Hacker News Reader

Comment by walrus01 | Hacker News Reader

satellite (earth station) systems engineer here: I highly doubt that there will be access at anything other than layer 3. The CPE will have a 1000BaseT port on it, when you connect to it, it'll hand out a DHCP v4 and v6 IP address on it. Maybe if you pay an extra $15 per month you can get a single static v4 address. I am very doubtful that it will be possible for an individual end user with a rooftop terminal to get access to "the network" at what would correspond to layer 2 in the OSI model.

It is possible that a more expensive service tier offering might provide L2 ethernet fabric virtual wire adjacency. Similar to the general idea of an VPLS / EoMPLS virtual wire tunnel, where whatever you plug into one end will see the single MAC address (or set of MAC addresses) of your other equipment on the other end. But this will be a managed service and not something that you can build arbitrarily.

View on HN

Thank you. You're the only "hacker" here among evangelicals and dreamers. Thanks for your informative comments on "hacker" news.

Without evangelicals and dreamers major new ISP infrastructure wouldn't get built. But it has to run into practical realities of what a network will look at all the way down the OSI model stack. People want gigabit FTTH to every home in the USA but the construction costs could be $5,000 per house on average.

I followed this closely and thank you for pointing out the cost implications per house connection. While that's a significant amount in the USA,it's far monstrous in the developing countries/rural areas;which from this forum are touted as the major targets of the SpaceX system. Do you see a way of reducing this cost that can practically serve a rural town in, for example, Africa? (I'm from East Africa and trying to make sense of what this new development means for local ISPs)

If you have a town/city of population 2,500 to 20,000 that has very poor internet connectivity in general at present, this will probably be most useful in its larger version to bring one relatively high capacity DIA/IP link (like 150 Mbps downstream x 50 Mbps upstream) to a single central location, as an ISP POP in the town, and then local distribution throughout a US county-sized area by point-to-point and point-to-multipoint terrestrial microwave in the 2.4, 3.5, 3.65, 5.2 and 5.8 GHz bands. Possibly with some use of low-cost equipment in the traditional licensed (6, 11, 18, 23) GHz bands, unlicensed 24 GHz band PTP, 60 and 80 GHz, etc.

Much appreciation for this feedback. I definitely need to dig further, but your comments are painting quite a discernable image of what the system means from an end user POV. We still got almost 5 years to its realization but something worth keeping tabs on.

Reply on news.ycombinator.com