HPSee HPSee
down

HPSee Requirements

Find Your HPSee Setup

Find the right HPSee setup for your infrastructure using the decision tree below. Select a setup to view its recommended hardware and software requirements, then share the direct link with your IT team. Actual requirements depend on your workload.

Your Setup
Managed cluster on AWS

1. Where do you want to run HPSee?
On premises
In the cloud

2. How is your on-premise hardware organized?
Machines dedicated to HPSee
A cluster controlled by a scheduling system
Currently supported: Slurm only

3. How many machines do you dedicate to HPSee?
One machine
Docker Compose
Multiple machines
Docker Swarm

Deployment Guidance

Deployment guidance

  • All values are guidance; actual requirements depend on your workload, such as the number of jobs running at the same time.
  • Dedicated machines need hardware reserved for HPSee.
  • Available deployment options depend on your HPSee version.

Next step

The HPSee Deployment Guide describes the installation for each setup. Requirements are refined as HPSee evolves.

Your Setup: AWS

HPSee runs as a managed cluster on AWS, with Amazon Aurora PostgreSQL as the database. Worker node requirements follow the specifications in BioSolveIT’s Terraform Template.

To share these requirements with your IT team, copy the link below and forward it.

Copy requirements link

API

4 vCPU
8 GB RAM
≥ 20 GB Disk

Access to HPSee Platform

Database

Aurora

Amazon Aurora PostgreSQL

EventBroker

1 vCPU
1 GB RAM
1 GB Disk

Connects to AWS Batch

Worker nodes

Up to 300

Managed cluster on AWS
Autoscaled starting with HPSee 3.0

16 vCPU
32 GB RAM
20 GB Disk

Per worker node. Can be adjusted in Terraform.

Cluster Setups:
High Scale, High Performance

The setup consists of worker nodes and a dedicated manager node that hosts the API and can also host the database. In Slurm setups, the manager node additionally runs one EventBroker replica. Docker Swarm setups do not require an EventBroker.

To share these requirements with your IT team, copy the link below and forward it.

Copy requirements link

Docker Swarm

Dedicated worker nodes plus a dedicated manager node to host the API. The database can also be co-located on the manager node (recommended).

Slurm

Worker nodes according to the scheduling system. A dedicated manager node is required to host the API and EventBroker. The database can be co-located on the manager node.

Scaling the setup

Scale your setup according to the size of your cluster. Two size suggestions can be found below.

Cluster sizes

Small Cluster: < 50 machines
Large Cluster: > 50 machines

Select Your Cluster Size

Small Cluster Large Cluster

Your Setup: Small Cluster

The setup consists of worker nodes and a dedicated manager node that hosts the API and can also host the database. In Slurm setups, the manager node additionally runs one EventBroker replica. Docker Swarm setups do not include an EventBroker.

To share these requirements with your IT team, copy the link below and forward it.

Copy requirements link

API

4 vCPU
8 GB RAM
≥ 20 GB Disk

Access to HPSee Platform

Database

16 vCPU
64 GB RAM
2 TB Disk

PostgreSQL. The 48 GB container memory budget is included in the RAM allocation.

EventBroker

1 vCPU
1 GB RAM
1 GB Disk

Connects to Slurm

Worker nodes

< 50

8 vCPU
16 GB RAM
20 GB Disk

Worker requirements apply to each concurrent job.
Scale for faster processing.

Manager node

The API, database and EventBroker can run on the same machine.

If these services share the manager node, its capacity must at least
cover their combined resource allocations shown above.

Your Setup: Large Cluster

The setup combines worker nodes with a dedicated manager node, which hosts the API and can also host the database. For Slurm setups, the manager node also runs one EventBroker replica. Docker Swarm setups do not include an EventBroker.

To share these requirements with your IT team, copy the link below and forward it.

Copy requirements link

API

16 vCPU
32 GB RAM
≥ 20 GB Disk

Access to HPSee Platform

Database

32 vCPU
128–192 GB RAM
4 TB Disk

PostgreSQL. The 48 GB container memory budget is included in the RAM allocation.

EventBroker

1 vCPU
1 GB RAM
1 GB Disk

Connects to Slurm

Worker nodes, per job

> 50

8 vCPU
16 GB RAM
20 GB Disk

Worker requirements apply to each concurrent job.
Scale for faster processing.

Manager node

The API, database and EventBroker can run on the same machine.

If these services share the manager node, its capacity must at least cover their combined resource allocations shown above.

Your Setup: Single Machine

One dedicated machine runs the database, the API and the worker on the same hardware. No separate manager node is needed.

To share these requirements with your IT team, copy the link below and forward it.

Copy requirements link

CPU

20+ vCPU

Database: 8+ cores
API: 4 vCPU
Worker: 8 vCPU

RAM

56 GB

Database: 32 GB
API: 8 GB
Worker: 16 GB

Disk

620 GB+

Database: 500 GB+ NVMe
API: 100 GB+
Worker: 20 GB NVMe

Network

1 GbE

10 GbE if users download whole libraries.
Database uses loopback.

Recommended sizing example

  • Docker Compose with one worker.
  • CPU, RAM and disk add up: all services share one machine.

Planning faster compute?

  • Allocate additional CPU cores for faster processing, with 2 GB RAM per worker core.
  • Allow 2 GB RAM per core.

Important note:

Reserve the machine for HPSee: It does not throttle itself for other jobs.

Operating Systems and SeeSAR Client Requirements

To share these requirements with your IT team, copy the link below and forward it.

Copy requirements link

HPSee Server

Linux

  • Supported.
  • Docker Engine or Apptainer, depending on your setup.

Windows

  • Supported.
  • Windows 10 (version 1903, build 18362) or later.
  • WSL2 and Docker Desktop.

macOS

Server installations are not currently supported.

Applies to installations on your own machines. AWS setups follow the AWS deployment.

Client

SeeSAR

16 GB RAM

Required for working with C-S-D results in SeeSAR.

Fully Supported

  • Linux
  • macOS
  • Windows