Elastic High Performamce Computing (E-HPC) is billed on the basis of your usage. It mainly includes the following services: ECS instances, E-HPC, Network Attached Storage (NAS), and outbound traffic through a logged on node.
Note: To deploy a large cluster, open a ticket to avail better deals.
In an E-HPC cluster, each node is an ECS instance, and its price depends on the specifications you select. An ECS instance is created when you deploy an E-HPC cluster, so you do not need to have one in advance. Any ECS discount you have also applies. For more information, see ECS Pricing and ECS Pay-As-You-Go.
E-HPC provides the multi-dimensional management of your cluster, instance resizing option to meet your business demands, and integrates with the features including OpenAPIs, monitor, and alarms, O&M tools, and automatic server O&M in the background.
Billing is according to your core-hours usage. So, if your total number of cluster cores in a region is smaller than 200 (vCPU), you are not billed. If the total number is greater than 200 (vCPU), 0.003 USD is charged for each extra core-hour.
NAS is used as the shared storage space for all nodes in an E-HPC cluster. For more information, see NAS pricing.
The node you have logged on is allocated a public bandwidth and linked with an EIP by default. You only pay for the outbound traffic, which is billed by traffic per hour.
For example, if the outbound traffic usage is 10 GB, the pricing is 10 GB * price (per GB). The traffic pricing differs according to the regions. For more information, see EIP billing.
For basic information about EIP, see Elastic IP address.
Total price of an E-HPC cluster = ECS instances price + E-HPC services price + NAS services price + Outbound traffic price.