Get the Most Out of MetaDefender Core

Performance Optimization and Fine-Tuning Guide

Customize Your MetaDefender Core Experience

While there’s no “one size fits all” approach to every use case, let’s make sure you’re getting the most from MetaDefender Core.

Stronger Core Resilience

Shorter Processing and Scanning Times

Increased Throughput

More Efficient CPU Usage

On-Demand Scalability

Real-World Results
for Cross Domain Security

Run on the same system—but optimized—MetaDefender Core performs significantly more efficiently compared to baseline output.

A typical use-case for cross domain solutions is scanning OS patches—usually in the form of archive files (e.g. CAB, MSU, DE, etc.). In this example, see how MetaDefender Core version 5.10.0 with 20 antivirus engines ran in an optimized environment.

support-charts performance-optimization-and-fine-tuning-screenshot-test-1.jpeg

Throughput

150 Objects/second *

*Scanning Windows patches totaling over 14M objects

Metrics That Matter

While processing files on MetaDefender Core, the service performance is measured by various metrics, some of those commonly used to define performance level include:

Throughput, or the number of processed objects per hour vs. number of processed files per hour
Submission load, or the number of successful requests per second
Average processing time per object
Total processing time against certain data set.

Performance Optimization and Fine-Tuning

It is crucial too for any high-performance system to have monitoring mechanisms that ensure all metrics are kept in control, and if not, are identified and addressed efficiently and quickly. In the context of MetaDefender Core, these parameters can be monitored for performance measurement:

System resources (CPU, RAM, Disk, etc.)
MetaDefender Core active performance
MetaDefender Core health check

Fine-Tuning MetaDefender Core

During its operation, MetaDefender Core and its engines are expected to utilize system resources as much as possible to boost throughput while not exceeding pre-defined thresholds continuously for long periods of time.

Fine-tuning Meta Defender Core based on your use-case and system hardware can maximize overall performance with optimization broken down into two categories: System Tuning and Application Tuning.

System Tuning

Application Tuning

Factors	Recommendation
CPU	Review server specification requirements: https://docs.opswat.com/mdcore/installatio Number of CPU cores, number of CPU sockets, and CPU frequency should be factored in Use CPU benchmarks (e.g. Geekbench 6) for accurate CPU performance
RAM	Review RAM capacity specification requirements: https://docs.opswat.com/mdcore/installation It is recommended to use DDR4 or DDR5 for optimal random memory access speed
Disk	Review disk space requirements: https://docs.opswat.com/mdcore/installation The use of a high disk speed (M2, SSD) is highly recommended
Environment	Power plan: Set to "high performance". Avoid using "power saving" mode since it may throttle CPU power and capacity. BIOS settings: Ensure all processors are enabled for turbo boost when applicable, and maximum performance for memory frequency and CPU Operating system: Windows Server 2022; Linux Red Hat Enterprise 9, Debian 11, or Ubuntu 20.04

Factors	Recommendation
Max file sizes setting	Larger files likely mean longer processing time, at least for temporary file creation and access, and file hash calculation. For security and performance reasons, each workflow has a default maximum file size setting for file scans. Adjust this limit as needed based on your dataset.
Archive settings	Large files are usually in archive compressed file format so it will require MetaDefender Core and its engines to extract and process nested files within. Increase archive settings based on your dataset.
File Type	File type timeout should be high enough as archive file could be large.
MetaDefender Core queue size	Consider increasing the MetaDefender Core queue size when scanning large quantities of archive files.
Engine parallel count	Increase this limit to push more tasks to engines when applicable. NOTE: increasing this limit does not always mean higher throughput produced by engines, as too many threads opened might overload engines.
Skip hash calculation setting	Hash calculation may require significant CPU resources and add overhead to the overall processing time—especially for large files.
Engine scan timeout	Lowering down this timeout threshold to seconds will help MetaDefender Core move on quickly without pending on any unexpectedly slow scanning from any engine at any time for any reason.
For guidance on standard settings please visit https://docs.opswat.com/mdcore/installation

MetaDefender Core Knowledge Base

Looking for more documentation for advanced MetaDefender Core usage? Visit our in-depth resource repository now for more information.

See Documentation