Live data from Hacker News

Map of worldwide undersea cables

cablemap.info

21–30 of 51 posts

Re: Map of worldwide undersea cables

#22
Not very surprising, but there is a perfect correlation between the major "junctions" and Amazon's 4 data centers spanning the world, situated in North California (us-west-1), North Virginia (us-east-1), Ireland (eu-west-1) and Singapore (ap-southeast-1).

Recently I published an article describing how we find the closest data center using GeoIP and indexing, might be of interest. http://www.turnkeylinux.org/blog/geoip-amazon-regions

Re: Map of worldwide undersea cables

#23
post #4

If this interests you, Neal Stephenson's chronicle the FLAG cable getting laid is required reading: http://www.wired.com/wired/archive/4.12/ffglass_pr.html

I just wanted to post that too, you beat me to it.

Seriously, go read it. It's an amazing report.

Re: Map of worldwide undersea cables

#24

Pretty interesting stuff. I was hoping to correlate some of my traceroutes with these cables to figure out how my own international traffic is routed, but it seems that with the rise of optical routing, not enough intermediate hops show up for me to figure it out. For example, San Jose to Copenhagen seems to be a single hop for me, so I have no idea what transatlantic cable my data takes: xe-9-3-0.sjc10.ip4.tinet.net…

It could be Layer II routed - then the intermediate hops won't show up in traceroute (a Layer III protocol)

Re: Map of worldwide undersea cables

#25
post #11

what's the diameter of one of these cables? are they literally laid on the ocean floor?

You can see one at http://www.itnewsafrica.com/?p=3111 And yes, they are laid on the floor, mostly. In shallows they may be armored against sharks or plowed in to protect against anchors. The Neal Stephenson, WIRED article is an excellent read, but honestly unless you have a new Safari or some bookmarklet to extract the content from the surrounding flashing lights and popping ads you will probably go blind before you…

Printable version: http://www.wired.com/wired/archive/4.12/ffglass_pr.html

Re: Map of worldwide undersea cables

#26

Why is there a 5TBps line from the northern part of mainland Norway up to Svalbard? Is there much telecomms there?

I was going to comment on that one too: looks like they have a lot of bandwidth for the population.

Aside from the "research" answers, it's probably the government trying to give them something to do besides drinking all winter long:-)

From Wikipedia: "The most notable feature of Longyearbyen's climate is the lack of sunlight for over 4 months of the year in winter"

Re: Map of worldwide undersea cables

#29
post #18

Why is there a 5TBps line from the northern part of mainland Norway up to Svalbard? Is there much telecomms there?

From Svalbard you can see pretty much every polar orbit once per orbit, plus Norway is politically stable and very liberal with research licenses there, plus it’s warm for the latitude. So there’s demand from the kind of people who have polar-orbiting satellites that produce a lot of data.

Just to confirm your arguments, I'm quoting from from http://www.aftenposten.no/english/local/article545833.ece: "Although stretching cable for about 3,000 kilometers from Andoeya to Longyearbyen seems inconvenient, the spot has been carefully chosen. Svalbard Satellite Station (SvalSat) has specialized in retrieving data from satellites in polar orbit..."

And then from http://www.ksat.no/Products/Svalsat.htm: "The satellite coverage at this latitude holds unique opportunities and SvalSat is the only commercial ground station in the world able to provide all-orbit-support (14 of 14 orbits) to owners and operators of polar orbiting satellites."

Re: Map of worldwide undersea cables

#30
post #24

Pretty interesting stuff. I was hoping to correlate some of my traceroutes with these cables to figure out how my own international traffic is routed, but it seems that with the rise of optical routing, not enough intermediate hops show up for me to figure it out. For example, San Jose to Copenhagen seems to be a single hop for me, so I have no idea what transatlantic cable my data takes: xe-9-3-0.sjc10.ip4.tinet.net…

It could be Layer II routed - then the intermediate hops won't show up in traceroute (a Layer III protocol)

Ah yeah, that's probably more likely than layer 1 routing. There does seem to be an increase in it, though; unless I'm mistaken, I seem to recall that 10 years ago it was a bit easier to get a rough idea of the geographical hops a packet was taking. It seems quite a few large providers these days only do one layer 3 hop, so all you see is the entry and exit at the borders of their network.
Post reply on HN