Routing Active Testing server

How sensitive is the management connection to the Routing Active Testing server?

The Test Agents are very robust to network disturbances affecting the management connection from the Test Agent to the Routing Active Testing server. However, like all applications that traverse an IP network, the underlying communication protocols (in this case TCP) do pose some requirements on the network connection.

Our internal tests show that Routing Active Testing starts to become affected at around 10% loss in combination with 100–200 ms one-way delay (200–400 ms round-trip delay). Even higher loss can usually be tolerated without noticeable performance degradation, provided that the delay stays low (one-way delay in the order of tens of milliseconds – not hundreds).

Conditions degrading the performance of the management connection are only seen in networks with significant problems or on low bit rate satellite connections. You will then notice things like slow and/or asynchronous updates of measurement graphs.

Is my measurement performance affected by the geographical distance to my Routing Active Testing server?

No, it is not.

No measurement traffic ever occurs between your Test Agents and the Routing Active Testing server. The Test Agents communicate with the server over an encrypted link, sending collected measurements and receiving control traffic.

Most of the post-processing of the measurement data is done on the Routing Active Testing server. All packet-level processing, however, is done in real time in each Test Agent to achieve best performance and accuracy. The Test Agents periodically upload their data to the Routing Active Testing server for further post-processing and storage.

Where are the Routing Active Testing cloud servers located geographically?

We use Amazon’s globally available data centers to host our cloud servers. We are continuously adding more servers on different continents as our customer base grows.