Servers power up shown in the web monitor

[Staff 9972]

3 hours ago, revsplus said:

 in AirVPN datacenter for which this info has been made public, for example Alblasserdam, the servers are connected with 4 (four) or 2 (two) 1 Gbit/s links, or 2 (two) 10 Gbit/s links to 40 Gbit/s switches/uplinks and (two? I don't remember) high speed routers towards AMS-IX.

Hello!

You have the elephant's memory or you dug into datacenters' features. Let's not exaggerate though, the links are normally two and Juniper or Arista switches can start at 10 GbE, only some are 40 GbE or more. However, the stated bandwidth is guaranteed and dedicated as always. Now that the CPU is no more the first bottleneck, usually the limit is enforced on port. Specifically for AMS-IX we don't know whether our traffic relies on one high speed router or more, last year it was one. The total guaranteed bandwidth we pay for in the Netherlands is currently something around 150 Gbit/s in total.

Kind regards
 

[Staff 9972]

17 hours ago, arteryshelby said:

Take this example:
A server has 200 mbit outbound and 800 mbit inbound usage
with your accounting the usage would be 200 + 800 = 1000 from 2000 (halfduplex)
with my account the usage would be 800 (greater value) from 1000 (fullduplex) 

Hello!

The asymmetry you describe is an anomaly caused by malicious activity which is already foreseen. The upload/download ratio is monitored since 2011. If the absolute value of its difference from 1 is greater than a pre-determined safety parameter the alert system kicks in. The situation takes place only when malicious traffic breaks in through the perimeter defense. Users are alerted by the real time server monitor or by our software accordingly. A server under the condition you describe must not and will not operate.

Kind regards
 

[arteryshelby 25]

1 hour ago, Staff said:

Hello!

The asymmetry you describe is an anomaly caused by malicious activity which is already foreseen. The upload/download ratio is monitored since 2011. If the absolute value of its difference from 1 is greater than a pre-determined safety parameter the alert system kicks in. The situation takes place only when malicious traffic breaks in through the perimeter defense. Users are alerted by the real time server monitor or by our software accordingly. A server under the condition you describe must not and will not operate.

Kind regards
 

But asymmetric loadcase was the hole argument for why my calculation would be wrong but yours is right? The choosen example is a little extreme but the same principle applies for every asymmetric loadcase, like 150 mbit in / 200 mbit out. Will the system also detect this slight differences?
 

Quote

You may have asymmetries which make this quoted statement false. Even a scriptkiddie modest flood counts significantly, in this case, either inside the VPN, or incoming from outside the VPN itself.

I see the 800 mbit vs 1000 mbit difference but still think that my calculation is way more valuable from a customer perspective as you can easilly identify forwarding capacity without need to check inbound or outbound individually. Also the only way to check inbound and outbound usage is the graphs which are way behind in time (didnt really pay attention but they are perhaps updated every 15 minutes or so?)

Please discuss this internally.