Cloud-ready or on-prem?

SparkBeyond enables users to use cloud-native, multi-cloud, and on-premise platforms that seamlessly work with existing data storage, enterprise source systems, tools, and underlying infrastructure.

SparkBeyond's product suite is designed to seamlessly connect to and work in harmony with major cloud infrastructure and service providers without compromising stringent data governance, compliance, and security requirements. ​

‍
Microsoft Azure™ Services

• Use the C3 AI Suite to leverage the full power of Azure services for the rapid design, development, deployment, and operation of next-generation AI applications​
• Drive predictive insights, add business value, and solve previously unsolvable problems

AWS™ Services

• ​Leverage the power of Amazon Web Services as part of a single-, multi-, or hybrid-cloud environment for developing, deploying, and operating next-generation AI applications​
• Scale across a multiplicity of data sources to generate and operationalize predictive insights

Google Cloud™ Services

Seamlessly deliver the C3 AI Suite on Google Cloud Platform, leveraging its cutting-edge infrastructure and AI services as part of a multi or hybrid-cloud environment​.

Operating Systems

The following Linux distributions are supported:

• CentOS 7.x
• RHEL 7.x
  

Note: It is highly recommended to always run the latest version

Deployment Options

• On-Premise
• Client’s Public Cloud Instance (Azure, AWS and GCP)
• Client’s Private Cloud Instance
• SparkBeyond-Managed Instance (Azure, AWS and GCP)

Delivery Methods

• Amazon Machine Image (AMI)
• Azure VM Image
• Google GCP VM Image
• VMware Image
  

Deployment time estimates

• 1-2 hours if deployed using a disk image (Azure, AWS, GCP and VMware)
• 1-2 business days if deployed on-premise*
  

Note: Deployment on hardened systems may take more than two business days

Example: 4 Users

On-Prem configuration for up to four concurrent jobs per machine

• 160 GB RAM
• 2 TB attached storage (SSD drive is highly recommended; storage size must be at least 10 times greater than the total estimated data size)
• 20 vCPUs (10 cores with hyper-threading)
  

Cloud configuration for up to four concurrent jobs per machine

• AWS:  minimum instance type is m4.10xlarge
• Azure:  minimum instance type is E20s_v3
• Google Cloud:  minimum instance type is n2-standard-48
• For cloud deployment, the minimum instance specs are:
  
• 160 GB RAM
• 2 TB attached storage
• 32 vCPUs (16 cores with hyper-threading)

Note: When it is not possible to allocate a single disk for the whole root filesystem (e.g Azure), the installation can be divided between two disks:

1. A root disk (with a minimum of 32GB) mounted under /. In addition, it is highly recommended to allocate a sufficient amount of space for the /tmp directory (around 200GB),  so a root disk with 250GB in total would be ideal.
2. An installation disk, mounted under /opt/sparkbeyond with the recommended size (500GB minimum).