Companies like Gaikai and OnLive are doing acceptably low-latency streaming over the public internet on dinky home broadband connections. Are you sure that they can't do something relatively low-latency using GPU encoding over a gigabit ethernet home network?
Valve has talked about there being multiple tiers of steambox, with the lowest one being basically a thin client, for the streaming. In that case, yeah, your gaming desktop is going to be much more powerful than your terminal. And since most modern GPUs have dedicated hardware for video encoding, it's not going to cause much of a performance impact either.
But won't it, eventually? It's a chicken-and-egg problem. Few native Linux games because of a virtually non-existent install-base, a virtually non-existent isntall base because there are few native Linux games.
This could be an end-run around that. If the streaming works well enough, it could help get a lot of SteamOS boxes in the wild, which builds the install base, making it a more attractive target for native development.
Then as a serious question, how does this thing compare to existing embedded Atom or NUC boards? There's a Celeron NUC kit out there for $180ish and it's smaller than the Minnow to boot. There are also lots of fanless Pico-ITX boards with Intel chips.
Which is why raidz3 might be a better option for people with super critical data than raidz2:) I do get your point, though. It was for that reason that I migrated from raidz to raidz2; I didn't like the idea that if a single disk failed, I had no protection from read errors during a rebuild.
It's also worth commenting on what a few other people have said, which is that copies=2 doesn't put the data on different disks (it's not the same as mirroring). This is partially true; ZFS tries hard to get the data onto different disks and different vdevs, it's just not guaranteed (it depends on available space).
To be fair, Intel's Sandforce-based SSDs have been quite reliable, as have the ones used in Apple products. Are you sure you're not meaning "OCZ" when you say "Sandforce"?
The fingerprint readers in laptops work rather differently... and poorly. They're optical readers, and they work more like scanners in that they just capture a strip, and you have to swipe your finger over it. Having experienced fingerprint readers on a few different laptops, they don't work well (they're finicky and rarely want to read your fingers unless you swipe them just right).
Apple's approach is for a 2D sensor that doesn't use swiping. From the videos they've posted, it also seems to be much more willing to accept matches from different areas of the finger (so you don't have to position your finger very precisely). If it works as well as it is promised to, this would be a huge improvement over the fingerprint readers shipping in business laptops... but I'm still somewhat skeptical it will work as well as they claim. We'll see, I'm interested.
In the case of most drives, the key they ship with is randomly generated at the factory, unless you enable ATA passwords in your BIOS, which will prompt a new key to be generated, secured by that password. This is the typical behaviour; encrypt everything by default using the built-in key, and most support various external interfaces for securing that. Some even support eDrive, which integrates with BitLocker. Anandtech has a nice article about that:
Perhaps you're unaware that most modern SSDs these days do controller-level AES encryption of all data? Intel's drives do (as do any others based on Sandforce controllers), Samsung's newer ones do, Crucial's newer ones do... and the keys are stored in the controller, not the NAND.
It's kind of odd for you to say I'm on drugs for saying things that are on the spec sheets of the drives themselves...
Um, no, it doesn't. That would be dumb, as well as pointless, since the entire point of SSD encryption is to prevent the flash from being directly accessed like that. SSD controllers store the keys internally, not on the external flash.
What makes you think you can't take FLASH devices and access them in a similar way to platters?
Because on most SSDs, the data is encrypted, and on all SSDs, the pages are in an effectively random order. If you've lost the controller, you've lost both the encryption keys and the table that enables a logical platter-style presentation of the pages. No amount of soldering is going to fix those problem.
400G would definitely resolve that complaint, but shoving an enormous HDD into a tablet (roughly 35 cubic centimetres) is silly. That's a laptop-sized drive that would be too big for an ultrabook, let alone a tablet. Shrinking it down from 7mm to 5mm doesn't magically make it appropriate. That's many many times more volume than the eMMC in tablets consume... that's enough for an extra ~26 Wh of battery capacity, which is more than half the battery capacity of an iPad.
At the default threshold of 1.2, I've never seen it do any of those things (have any negative impact on low-contrast areas or lineart). Even on very low-contrast textures. And yet at those same settings it significantly reduces the appearance of banding. Perhaps you're applying it at a much much higher threshold than I am?
To be fair to them, animated content with its frequent use of simple gradients does get a bigger advantage out of 10-bit encoding (avoiding banding) than film content does. Then again, the ffdshow deband filter (GradFun2DB) does a fantastic job at sorting out banding issues on playback without degrading image quality regardless of content type, so that certainly limits the utility of 10-big encoding.
The Kinect does not use stereoscopy. It projects a IR laser grid (well, randomized) and then uses a single monochrome IR camera to see the grid and calculates depth from that. There are two cameras in the Kinect, an RGB camera and a monochrome IR camera, but only one is used for depth.
Tesla already has a manufacturing deal with Panasonic, who makes all their cells. They don't buy the thing on the open market. I'm pretty sure that if Tesla wants more batteries, Panasonic can and will ramp up production.
What's your point? Their definition of "internet connectivity" counts a datacenter as a broadband ISP... I've already got 50 megabit VDSL2 service that I "PAY" for.
All of whom are limited to offering the same services offered by Bell and Videotron. Sometimes they can't even match the incumbent pricing due to the ridiculous CBB rates approved by the CRTC.
Right, except OVH in Montreal is a datacenter. They really do have an obscene amount of bandwidth to the rest of the internet. The problem is that the "list" is counting a datacenter as a broadband ISP. They're not, they're a datacenter.
Plunking an enormous datacenter down next to a city doesn't suddenly make it a futuristic super internet city...
They list Montreal, and their primary reasons are laughable.
They say Montreal does very well in speedtests because of... OVH. Wait, what? That's a dedicated server and cloud services provider, they have nothing at all to do with consumer broadband in Montreal. Maybe this is a positive for businesses, but it has zero bearing on your average Montrealer. The second reason is the Ile Sans Fil people, who install free wifi access points... except their coverage is non-existent. They've got 260 access points. There are at least that many access points in my apartment building alone; 260 access points in a metro area of nearly 4 million people means that you can wander the city and will probably never see an Ile Sans Fil access point. I've seen them on rare occasions, but never successfully connected to one (I've tried).
Including Montreal in a list of "top internet cities" pretty much invalidates your entire list...
It's somewhat confusing that there are now two separate and mostly unrelated TOS continuations (well, Vic Mignona is involved in both) shooting on the same soundstages (specifically the ST:NV ones), though. ST:NV has changed most of the actors one or more times anyhow, so it's not entirely clear to me why they're not collaborating more directly seeing as how they're already using the same facilities.
He's playing Doc Zimmerman. He didn't want to reprise the role of The Doctor because he's aged too much, but when it was suggested that Zimmerman would have aged the same as he did, he was onboard.
Slashdot Mirror
Companies like Gaikai and OnLive are doing acceptably low-latency streaming over the public internet on dinky home broadband connections. Are you sure that they can't do something relatively low-latency using GPU encoding over a gigabit ethernet home network?
Valve has talked about there being multiple tiers of steambox, with the lowest one being basically a thin client, for the streaming. In that case, yeah, your gaming desktop is going to be much more powerful than your terminal. And since most modern GPUs have dedicated hardware for video encoding, it's not going to cause much of a performance impact either.
But won't it, eventually? It's a chicken-and-egg problem. Few native Linux games because of a virtually non-existent install-base, a virtually non-existent isntall base because there are few native Linux games.
This could be an end-run around that. If the streaming works well enough, it could help get a lot of SteamOS boxes in the wild, which builds the install base, making it a more attractive target for native development.
Then as a serious question, how does this thing compare to existing embedded Atom or NUC boards? There's a Celeron NUC kit out there for $180ish and it's smaller than the Minnow to boot. There are also lots of fanless Pico-ITX boards with Intel chips.
Which is why raidz3 might be a better option for people with super critical data than raidz2 :) I do get your point, though. It was for that reason that I migrated from raidz to raidz2; I didn't like the idea that if a single disk failed, I had no protection from read errors during a rebuild.
It's also worth commenting on what a few other people have said, which is that copies=2 doesn't put the data on different disks (it's not the same as mirroring). This is partially true; ZFS tries hard to get the data onto different disks and different vdevs, it's just not guaranteed (it depends on available space).
What are the chances of the exact same sector being corrupt on at least three disks in a raidz2 vdev? This doesn't seem like a plausible scenario.
To be fair, Intel's Sandforce-based SSDs have been quite reliable, as have the ones used in Apple products. Are you sure you're not meaning "OCZ" when you say "Sandforce"?
The fingerprint readers in laptops work rather differently... and poorly. They're optical readers, and they work more like scanners in that they just capture a strip, and you have to swipe your finger over it. Having experienced fingerprint readers on a few different laptops, they don't work well (they're finicky and rarely want to read your fingers unless you swipe them just right).
Apple's approach is for a 2D sensor that doesn't use swiping. From the videos they've posted, it also seems to be much more willing to accept matches from different areas of the finger (so you don't have to position your finger very precisely). If it works as well as it is promised to, this would be a huge improvement over the fingerprint readers shipping in business laptops... but I'm still somewhat skeptical it will work as well as they claim. We'll see, I'm interested.
LSI, who make the Sandforce chipsets, says otherwise:
http://www.lsi.com/technology/duraclass/Pages/Automatic-Encryption.aspx
Fuse-based OTP (one time programming memory) for unique master key
In the case of most drives, the key they ship with is randomly generated at the factory, unless you enable ATA passwords in your BIOS, which will prompt a new key to be generated, secured by that password. This is the typical behaviour; encrypt everything by default using the built-in key, and most support various external interfaces for securing that. Some even support eDrive, which integrates with BitLocker. Anandtech has a nice article about that:
http://www.anandtech.com/show/6891/hardware-accelerated-bitlocker-encryption-microsoft-windows-8-edrive-investigated-with-crucial-m500
Perhaps you're unaware that most modern SSDs these days do controller-level AES encryption of all data? Intel's drives do (as do any others based on Sandforce controllers), Samsung's newer ones do, Crucial's newer ones do... and the keys are stored in the controller, not the NAND.
It's kind of odd for you to say I'm on drugs for saying things that are on the spec sheets of the drives themselves...
Um, no, it doesn't. That would be dumb, as well as pointless, since the entire point of SSD encryption is to prevent the flash from being directly accessed like that. SSD controllers store the keys internally, not on the external flash.
What makes you think you can't take FLASH devices and access them in a similar way to platters?
Because on most SSDs, the data is encrypted, and on all SSDs, the pages are in an effectively random order. If you've lost the controller, you've lost both the encryption keys and the table that enables a logical platter-style presentation of the pages. No amount of soldering is going to fix those problem.
When you're talking about 8mm thick tablets, that's a pretty hefty increase to get some spinning rust in there.
400G would definitely resolve that complaint, but shoving an enormous HDD into a tablet (roughly 35 cubic centimetres) is silly. That's a laptop-sized drive that would be too big for an ultrabook, let alone a tablet. Shrinking it down from 7mm to 5mm doesn't magically make it appropriate. That's many many times more volume than the eMMC in tablets consume... that's enough for an extra ~26 Wh of battery capacity, which is more than half the battery capacity of an iPad.
At the default threshold of 1.2, I've never seen it do any of those things (have any negative impact on low-contrast areas or lineart). Even on very low-contrast textures. And yet at those same settings it significantly reduces the appearance of banding. Perhaps you're applying it at a much much higher threshold than I am?
To be fair to them, animated content with its frequent use of simple gradients does get a bigger advantage out of 10-bit encoding (avoiding banding) than film content does. Then again, the ffdshow deband filter (GradFun2DB) does a fantastic job at sorting out banding issues on playback without degrading image quality regardless of content type, so that certainly limits the utility of 10-big encoding.
The Kinect does not use stereoscopy. It projects a IR laser grid (well, randomized) and then uses a single monochrome IR camera to see the grid and calculates depth from that. There are two cameras in the Kinect, an RGB camera and a monochrome IR camera, but only one is used for depth.
Better get on the horn to Panasonic and Samsung.
Tesla already has a manufacturing deal with Panasonic, who makes all their cells. They don't buy the thing on the open market. I'm pretty sure that if Tesla wants more batteries, Panasonic can and will ramp up production.
What's your point? Their definition of "internet connectivity" counts a datacenter as a broadband ISP... I've already got 50 megabit VDSL2 service that I "PAY" for.
All of whom are limited to offering the same services offered by Bell and Videotron. Sometimes they can't even match the incumbent pricing due to the ridiculous CBB rates approved by the CRTC.
Right, except OVH in Montreal is a datacenter. They really do have an obscene amount of bandwidth to the rest of the internet. The problem is that the "list" is counting a datacenter as a broadband ISP. They're not, they're a datacenter.
Plunking an enormous datacenter down next to a city doesn't suddenly make it a futuristic super internet city...
They list Montreal, and their primary reasons are laughable.
They say Montreal does very well in speedtests because of... OVH. Wait, what? That's a dedicated server and cloud services provider, they have nothing at all to do with consumer broadband in Montreal. Maybe this is a positive for businesses, but it has zero bearing on your average Montrealer. The second reason is the Ile Sans Fil people, who install free wifi access points... except their coverage is non-existent. They've got 260 access points. There are at least that many access points in my apartment building alone; 260 access points in a metro area of nearly 4 million people means that you can wander the city and will probably never see an Ile Sans Fil access point. I've seen them on rare occasions, but never successfully connected to one (I've tried).
Including Montreal in a list of "top internet cities" pretty much invalidates your entire list...
The problem is that L5 did one pilot, it ended on a cliffhanger, and a year and a half later there's no indication of any work to continue it.
It's somewhat confusing that there are now two separate and mostly unrelated TOS continuations (well, Vic Mignona is involved in both) shooting on the same soundstages (specifically the ST:NV ones), though. ST:NV has changed most of the actors one or more times anyhow, so it's not entirely clear to me why they're not collaborating more directly seeing as how they're already using the same facilities.
He's playing Doc Zimmerman. He didn't want to reprise the role of The Doctor because he's aged too much, but when it was suggested that Zimmerman would have aged the same as he did, he was onboard.