shaklee3

27 ms·

301.
by shaklee3 4y ago
that's partially true. the data rates have a lot to do with frequency reuse, which means more complicated designs. starlink is complicated in the dynamic conditions, but the beamforming is relatively simple.
302.
by shaklee3 4y ago
thanks. couldn't have said it better
303.
by shaklee3 4y ago
I think you conveniently forgot to include NRE and OPEX
304.
by shaklee3 4y ago
I know you're usually here to promote starlink, but all ookla data shows starlink is 40+ms on average. it's dishonest to pretend 20ms will ever be normal.
305.
by shaklee3 4y ago
what you're experiencing is what everyone in the industry already knew, but we got slammed by musk fans for pointing it out years ago. the ping will never be 20ms for most people. path loss is only a small part of a ping.
306.
by shaklee3 4y ago
as mentioned previously, I worked on exactly this for 15 years, and I also don't understand how people with knowledge of the industry believe SpaceX is going to be profitable on starlink without some course correction. nobody is debati
307.
by shaklee3 4y ago
this was the industry I worked in for over a decade and your comment comes off as someone that doesn't know the economics. there have been many, many articles of people in this industry showing that SpaceX cannot sustain this burn rate
308.
by shaklee3 4y ago
iridium rates compared to SpaceX are not comparable and not the same technology
309.
by shaklee3 4y ago
it's extremely hard to do at those data rates. if it wasn't there would be existing examples. Facebook tried and failed.
310.
by shaklee3 4y ago
bandwidth per dollar is what matters, and Leo is not as good as geo in that regard. if you threw the same amount of money SpaceX is putting into starlink into a viasat-3 you'd have more bandwidth and better coverage.
311.
by shaklee3 4y ago
which is exactly what starlink will be when it has enough customers. it's already having a slow decline in speeds as the network fills, and it's not even close to capacity in most areas.
312.
by shaklee3 4y ago
geo companies don't care much about launch cost since it's infrequent, and SpaceX has not shown starlink has a path to be profitable. I know it's the cool thing to do to love SpaceX, but separate the launch from the satellite
313.
by shaklee3 4y ago
in terms of bandwidth they own more than all those others. it doesn't matter how many satellites you have if the capacity of each is tiny.
314.
by shaklee3 4y ago
this is completely incorrect. viasat's technology is far more advanced than starlink. just because it's not LEO doesn't mean it's simple. and starlink doesn't have dramatically increased capacity. they have a modera
315.
by shaklee3 4y ago
vpp uses dpdk
316.
by shaklee3 4y ago
dpdk has rdma/GPUdirect now as well
317.
by shaklee3 4y ago
this is completely wrong. as the article points out, there are plenty of tcp stacks and other user space stacks available. dpdk is used in places where performance matters, and just because you haven't worked in that area doesn't
318.
by shaklee3 4y ago
dpdk is more equivalent to xdp. ebpf is not really comparable since it can't do large programs within a bpf kernel. dpdk is much, much more than just a packet processing framework.
319.
We discovered severe thermal throttling with Apple's new M2 MacBook Pro(twitter.com)
10 points by shaklee3 4y ago|2 comments
320.
C++: Assignment for Optional(brevzin.github.io)
4 points by shaklee3 4y ago|0 comments
321.
by shaklee3 4y ago
there is no cache. it's an analog repeater.
322.
by shaklee3 4y ago
this is not the case in the USA either. this sounds either exaggerated or the OP is leaving something out.
323.
by shaklee3 4y ago
Rust does not have a memory model.
324.
by shaklee3 4y ago
valgrind came out 20 years ago, so yes, it definitely existed 10 years ago.
325.
by shaklee3 4y ago
and they hired Travis Downs, who is probably even more knowledgeable about optimization.
326.
by shaklee3 5y ago
they use it for the control plane. it's clearly not used for actually submitting trades.
327.
by shaklee3 5y ago
you could also write a similar story about the predatory licensing of cplex and gurobi. when gurobi started their model was "we're the opposite of cplex, and we don't count cores". but they realized that wasn't maki
328.
by shaklee3 5y ago
cplex comes with a great heuristic now (odh), and gurobi too.
329.
by shaklee3 5y ago
gurobi and cplex agree light-years ahead of the open source solvers. those will work on small problems, but you need the expensive ones for complex models.
330.
by shaklee3 5y ago
can confirm. used to work for an ISP and they game it, and the reason is that everyone else does. you aren't supposed to game the fcc broadband site, though, since they use they as a proxy for internet speeds in the US.

More ›