Skip to main content

Agent System Requirements

System Requirements​

To ensure that the Agent is able to communicate with Ganymede Cloud, ensure that your computer meets the following specs:

Hardware Requirements​

  • Intel i3 processor or newer, with x86/x64 architecture
  • At least 4GB of RAM, though more may be needed for heavy workloads.
  • At least 10GB of free disk space

Network Requirements​

  • Ability to access relevant network-limited resources (network drives, VPN, databases, etc.)
  • has outbound HTTPS (port 443)

OS Requirements​

  • Either Windows 10 or newer with display, or Ubuntu/Debian Linux prior to EOL
  • Operating system should be up to date with latest patches and updates

Installation permissions​

Administrative access is required so that the Agent can be installed to run as a service, which enables the Agent to launch automatically on system startup. Once configured, users are not required to log in as an admin.

Testing Connectivity​

If direct access to the internet is allowed (no proxy), configure your Firewall to allow Outbound HTTPS (port 443) connections from the PC where the agent is running. If direct access to the internet is not allowed (proxy), configure your Proxy to allow the following connections from the PC where the agent is running.

  • storage.googleapis.com on port 443
  • pubsub.googleapis.com on port 443
  • <tenant_name>.ganymede.bio port 443

Windows Agents​

For Windows Agents, connectivity can be tested by running the following commands in PowerShell:

$ IWR -URI storage.googleapis.com
$ IWR -URI pubsub.googleapis.com
$ IWR -URI <tenant_name>.ganymede.bio

Linux Agents​

For Linux Agents, connectivity can be tested by installing netcat and running the following commands in a terminal:

$ nc -vz storage.googleapis.com -Port 443
$ nc -vz pubsub.googleapis.com -Port 443
$ nc -vz <tenant_name>.ganymede.bio -Port 443

Agent Configuration Considerations​

Once basic system requirements are met, the core areas requiring consideration are how to configure Agents to suit your operations. Below are some questions to consider when configuring your Agents:

User Interaction with Lab Instruments​

Frequently, the data producer is not the same person who is configuring Agents. Some questions for the data producer are:

  • Does the instrument software have an option to auto-export files? If not, could data be exported to a consistent location? Ensuring process consistency is critical for ensuring data consistency.
  • What are the file formats that are being exported? Are they standardized?
  • Which of the exported files should be captured in Ganymede Cloud?
  • Is there a way to instruct the instrument to include experiment/sample ID within the file contents or file name? If not, how can this information be supplied?

File Exports​

  • Is it feasible to replicate data from an instrument to a network drive? This may be helpful if there is sensitivity around installing software on instrument PCs, and potentially reduce effort required to manage Agents installed across many machines.
  • What are the folder location(s) for the file exports?
  • Does the instrument output new files for each run, or does it output a single file for all runs? If the latter, what is the desired impetus for save (e.g. - last modified time > 5 minutes ago)
  • For instruments with multiple users, does the instrument export to the same folder location for each user?

Instrument Run Metadata​

Tagging files with relevant characteristics is critical for ensuring that data remains appropriately organized and accessible once captured in Ganymede. By default, Ganymede captures PC information as shown in the Monitoring Agent Connections section of the Agent documentation, but instrument data often can be shaped to a degree by end-user needs. Some questions to consider when tagging files are:

  • What are the metadata fields that are relevant to the instrument? Some examples include:

    • Instrument name / vendor
    • Instrument function
    • Instrument location
    • Software name / version
    • Team
    • Program
    • Experiment
    • Sample
    • User

  • Where do the metadata fields live? Are they in the file name, file contents, or contents of the captured files?