MTU / MRU and bandwidth throughput?

Hi,

If anyone has previous experience/information on this issue I’d appreciate the input?

I am running a VPN between two sites over DSL/PPPoE lines.

The DSL Upstream Speed is 512 kbps, so the maximum throughput between the two VPN sites could be anything up to 512 kbps.

The pppoe interface reports on the status screen of an MTU = 1480 & MRU = 1492.

The mikrotik box has automatically created mangle rules which downsize the MTU = 1440 & MRU =1452.

This is obviously done to accommodate certain sites i.e. secure banking etc. that may require this.

If I only communicate intra VPN and not with the ‘outside world’ do I need this mangle rule?

Using the mikrotik bandwidth test client/server the average throughput with the mangle rule is 480 kbps, this is between the real ip addresses on the pppoe interfaces that the ISP is providing.

Could the mangle rule that reduces MTU & MRU sizes also be decreasing throughput?

The bandwidth throughput gets worse as I am running an EOIP setup over these DSL/PPPoE lines, as I progressively test deeper into the EOIP environment the more degraded the throughput gets i.e.

pppoe > pptp > eoip > bridge > ether - ether < bridge < eoip < pptp < pppoe.

Thanks

The problem as i see it, is that the biggest package you can send without fracturing it, is equal to the smallest MTU on your entire link from PPPoE to PPPoE, whether PPTP or ether and so forth. What MT do is to automatically test when this start happening and set your MTU for the appropriate size to prevent this. If your MTU is set up correctly on your interfaces, you don’t need it.