From: damarkley@gmail.com   
      
   On Wednesday, August 19, 2020 at 5:28:55 PM UTC-4, JF Mezei wrote:   
   > On 2020-08-19 16:13, Jeff Findley wrote:   
   >   
   > > Since when? It's pretty bog standard to run a speedtest like ookla. It   
   > > tests the bandwidth of your Internet connection, measured in Mbps over a   
   > > short span of time.   
   > A beta with a couple of users is not representative of what it will be   
   > like in production because it does not represent the level of   
   > oversubscription that Starlink intends to have to make the service   
   > profitable.   
   > > Which is why users will no doubt be running speed tests periodically to   
   > > see what they're getting. Obviously SpaceX already tracks all that   
   > > information.   
   > A fixed number of beta testers in a region with fixed number of base   
   > station(1) will yield the same results and show stable service. They   
   > already know how many periuods during the day there would be   
   > interruptions in service and already know how each launch will reduce that.   
   >   
   > But this beta allows Starlink to tests the residential base station   
   > hardware and software that tracks satellites to orient antenna and   
   > impact of westher. So Starlink is far more interested in signal strength   
   > vs weather radar images than it would be of a speed test.   
   >   
   > Note also that Starlink allowed base statiosn to connect to satellites   
   > far lower above horizon in order to decreate the numebr of gaps in   
   > service between satellite passes. So Starlink will be very interested in   
   > knowing about signal strength at variious elevation of staellite vs   
   > residential base station as this may help determine the eventual true   
   > limit of how low above horizon a base station will attempt to look for a   
   > staellite.   
   > > Also, my cable Internet service is shared with everyone else in the area   
   > > that's also using that same service.   
   > And the cable company has extact statistics on node traffic for your   
   > neighbourhood and when it reaches a certain level, begins process of   
   > splitting the node into 2 nodes so that this happens before se5vice on   
   > the 1 node degrades below its service standards.   
   >   
   > Some cable companies have better standards than others.   
   >   
   > Complaints from customers on slow speed tests do nothing since the   
   > cableco has exact statistics on node usage taken at the CMTS port and   
   > they know very well once customers start to see service degradation.   
   > > Latency to servers users actually use (e.g. a gaming server) will get   
   > > better when the laser interlinks between satellites come online.   
   > Since the current fleet of satellites do not have the lasers physically   
   > installed, it will be a long while before inter satellite links are of   
   > use. And this will not really change latency. It is very doubtful that a   
   > user in Port Henry NY connecting to a game server will have satellites   
   > be aware that the destination is in Seattle and use intersatellite links   
   > to bounde to a satellite that has view of the Seattle ground station   
   > whyere it can had short ground hop to gaming server.   
   >   
   > Such optmizations will likely be possible for permanent visrtual   
   > circuits for enterprise customers (point to poing connections that   
   > provide layer 2 connectivity) compared to normal Internet serfvice at   
   > layer 3 which would require Satellites to have full Internet BGP routing   
   > table and understand the concept of anycast to know where the nearest   
   > servre is.   
   > > That's still going to be a *lot* better than satellite Internet from a   
   > > geosynchronous satellite.   
   > These services are grossly oversubscribed in order to make them   
   > profitable. And like with Starlink, they are limited in capacity of   
   > ground ulink to statellite which is shared by all users. This is why   
   > the numebr of ground stations that Starlink will deploy will be the   
   > determining factor of how much capacity the system has. And if you are   
   > in Port Henry and you are directed to a ground station in Albany, you   
   > then travel via ground fibre to Seattle and will have worse service than   
   > ground based Internet from your home which doesn't have the roughly   
   > 1000km hop up/down from satellite.   
   >   
   > There is no question that LEO can theoretically beat the pants off a GEO   
   > satellite service. But when you introduce the business aspect of how   
   > much revenue you get and how much its costs, the question is still not   
   > clear.   
   >   
   > There are goo reasons why Starlink or any other satellite service   
   > publish the uplink capacity for for their service. It is a key component   
   > of how much oversubscription is needed to make service profitable and   
   > how terrible the service will be.   
      
   Have you considered starting your own group "I Hate Elon Musk and Everything   
   He is Affiliated With"?   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)