Monitoring Statistics on ILP


Hi all,

I’m working on a project to gather statistics on the Interledger livenet. This would be a service which both clients and connectors could use to measure statistics such as latency, speed, and exchange rates. I’m fairly new to the network and would love to have your feedback on a few questions.

  • How could I use the ping/echo protocols in the connector implementations?

  • How exactly does peering work in the network and how could I structure peering with connectors or other clients for my service?

  • Could this be used to discover the centrality of certain nodes and the existence of certain routes and their qualities, and thus determine who a new node might want to peer with?

  • Related to the above, what kinds of metrics could be used to determine the “quality” of a route?

Looking forward to hearing your thoughts!

Community Call - April 17, 2019

Hey - this sounds awesome, and super useful. Working on a simple public connector uptime site, would be great to collaborate. Join the slack channel for open network tools? ( --> #open-network)

Will dm to follow up.

1 Like

This sounds pretty useful. I’m pretty sure there is something that helps with an element of this either when peering or when sending a payment. I remember @adrianhopebailie talking briefly about packet routing and some of this seems to be taken into consideration if I remember correctly.

Available liquidity and trust limits may be something else people may want to review and consider when making peering decisions and an index of these values would be great.

1 Like

Ping / echo – you could identify the ilp address of each connector and get a measure of the latencies by pinging them. This would give you a standard to compare the latency of each path from the viewpoint of your connector.

Peering – currently this is a very manual process and you would need to reach out to the various parties hosting connectors and ask them to set up a plugin for you. Then you would add a plugin on your connector that connects to the plugin that they set up on theirs. The best plugin to use for the open network right now would likely be this.

Centrality / routing – yes, the more peers you have the greater visibility you will have in the network.

Quality of route – the two primary metrics one would likely be concerned with in an ILP route would be exchange rates and latency.


Great idea – this would be very helpful also for tutorials and documentation purposes such that they always list available connectors.

While debugging several moneyd instances on the testnet in the last few days, I’ve created a small tool to ping nodes/connectors: ilp-ping
The code is basically a cli-wrapper around the ping functionality of moneyd-gui. It’s in a very early stage but it can measure the min/avg/max/stdev latency to a connector. Feel free to use/adapt it for your monitoring approach.


I think another interesting statistic would be some kind of measure of bandwidth on a per-path basis. For example, if a particular path is super-fast in terms of packet latency, and has good FX, but only allows $1 per hour, then it might not be useful if I want to send, for example, $100 in the next minute.