Instantaneous transfer rate is always either the line rate of your network interface or 0. Anything meaningful is going to be some kind of average over some kind of interval.
I think your opinion of American cars isn't much more factual than the previous poster's opinion of French ones. Take a look at the new Ford Fusion or Dodge Dart for example.
It's not surprising that people living in an area with high crime would like to be able to legally defend themselves. Gangs and drug dealers don't care if it's legal or not.
There is a word for that.MTIE, packet delay variation, or just http://en.wikipedia.org/wiki/Jitter. Latency on any long haul route is going to be measured in Milliseconds whereas Jitter will be measured in Microseconds if you use the right equipment.
Because, they have not demonstrated that it can be mass-produced cheaply. They're still doing these in a lab. They may be using standard 90mm lithography process, but they're using non-standard wafers with some exotic bits stuck in there like germanium and carbon nanotubes. Whether they can produce this with the kind of success rate needed to make it worthwhile is yet to be seen.
I say misleading because both TFS and TFA headlines say commercially viable. When you actually dig into it you find out this has never left the lab. It also implies that something like this hasn't been done before. It has, just not on silicon. And, IBM still can't do it purely on doped silicon. If you read deeper the wafer has to include a germanium layer and carbon nanotubes for the optical components.
The hypothetical IBM chip is not a competitor to the Infinera device, it's a competitor to CFPs and CXPs produced using discrete optical and electronic components.
Infinera was at commercial scale around 7 years ago, at 100Gigabit speeds (10x10Gbit/s). They're very expensive, but cheaper than 10 discrete OTU2/OC192/10GbE LAN-PHY transponders with optics. From what I've read in article, IBM may possibly be able to use this to lower the cost of LR4 optics in routers, at least that's what they seem to be aiming at. It won't give us the ability to do anything we can't already do today, though.
It's also remarkably misleading. Infinera has been doing Photonic Integrated Circuits for a while now, but they're definitely not cheap. The only thing IBM may have pioneered is doing it on Silicon. Infinera uses Indium Phosphide.
I suspect they are actually doing some kind of fiddling in the electrical domain and calling it FDM, while the laser is still either On Off Keying or maybe Phase Shift Keying. Since DP-QPSK transmitters and receivers still cost about as much as a luxury car they're hardly COTS.
Nobody said anything about being a modern day competitive chip. I'll agree they've probably shrunk the die and increased the clock speed. But, since they haven't increased the L1 and L2 caches over the late 90s DEC version, I doubt they've done anything else radical either.
Next time you're driving look around and see how many vehicles have more than one passenger. Most of the time you'll have trouble finding a single one.
5 microseconds per kilometer tends to be a pretty good approximation, depending on the transport gear. Things like FEC, EFEC, dispersion compensation modules (non-bragg grating type), frequent OEO regens can add up and make it worse.
That would give you a ballpark of 11ms for a 1450 mile circuit.
Slashdot Mirror
On the Internet, it's Nazi's all the way down.
There is also the fact that nobody else who has reviewed the car had these issues, unless you count Top Gear who also fabricated their results.
Instantaneous transfer rate is always either the line rate of your network interface or 0. Anything meaningful is going to be some kind of average over some kind of interval.
They aren't a fabrication, they're a vocal minority that gets more coverage than they are worth because it helps people like you feel superior.
I think your opinion of American cars isn't much more factual than the previous poster's opinion of French ones. Take a look at the new Ford Fusion or Dodge Dart for example.
It's not surprising that people living in an area with high crime would like to be able to legally defend themselves. Gangs and drug dealers don't care if it's legal or not.
random variation in transmission speeds
There is a word for that.MTIE, packet delay variation, or just http://en.wikipedia.org/wiki/Jitter. Latency on any long haul route is going to be measured in Milliseconds whereas Jitter will be measured in Microseconds if you use the right equipment.
Because, they have not demonstrated that it can be mass-produced cheaply. They're still doing these in a lab. They may be using standard 90mm lithography process, but they're using non-standard wafers with some exotic bits stuck in there like germanium and carbon nanotubes. Whether they can produce this with the kind of success rate needed to make it worthwhile is yet to be seen.
How much does what IBM has done help us towards being able to produce photonic logic?
None of it. They're just working toward miniaturizing and reducing the cost of these things. https://www.google.com/shopping/product/8819852028889869930?q=LR4%20CFP
I say misleading because both TFS and TFA headlines say commercially viable. When you actually dig into it you find out this has never left the lab. It also implies that something like this hasn't been done before. It has, just not on silicon. And, IBM still can't do it purely on doped silicon. If you read deeper the wafer has to include a germanium layer and carbon nanotubes for the optical components.
The hypothetical IBM chip is not a competitor to the Infinera device, it's a competitor to CFPs and CXPs produced using discrete optical and electronic components.
Infinera was at commercial scale around 7 years ago, at 100Gigabit speeds (10x10Gbit/s). They're very expensive, but cheaper than 10 discrete OTU2/OC192/10GbE LAN-PHY transponders with optics. From what I've read in article, IBM may possibly be able to use this to lower the cost of LR4 optics in routers, at least that's what they seem to be aiming at. It won't give us the ability to do anything we can't already do today, though.
It's also remarkably misleading. Infinera has been doing Photonic Integrated Circuits for a while now, but they're definitely not cheap.
The only thing IBM may have pioneered is doing it on Silicon. Infinera uses Indium Phosphide.
Man, that was a painful read. I'm starting to feel sorry for Autonomy. Not everything has to be persecution, sometimes people are just a bad fit.
Some 80% of all first businesses fail, but only 20% of second businesses fail. That's because after the first business, you learn from your mistakes.
Or is it because the second businesses are started by the people who didn't fail the first time?
It doesn't give you man tits? Also, the million dollar electronics in a modern MotoGP or WSBK motorcycle make this app look like a childrens toy.
I suspect they are actually doing some kind of fiddling in the electrical domain and calling it FDM, while the laser is still either On Off Keying or maybe Phase Shift Keying. Since DP-QPSK transmitters and receivers still cost about as much as a luxury car they're hardly COTS.
It will still be encrypted after we have figured out what the plain text is. It will no longer be undecoded, though.
I haven't seen MIPS mentioned outside of TFSummary, so I think that was an editorial brain fart.
Nobody said anything about being a modern day competitive chip.
I'll agree they've probably shrunk the die and increased the clock speed.
But, since they haven't increased the L1 and L2 caches over the late 90s DEC version, I doubt they've done anything else radical either.
You're right, but it looks like they've done the latter. http://laotsao.wordpress.com/2011/10/29/sw1600-and-alpha-21164/
Next time you're driving look around and see how many vehicles have more than one passenger. Most of the time you'll have trouble finding a single one.
5 microseconds per kilometer tends to be a pretty good approximation, depending on the transport gear.
Things like FEC, EFEC, dispersion compensation modules (non-bragg grating type), frequent OEO regens can add up and make it worse.
That would give you a ballpark of 11ms for a 1450 mile circuit.
I couldn't say because the network I work on doesn't go that far west, but 350 E Cermak is the center of the Internet, if it has one.
The largest datacenter in the world is at 350 E Cermak Road in Chicago.