertise of platform administrators and data engineers to create, configure and maintain clusters capable of running Starburst Enterprise. Your organization must also have people who can connect the desired data sources and ensure that data security and governance requirements are met. Typically, people with these skill sets sit within your IT organization.

Starburst Enterprise offers the most control of your deployments at the cost of being the most complex to install and maintain.

With limited availability using a public cloud provider or even marketplace offering can help reduce the workload.

Data sources #

The data sources you plan to query have a large impact on your choice. You need to understand what databases, object storage or others systems your users need to query. In addition, you need to know where these system are deployed to ensure that Starburst Enterprise or Starburst Galaxy can access them with sufficient network performance and capacity.

For example, if all your data is stored in your private network and data center, you should run Starburst Enterprise there as well. However, if all your data is hosted on a public cloud provider, you can choose to run Starburst Enterprise on the same cloud provider yourself, use a marketplace offering or even use Starburst Galaxy.

Starburst Enterprise includes a very large variety of connectors to support the most common data sources. They include many RDBMSs, Hadoop/Hive and other object storage systems, and commercial platforms such as Snowflake or Teradata. In addition, Starburst Enterprise can be operated anywhere. If your organization needs to query large collection of data sources, Starburst Enterprise is the right choice for you. You can choose various options on how to run it, and ensure it is located closely to your data sources.

Location #

If you run exclusively on-prem, then your choice is easy - Starburst Enterprise. You can run it on bare metal servers, on virtual machines, in private clouds, or even in your Kubernetes clusters on supported cloud platforms. Your choice depends entirely on your requirements, data sources, people and skills.

If you are using a single cloud provider, you can run Starburst Enterprise, again with the options to manage the virtual machines yourself and use the tar.gz or RPM archive.

You can also use a Kubernetes offering from cloud providers, and combine it with the our Kubernetes support with Helm charts.

Alternatively, you can use cloud-specific offerings such our Amazon Cloudformation support.

Starburst also makes Starburst Enterprise available in all major marketplaces.

If you use a multi-cloud or hybrid cloud strategy, the locality of your data should be your first consideration to reduce data transfer costs. You can even run multiple Starburst Enterprise clusters, and potentially connect them with Starburst Stargate.

Here are some things to consider in choosing a location for your Starburst Enterprise cluster, keeping minimizing intra-cloud data transfer as low as possible front-of-mind:

  • Can you place your cluster in a cloud that contains multiple data sources?
  • What intra-cloud data sources are most likely to be federated? Where is the largest of them?
  • Is one of your cloud provider pricing models more favorable than the others?

The closer you can put Starburst Enterprise to the bulk of your data, the more you can reduce the amount of data being returned, and save on data transfer costs.

Security #

What does data security and governance look like at your organization?

Starburst Enterprise supports a wide range of security features.

  • Different authentication platforms such as LDAP, Kerberos or OAuth
  • Data access control with user impersonation, credential passthrough and others
  • Authorization management with Ranger or Privacera platform
  • Query logger for auditing and tracking
  • and others

Our security guide provides a good introduction.

The location and mode of running Starburst Enterprise specifically determines which security features can be used. You need to specifically verify if a feature can be used based on your determined location.

For example, if your company relies on Apache Ranger or Privacera platform to secure your data, then a self-managed Starburst Enterprise in a private cloud or on-prem is most likely the right choice for you.

Depending on your particular security needs, one or more marketplace offerings may also work. Starburst Enterprise includes Helm charts for Apache Ranger installation and usage on Kubernetes.

Starburst Galaxy #

As our most convenient solution, Starburst now offers a hosted and managed solution - Starburst Galaxy.

Starburst Galaxy is the most hands-off approach as it provides a simple user interface and most complexity, including deployment and upgrade, is completely taken care of for you by Starburst.