Server Status Load Indicator

[jipjipjuhee 0]

Hello, 

 

i have noticed, that the load indicator under server stats are the combined speed of incoming and outgoing traffic.  

e.g. Load=500Mbit -> rrd stats are showing 250MBit in and 250MBit out traffic... 

 

 

Why airvpn are using the combined speeds?

In my opinion most of the dataceners provide full duplex connections, so that the network interface should send and receive with full link speed simultaneously...? 

Are the connections limited in total used bandwith ?

 

 

[Staff 9893]

Hello!

 

The limit is given by the "weakest" element between network card, uplink port, line, router,,, In our case our servers, even if connected to a 10 Gbit/s port, have anyway a 1 Gbit/s network card, or have a maximum bandwidth allocation of 1 Gbit/s by contract. Or they are connected to a 1 Gbit/s port. Therefore reporting more than 1 Gbit/s would not be fair and correct toward our customers, because in reality they could never beat 1 Gbit/s.

 

Kind regards

[jipjipjuhee 0]

Hello Staff, 

 

thank you for your fast answering. But the root of my question was not answered. 

I'm understanding the diffrences between network interface card / port capacity / bandwith limit / contracts. 

 

I'm also aware that a connection can not exceed the bandwith limit.

 

But usually a network connection in full duplex mode is able to send and receive simultaneously (which normally allows sending and receiving simultanously at maximum connection speed e.g. GBit Connection allows to send 1GBit and receive 1GBit same time) 

 

Airvpn load indicator is showing "sum(in-bandwidth+out-bandwidth)" of the server. Which would be a maximum of 500MBit bandwidth for the user (user --[500MBit]>> airvpn server --[500MBit]>>  destination == 1000MBit total bandwidth usage).

 

But the default ethernet connection allowing 1000MBit per direction (in full duplex mode), so the load indicator in mind in my opinion should be "greatest(in-bandwith-usage, out-bandwidth-usage)" to display the corret limit of the connection.

 

so the question is: are the connections limited by total used bandwidth or by maximum connection speed (in half or full duplex mode)?

 

 

Kind regards

[zhang888 1066]

The servers are connected to 1Gbps ports in most of the locations. Since VPN users generate both uplink and downlink traffic each and every time, that speed has to be divided:

Imagine that you have to download a 100MB file, this means that during the download the server will actually transfer 2x of that data, first to get it from the origin and second to transfer it to you.

Basically, you are right - full duplex 1Gbps means 2Gbps maximum ideal transmission. Depends, as always, on all the components: NIC, cabling, and switches. Practically, this is never the case.

Ethernet duplex modes has nothing to do with that - the switches in the data-centers are 1Gbps both directions at all times, so the most correct representation would be the one you see in the

status page. There are no data-centers that have a different SLA model - the typical ones are either capped 1Gbps (sometimes to 100TB/month) or unlimited, unshared 1Gbps ~312TB/month max.