Bit surprised this works. Latency variability is huge and sometimes quite disconnected from geo location. I recall talking to someone in NL and realised I've got better latency to NL content from the UK than he did. Presumably better peering etc.
I work for IPinfo. We are launching a collaborative project with IXPs and major internet organizations to share raw measurement for routing and peering data for this purpose.
Latency variability is a huge issue. We run both traceroute and ping data, and we observe that there are entire countries that peer with IXP thousands of miles away in a different continent.
We bought a server from the oldest telecom company in the country and recently activated it. Currently, there is a 20 ms latency when traffic is directed towards the second oldest telecom. The packets have to travel outside the country before coming back in. This is a common phenomenon that occurs frequently. So, we usually have multiple servers in major cities since various ASNs have different peering policies.
For us we can map those behaviors and have algorithms and other data sources, make measurement-based geolocation perform well.
We are hoping to support IXPs, internet governance agencies, and major telcoms in identifying these issues and resolving them.
What is your path towards 'resolving' these issues?
I've done some mapping while comparing turn servers my org hosted on cloud vms vs a commercial offering, and it's pretty easy to find very different routing from point A to point B, but sometimes it's pretty clearly that not every transit network has access to every submarine cable, so traffic from say Brazil to South Africa might go from Brazil directly to Africa, or it might go to Florida, then Europe, then Africa. It'd be nice to take a more direct route, but maybe the Brazil -> Africa hop doesn't transit all the way, so BGP prefers the scenic route as it has a shorter AS path.
I didn't have any leverage to motivate routing changes though, so other than saying hmm, that's interesting, there wasn't much to do about it.
From our data side, we focus on network diversity and conduct continuous measurements. Due to the nature of our measurements and our knowledge of the precise locations of all 1330 servers, we understand how network packets travel across the internet. We simplify this information into algorithms and know how to accommodate detours that packets may take. There are specific patterns that we can identify and map, like some African servers route their traffic through LINX or a French IXP. If you are not connecting to private networks or even major telecoms on EU-based IXPs.
To help the system, we are reaching out to IXPs, major telecoms and peering agencies to advise them on how to peer and make critical internet routing decisions. We want to tell them on how to engage in data-focused peering, how their IXP is perceived from a broader internet data perspective, and how their packets from the IXP travel across the internet. We hope this colloboration will bring much needed efficiency in internet routing.
Could just be local loop latency, in VDSL or DOCSIS you can get 5-15ms of latency just in your first 1KM. London (e.g Telehouse) > Amsterdam is only about 7ms.
Wouldn't you just be closer to the closest PoP and requesting mostly cached content? With how connected amsterdam is they couldn't be around there. Also depending on when it was up until like 7-8 years ago even in major city centers there was no fiber in most places in NL. Now it's mostly covered.
Was a while back so bit fuzzy on what precisely we were measuring, but no wasn't something cached/CDN'd. Maybe a VPS or something not sure.
I was on a better connection (gigabit FTTC) and in a better peered location (central London).
>amsterdam
Don't know where precisely in NL they were or what connection type. I'd certainly expect a like for like amsterdam wired connection to win so this was probably something more pedestrian & rural
Yup. For example from my city to one of my dedicated server whose location is fully well-know (in France), I know there's 250 kilometers as the crow flies. Yet if I ping that server and draw a circle around my place (considering ping travels as fast as light in a vaccuum, which we know ain't happening but, hey, it's something) I get a radius of 2000 kilometers. About 8x the distance. I can prove that my IP ain't in the US but that's still not very precise.
And indeed many servers in the UK, which is 2x the distance than my server is, gives me constantly a lower ping.
TFA's approach, especially with the traceroute instead of Ping, is nice.
