Embedded Files
Deploying CARPL
Deploying CARPL
CARPL is built for flexibility - it can be deployed on Cloud or On-Prem, with a configuration that is in line with your specific deployment needs. For custom deployments outside of CARING's infrastructure we require Secure SSH access to the machine where CARPL will be deployed.
On Premise Deployment
On Premise Deployment
We would require a server (whether virtual or physical) on customer’s premises to install the CARPL Application. Basic configurations are given below. The following configuration can handle 30-40 number of concurrent requests:
CPU : We would require a server (whether virtual or physical) on customer’s premises to install the CARPL Application. Basic configurations are given below. The following configuration can handle 30-40 number of concurrent requests:
RAM : 32 GB
STORAGE: Minimum: 300 GB and Recommended: 2 TB (as per user requirements)
Note: GPU: To run On-CARPL inference, the GPU requirements are in line with model architectures. CARPL supports Nvidia GTX1060, GTX1080ti, RTX2080 and Titan range of GPUs. (Not required for API-based inferencing on remote systems)
Third Party Software Requirements:
Operating System(Preferred): Ubuntu , Redhat and CentOS
Docker 20.10.2 or more
Docker-compose 1.25.5 or more
Note: The latest Third Party Software Requirements version information and compliance should be established against the latest release document of CARPL products. These software can be installed either by CARING Team or by the customer’s IT team depending on the requirement and permission.
Internet Requirement : Internet access with upload speed: 1MBPs or 8Mbps and download speed: 2MBPs or 16Mbps is required for remote support, configuration and monitoring. Ex: Cisco or Aruba routers.
Note: In case customer prefers to use windows instead of ubuntu the above server requirements remain the same. Customer side requires an additional license for Windows.
Whitelist IPs (Incase VPN not enabled): Please whitelist 1.22.91.90
Network Firewall Requirements :
On Cloud Deployment
On Cloud Deployment
CARPL can be deployed on customer’s choice of cloud provider e.g. Amazon, Azure, Google Cloud Platform etc. It can also be deployed on private clouds as well. We require a virtual machine with the following configuration:
CPU : t2.large (equivalent or higher)
RAM : 32 GB
STORAGE: We use block storage attached to the VM (eg. EC2 instance). Minimum: 300 GB Recommended: 2 TB (as per user requirements)
Note: GPU: To run On-CARPL inference, the GPU requirements are in line with model architectures. CARPL supports Nvidia GTX1060, GTX1080ti, RTX2080 and Titan range of GPUs. (Not required for API-based inferencing on remote systems)
Third Party Software Requirements:
Operating System(Preferred): Ubuntu , Redhat and CentOS
Docker 20.10.2 or more
Docker-compose 1.25.5 or more
Note: The latest Third Party Software Requirements version information and compliance should be established against the latest release document of CARPL products. These software can be installed either by CARING Team or by the customer’s IT team depending on the requirement and permission.
Internet Requirement : Internet access with upload speed: 1MBPs or 8Mbps and download speed: 2MBPs or 16Mbps is required for remote support, configuration and monitoring. Ex: Cisco or Aruba routers.
Note: In case customer prefers to use windows instead of ubuntu the above server requirements remain the same. Customer side requires an additional license for Windows.
Whitelist IPs (Incase VPN not enabled): Please whitelist 1.22.91.90
Network Firewall Requirements :
Page updated
Report abuse