Skip to main content

System Requirements

Below are the recommended system requirements for Trisul deployments. Trisul runs on commodity servers, either bare metal or virtual machines. Below you’ll find recommended hardware and resource guidelines based on the two data-processing modes and varying deployment sizes.

Choose the processing mode and size that match your network’s scale, then proceed to Installation.

Operating Systems

Trisul Network Analytics packages are available on the following platforms.

OSRecommendedNotes
Ubuntu LTS 22.04/24.04Ubuntu 24.04
RHEL 9/8RHEL 9.xCan also use OracleLinux, AmazonLinux, RHEL, CentOS versions 9/8

Processing Modes

Trisul’s four products fall into two technical categories based on how they process data.

  • Flow-Based Mode (For Trisul NetFlow Analyzer, Trisul ISP, Trisul IPDR)- needs lower CPU, Memory and resources. And, in the case of Trisul IPDR, additional storage.
  • Packet Capture Mode (For Trisul NSM) - needs more CPU, Memory resources

See below for typical requirements.

Flow-Based Mode Requirements

Flow-Based analytics sizing is based on number of devices and interfaces. Choose the sizing block that most closely matches your network’s scale.

Trying it out

If you are just trying out Trisul, it is recommended to start with a small version as below.

HardwareSystem Requirements
TypeVM preferred
CPU8 vCPU cores
Memory16GB RAM
Network1GbE interface that can be used for both NetFlow and Management access
Disk1TB SAS, this can store upto 6 months data, proportionately add more based on retention

Table: Minimum System Requirements for Small Enterprise

For Trisul ISP

Refer to product specific system requirements here: Trisul ISP System Requirements.

For Trisul IPDR DoT Compliance Solution

Refer to product specific system requirements here: Trisul IPDR System Requirements

Packet Capture Mode Requirements

Packet capture analytics sizing is based on total packet bandwidth. This is not the link speed but rather the actual utilization of the link. Choose the sizing block that most closely matches your network’s traffic volume.

HardwareSystem Requirements
TypeBare metal preferred
CPU8 vCPU codes
Memory16GB RAM
MechanismSPAN Port
Network1 GbE for receiving the packets via SPAN port and another for management access
Disk1TB high speed SAS or NVMe for PCAP and 1TB SAS/SATA for database
OSOracle 22.04 Jammy or RHEL 9

Advanced Scaling

This section goes deeper into how Trisul handles heavier workloads. Read it when you need to plan for scaling beyond the standard sizing levels.

Load Profile

The table below gives a quick view of how the two main Trisul components use system resources. This helps you understand which parts of the system grow with traffic volume and which grow with the number of probes you deploy.

Node TypeDescriptionLoad Profile
Trisul HubStores all analytics data and handles searches, reports, and long-term storage(Database node).Mostly disk I/O–intensive. Load increases as more probes send data.
Trisul ProbeCollects network traffic (flows or packets) and processes it in real time (streaming analytics) before sending the results to the HubMostly CPU-intensive. Load increases with traffic volume.

Trisul Hub Scaling Rules

The Trisul Hub stores all analytics data and serves queries from users and probes. Because the Hub handles long-term data and high-frequency lookups, disk performance and capacity are the main factors that affect its scaling.

ModeScaling Metric When Hub Load IncreasesAdditional Resources Needed
HubFor every medium volume probe + Every 5 concurrent users+1CPU core and +2GB RAM

Trisul Probe Scaling Rules

The Trisul Probe handles live traffic capture and real-time analysis. As traffic volume increases, the Probe needs more CPU and memory to keep up. The table below gives a simple guideline for scaling a Probe based on packet capture load.

ModeScaling Metric When Probe Load IncreasesAdditional Resource Needed
Raw PacketsFor every 200Mbps of traffic+1 CPU core(3Ghz) and +4GB DDR4 (RAM)

Scaling Flow-Based Mode

In Flow-Based Mode, the key scaling factor is the amount of flow data being sent to Trisul. Flow records (NetFlow, sFlow, IPFIX, etc.) usually represent a small fraction of actual link bandwidth.

A simple way to estimate flow traffic is:

  • Estimating Flow Traffic : Flow data is typically 0.5–1% of your total network bandwidth.
    For example, a 1 Gbps link usually produces 0.5–1 Mbps of flow data.

The table below gives a guideline for scaling Trisul in this mode:

ModeScaling Metric When Load IncreasesAdditional Resources Needed
Flow ModeFor every 10Mbps flow traffic+1 CPU Core and +2GB RAM