2.1.1 Increase capacity and resiliency

Single server capacity

Please note that in the capacity measurement below MetaDefender Core and MetaDefender Email Gateway Security were deployed to the same physical machine.

System profile

Email consists of

Throughput

CPU usage

  • 4 Core Processor

  • 16 GB RAM

  • 64 bit OS

  • MetaDefender Core 20 engines

  • 30% of emails with no attachment

  • 40% of emails with single attachment (sanitization is not enabled)

  • 12.5% of emails have PDF1 attachment (sanitization is enabled)

  • 12.5% of emails have DOCX1 attachment (sanitization is enabled)

  • 5% of emails have JPG1 attachment (sanitization is enabled)

300 emails / minute

50-100%

Load balancing MetaDefender Email Gateway Security

For mail proxy and hosted deployment, this is desired configuration to increase the resilience and performance.

In order to achieve this, the following options should be considered.

  • Using load balancer (e.g., elastic load balancer, barracuda load balancer)

  • Using SMTP failover mechanism on your email gateway (configuration varies depending on the email gateway solution).

Sizing examples

The following displays two examples in real production environment. This will provide reference to sizing of the hardware hosting MetaDefender Email Gateway Security. Neither of the two setups in the examples experienced latency (more than 100 emails pending for process).

Please note that in the examples below MetaDefender Core and MetaDefender Email Gateway Security were deployed to separate physical machines.

Common parameters

OS

Windows Server 2012 R2 Standard 64

Instances

2 (HA)

Core version

MetaDefender Core 4.7.2

Email Security version

MetaDefender Email Gateway Security 4.0.0

Hardware profile

 

Emails / hour

Hardware Specification

organization A

10,000

8 CPU cores, 16 GB RAM, 140 GB HDD

organization B

1000

2 CPU cores, 8 GB, 100 GB HDD

Footnotes

1 Files that we used for testing have the following sizes:

  • PDF: 8 KB

  • DOCX: 220 KB

  • JPG: 650 KB

The result may vary depending on data set you select. We strongly recommend to conduct testing based on your organization's email traffic statistics. If you need assistance on the tools for testing, contact OPSWAT support.