--- title: "Considerations for scalability" slug: "considerations-for-scalability-4" tags: ["Install and Get Started", "ePolicy Orchestrator – On-prem", "ePO 5.10.x", "ePO 5.10.0"] updated: 2026-09-07T08:10:37Z published: 2026-09-07T08:10:37Z canonical: "docs.trellix.com/considerations-for-scalability-4" --- > ## Documentation Index > Fetch the complete documentation index at: https://docs.trellix.com/llms.txt > Use this file to discover all available pages before exploring further. # Considerations for scalability You can scale your ePO - On-prem infrastructure in multiple ways. Scaling is needed if the environment managed by ePO - On-prem is growing beyond the capacity of the current ePO - On-prem infrastructure. To grow your ePO - On-prem infrastructure, you can: Move the ePO - On-prem SQL database to a larger and more powerful SQL Server, add more Agent Handlers, or increase CPU and memory to improve storage performance on the SQL Server. With ePO - On-prem software, you can scale your network vertically or horizontally. - ***Vertical scalability*** — Adding and upgrading to bigger, faster hardware to manage larger and larger environments. Scaling vertically is accomplished by upgrading your server hardware, and installing ePO - On-prem on multiple servers throughout your network, each with its own database. - ***Horizontal scalability*** — Increasing the size of the environment that one ePO - On-prem server can manage. Scaling horizontally is accomplished by installing additional Agent Handlers, all sharing a single database. Make sure the ePO - On-prem infrastructure is scaled to handle major peaks in outbreak situations. ### Managed systems and servers The number of systems your ePO - On-prem server manages dictates the number and size of the servers needed. It also dictates the recommended server sizing needed to manage these systems. | Option | < 1,500 systems | 1,500–10,000 systems | 10,000–25,000 systems | 25,000–75,000 systems | > 75,000 systems | | --- | --- | --- | --- | --- | --- | | Virtual ePO - On-prem server | Yes | Yes | Yes | Yes | Yes | | Virtual SQL database server | Yes | Yes | Yes | Conditional* | Conditional* | | ePO - On-prem server and SQL database on the same server | Conditional* | Conditional* | Conditional* | Conditional* | Not recommended | | Add distributed repositories | Not required | Conditional* | Conditional* | Yes | Yes | | Add Agent Handlers (virtual) | Not required | Conditional* | Conditional* | Yes | Yes | *For more information, see [Combining servers](/v1/docs/combining-servers). > [!TIP] > Tip > > We recommend one Agent Handler for every 50,000 systems. There is no hard limit on the number of systems ePO - On-prem can manage. The primary limitation is the SQL database performance, specifically disk performance (IOPS – I/Os per second). You can scale the SQL database, add distributed repositories, and add Agent Handlers to manage more systems as needed. ### Important sizing considerations - **Environment** — Estimates based on a ePO - On-prem server running the Trellix ENS products. - **Operating systems** — You must use a 64-bit operating system for the ePO - On-prem server and SQL Server. - **CPU cores** — Server class, minimum 2.2 GHz. - **RAM** — Add 16 GB of RAM to SQL Server for every 25,000 nodes. - **Storage capacity** — Estimated event retention period of 6 months. - **Storage performance** — Storage estimated event retention period of 6 months. ### Recommended hardware based on number of managed systems | Node count | ePO - On-prem server | | --- | --- | | CPU cores | RAM (GB) | Storage (GB) | | < 10,000 | 4 | 8 | 300 | | 10,000–25,000 | 4 | 8–16 | 500 | | 25,000–75,000 | 8 | 16–32 | 500 | | 75,000–150,000 | 12 | 16–64 | 500 | | 150,000 + | 16 | 16–64 | 500 | | Node count | Agent Handler | | --- | --- | | Number of Agent Handlers | CPU cores | RAM (GB) | Storage (GB) | Notes | | < 10,000 | — | — | — | — | You can use a single server or multiple VMs. | | 10,000–25,000 | 0–1 | 4 | 8 | 150 | | | 25,000–75,000 | 0–1 | 4 | 8 | 150 | | | 75,000–150,000 | 1–3 | 4 | 8 | 150 | | | 150,000 + | 3+ | 4 | 8 | 150 | | | Node count | SQL Server | | --- | --- | | CPU cores | RAM (GB) | Storage (TB) | Performance (IOPS) | | < 10,000 | 4 | 8–16 | 0.5–1.0 | | | 10,000–25,000 | 4 | 8–16 | 0.5–1.5 | | | 25,000–75,000 | 8 | 16–32 | 1.0–2.0 | >10,000 | | 75,000–150,000 | 16 | 32–128 | 2.0–3.0 | >30,000 | | 150,000 + | 32+ | 128–256 | 3.0 | >90,000 | > [!NOTE] > Note > > By default, IOPS is calculated with 4-KB sectors. Microsoft recommends that SQL storage solutions are allocated with a 64-KB sector size.