Most storage systems I've worked with either strongly advise or outright require drive sizes > 1 TB to use a double parity system like RAID-6 due to the lengthy rebuild times associated with large individual members.
Flash performance is so far ahead of its bulk capacity, though, that capacity can scale without performance appearing to be a factor in restoring redundancy.
and yes, the rebuild times for these will be tremendous and you can't really think about your data as "RAID sets" anymore, that model is quickly becoming outdated. You have to think about "storage nodes" as a single, really large hard drive.
A distinction without a difference. Whether the data is made device redundant by some kind of RAID system or some kind of network copying, drive performance plays in significantly in restoring redundancy after a node failure.
I think the value of these drives is dubious if their redundancy rebuild rate is measured in days due to read/write rates not scaling. They may have niche use cases (ie, containing a complete copy of some large storage quantity on a single medium) but I'd also worry that the density would call into question their durability.
If PV is so economically valuable, why does it need subsidies?
I'm always kind of boggled by sectors that get killed when their subsidies get removed. Sounds to me like the product or sector isn't economically viable if it needs a subsidy to remain competitive.
In terms of "winning" some industrial race to produce PV panels, does it matter? Even if the US ends up developing some next-gen panel design that's a quantum leap forward, it seems unlikely to get manufactured at scale here anyway. The same basic economics that causes it to work out to have China build TVs or refrigerators or everything else at massive scale would seem to wind up working for solar panels, too.
Even if China "owns" the designs, even the Chinese can only absorb so many panels and will likely see the wisdom in building them at scales that feed both domestic and export markets or licensing their technology to other manufacturers.
It's nice to dream about a groovy green energy industry in the US where we design, manufacture and install next-gen PV exclusively at home, but I think modern economics are too sophisticated these days. Capital will seek the lowest cost production for the product that produces the products that buyers identify as providing the utility they want for the money they will spend.
This is an impressive density, but my question is whether drive performance increases and what kind of bus will the drive have?
SATA 3.2 claims 16 GBits/second but my guess from glancing at the other 3.2 features would indicate this is mostly a flash-oriented spec that would be tied to M.2 slots, and that the ordinary SATA slots would still be SATA-3 @ 3 Gbps.
Without a bus and drive combination capable of moving 40 TB in a reasonable amount of time (4-6 hours), these drives will be a novelty and not useful in any storage system without double or triple parity redundancy. The rebuild times without drive speed improvements and a fast bus would make them just time bombs.
I thought there were efficiencies to be had from a self-driving car being able to optimize acceleration, braking, and with car-car communication, greatly reducing a lot of the stop/go in congestion. It should be able to reduce the power consumption of the engine/motors over the way humans accelerate, brake, and speed up and slow down too often.
These won't make up all the losses to self drive, but won't they offset some meaningful percentage? 5% or so power consumption from the drive system through efficient power application should provide a meaningful buffer for the self-drive system.
I also wonder if the self-drive system couldn't partially sleep some part of the sensor environment and processing through some complex decision system involving time of day, real-time congestion reports, road condition, weather, road geography and car-car reporting data. Basically if you're on a long, straight street with no detectable traffic at 3 am in clear weather, couldn't the car dial back the sensor operation to draw less power? Or even if it kept them all running, obtain some efficiency because there's just less change data?
Now that they are starting to re-use rockets and are successfully landing them, have they crossed some magic threshold where their launches are now much cheaper than their competitors using disposable rockets?
Or are they still having to charge a premium due to R&D investments into their system?
If they aren't starting to reap cheaper launch costs, when will they? I would think that while the reusable rockets is an interesting design goal, it would need to cut launch cost meaningfully to be really beneficial.
Amazon used to make available data on where a thing you were browsing was popular. I mostly remember books, and they would tell you the regions or areas where it was being sold. "#1 In Minneapolis, #5 in Cambridge" that sort of thing.
If they still had that, you could figure out where sex lube was most popular and grant that area the anal capital of america title.
I kind of agree on corporate cash hoards being out of circulation, although I think they do park the money in some kind of short term securities (like Treasuries or similar highly liquid short-term bonds) so it's doing something.
In the case of Apple specifically, I wonder if part of the reason Apple gets such a good tax deal from Ireland is that there is a quiet agreement that Apple will park a certain portion of it for a longer period in Irish banks, thus enhancing the lending base of Irish banks and ultimately benefiting the Irish economy.
The DOJ should have split Microsoft into Operating Systems and Severs & Applications.
The latter would have been freed to pursue other operating systems as host platforms for their Servers and client Applications.
The former would be able to stop crippling the OS for the server and application platforms as well as encourage other application and server platforms to be hosted on the OS.
Either way, both would be freer to innovate without constantly having to be compromised for their siblings needs.
Beyond that, they could drastically simplify their pricing and versioning schemes. I've often wondered if MS wouldn't actually make more money with a flat rate $100 per license for Windows. The overhead of various licensing schemes and product variations has to eat into profits
But with Apple in particular and some others, isn't it already lesson learned? They're not really innovating with their cash. No (or at least few) failed products introduced, and most of that cash tucked away and guarded by a Leprechaun somewhere in Ireland.
While I agree with you philosophically, I think in terms of an AV program on Windows you're dealing with a unique set of vulnerabilities and a black hat state organization would want to know every detection technique and evasion detection trick they could. It's kind of a fundamentally insecure environment to begin with.
I suspect that we're headed totally into rental, but not entirely driven by greed. I think part of it is driven by rapid software changes and immense code bases which defy traditional hard revisioning -- not technically, but the overhead of providing patches just becomes too great.
I'm also inclined to think that someone is going to figure out how to arbitrage "subscription" software into basically sub-letting subscriptions. Buy enough licenses and enforce micro-rental periods and re-rent the software out to people who want to use it for a day or an hour.
Actually, at those time periods I kind of like subscription software concepts -- I can use what I need for a small cash outlay and not get invested in more software then I really need. Like renting a truck or something, where you just need to do something once.
They don't care. Nobody in charge of anything cares because they are only concerned about next quarter and what comes after.
I actually have a theory that the short-horizon world we live in is some weird byproduct of the baby boomer generation. They're all zeroing in on retirement and are just looking to maximize their retirement income and they literally unconcerned about stuff past about age 80.
I think it will take about another 10 years, when the boomers are nearly all retired or dead, for anything meaningful to change.
...will turn out to be tricking its customer base into renting rather than owning its software. He bought off the fiefdoms by picking winners and turning them into rent-receiving franchises.
As long as few viable alternatives exist for Office and Exchange and Windows remains their nearly exclusive platform and all turn into a rent-seeking business, Microsoft will continue to make a lot of money.
The corollary to the superstar programmer is that more productive engineers are allowed to take bigger risks in code deployment,allowing them to skip some rote development that a "less productive" engineer may be required to go through.
When it works out, the more productive engineer is allowed to continue risk taking and becomes more productive.
Curative drugs that eliminate disease or produce significant remissions or reductions in illness aren't recurring growth income. You can take them once or very infrequently. Drugs which treat symptoms or only provide short-term gains have to be taken regularly and these drugs have greater per patient consumption and thus greater revenue for drug makers.
Drug makers this have a perverse incentive to merely maintain static illness in patients because it's the most profitable economic model.
Maybe it's their methods -- throw a ton of money at the top 25% of MBA candidates for internships and see what sticks. You gain a database of these people and presumably a fair percentage actually take the internship, so you get to test drive them, too.
Do this for 5 years and you have a pretty good idea of who's worth a damn in senior leadership positions for the next 20 years.
I'm willing to buy the argument that they were more easily exploitable because of their domestic Russian base -- that means vulnerable humans who can turned through the usual apparatus of spycraft and domestic security services, as well as increased general vulnerability because of their geographic location.
That being said, I think any software producer whose products are expected to run at "ring zero" of security should be thought of as vulnerable, regardless of where they are based. I'm sure the intelligence services and security services long ago made the conceptual leap that these were vulnerable targets that would give them direct entry into high value targets due to the nature of their functional security requirements.
I think the chain of trust anymore is pretty much broken and it's not really very paranoid to consider anything secure.
Don't get me wrong, the red man has taken it up the ass. Once the sole proprietor of a mighty hemisphere, he's now hawking gambling, cut rate cigarettes and and has-been rock bands after the white man fucked him over but good.
But this sovereign tribe thing is BS. You want to hold the patent? Fine, then it's only valid on the reservation. Unless the patent violation occurs on tribal land, you're outside your sovereign territory and authority.
If you want to wave your sovereignty around, do something brave. Legalize pot and tell the state legislature to fuck off.
I like aspects of its user interface, but it still can't find places by name for shit. I can put place names into Google Maps and it always finds what I'm looking for. Most of the time Apple Maps doesn't find anything or it finds something totally wrong, far away.
My only suggestion would be that the threshold for posts "disappearing" should be higher -- maybe -20 or something, or some obvious way to restore them for users who want to see them or some kind of threaded sort that ignores votes.
I hear this complaint about Reddit, but I don't really see it much.
The only place I see it is in a sub for my own local city. My city is admittedly very, very liberal, but this sub is overrun with bike advocates and antifa supporters. I've been downvoted into oblivion for having normal sentiments any property owner would have (like objecting to homeless people camping and peeing nearby), but nobody has banned me for it.
Most of the subs I visit, though, have very little political content. Usually you get downvoted just for shitposting or being way off topic, which is mostly fair.
I think Reddit's biggest problem isn't censorship but just high levels of shitposting and relentless memeposting, but mostly this is a "major" subreddit problem, not a niche subreddit problem. It's kind of no different than USENET was.
Slashdot Mirror
Most storage systems I've worked with either strongly advise or outright require drive sizes > 1 TB to use a double parity system like RAID-6 due to the lengthy rebuild times associated with large individual members.
Flash performance is so far ahead of its bulk capacity, though, that capacity can scale without performance appearing to be a factor in restoring redundancy.
and yes, the rebuild times for these will be tremendous and you can't really think about your data as "RAID sets" anymore, that model is quickly becoming outdated. You have to think about "storage nodes" as a single, really large hard drive.
A distinction without a difference. Whether the data is made device redundant by some kind of RAID system or some kind of network copying, drive performance plays in significantly in restoring redundancy after a node failure.
I think the value of these drives is dubious if their redundancy rebuild rate is measured in days due to read/write rates not scaling. They may have niche use cases (ie, containing a complete copy of some large storage quantity on a single medium) but I'd also worry that the density would call into question their durability.
If PV is so economically valuable, why does it need subsidies?
I'm always kind of boggled by sectors that get killed when their subsidies get removed. Sounds to me like the product or sector isn't economically viable if it needs a subsidy to remain competitive.
In terms of "winning" some industrial race to produce PV panels, does it matter? Even if the US ends up developing some next-gen panel design that's a quantum leap forward, it seems unlikely to get manufactured at scale here anyway. The same basic economics that causes it to work out to have China build TVs or refrigerators or everything else at massive scale would seem to wind up working for solar panels, too.
Even if China "owns" the designs, even the Chinese can only absorb so many panels and will likely see the wisdom in building them at scales that feed both domestic and export markets or licensing their technology to other manufacturers.
It's nice to dream about a groovy green energy industry in the US where we design, manufacture and install next-gen PV exclusively at home, but I think modern economics are too sophisticated these days. Capital will seek the lowest cost production for the product that produces the products that buyers identify as providing the utility they want for the money they will spend.
This is an impressive density, but my question is whether drive performance increases and what kind of bus will the drive have?
SATA 3.2 claims 16 GBits/second but my guess from glancing at the other 3.2 features would indicate this is mostly a flash-oriented spec that would be tied to M.2 slots, and that the ordinary SATA slots would still be SATA-3 @ 3 Gbps.
Without a bus and drive combination capable of moving 40 TB in a reasonable amount of time (4-6 hours), these drives will be a novelty and not useful in any storage system without double or triple parity redundancy. The rebuild times without drive speed improvements and a fast bus would make them just time bombs.
I thought there were efficiencies to be had from a self-driving car being able to optimize acceleration, braking, and with car-car communication, greatly reducing a lot of the stop/go in congestion. It should be able to reduce the power consumption of the engine/motors over the way humans accelerate, brake, and speed up and slow down too often.
These won't make up all the losses to self drive, but won't they offset some meaningful percentage? 5% or so power consumption from the drive system through efficient power application should provide a meaningful buffer for the self-drive system.
I also wonder if the self-drive system couldn't partially sleep some part of the sensor environment and processing through some complex decision system involving time of day, real-time congestion reports, road condition, weather, road geography and car-car reporting data. Basically if you're on a long, straight street with no detectable traffic at 3 am in clear weather, couldn't the car dial back the sensor operation to draw less power? Or even if it kept them all running, obtain some efficiency because there's just less change data?
Now that they are starting to re-use rockets and are successfully landing them, have they crossed some magic threshold where their launches are now much cheaper than their competitors using disposable rockets?
Or are they still having to charge a premium due to R&D investments into their system?
If they aren't starting to reap cheaper launch costs, when will they? I would think that while the reusable rockets is an interesting design goal, it would need to cut launch cost meaningfully to be really beneficial.
Could they have been linked with small network of connecting modules into a larger station?
Amazon used to make available data on where a thing you were browsing was popular. I mostly remember books, and they would tell you the regions or areas where it was being sold. "#1 In Minneapolis, #5 in Cambridge" that sort of thing.
If they still had that, you could figure out where sex lube was most popular and grant that area the anal capital of america title.
I kind of agree on corporate cash hoards being out of circulation, although I think they do park the money in some kind of short term securities (like Treasuries or similar highly liquid short-term bonds) so it's doing something.
In the case of Apple specifically, I wonder if part of the reason Apple gets such a good tax deal from Ireland is that there is a quiet agreement that Apple will park a certain portion of it for a longer period in Irish banks, thus enhancing the lending base of Irish banks and ultimately benefiting the Irish economy.
The DOJ should have split Microsoft into Operating Systems and Severs & Applications.
The latter would have been freed to pursue other operating systems as host platforms for their Servers and client Applications.
The former would be able to stop crippling the OS for the server and application platforms as well as encourage other application and server platforms to be hosted on the OS.
Either way, both would be freer to innovate without constantly having to be compromised for their siblings needs.
Beyond that, they could drastically simplify their pricing and versioning schemes. I've often wondered if MS wouldn't actually make more money with a flat rate $100 per license for Windows. The overhead of various licensing schemes and product variations has to eat into profits
But with Apple in particular and some others, isn't it already lesson learned? They're not really innovating with their cash. No (or at least few) failed products introduced, and most of that cash tucked away and guarded by a Leprechaun somewhere in Ireland.
While I agree with you philosophically, I think in terms of an AV program on Windows you're dealing with a unique set of vulnerabilities and a black hat state organization would want to know every detection technique and evasion detection trick they could. It's kind of a fundamentally insecure environment to begin with.
I suspect that we're headed totally into rental, but not entirely driven by greed. I think part of it is driven by rapid software changes and immense code bases which defy traditional hard revisioning -- not technically, but the overhead of providing patches just becomes too great.
I'm also inclined to think that someone is going to figure out how to arbitrage "subscription" software into basically sub-letting subscriptions. Buy enough licenses and enforce micro-rental periods and re-rent the software out to people who want to use it for a day or an hour.
Actually, at those time periods I kind of like subscription software concepts -- I can use what I need for a small cash outlay and not get invested in more software then I really need. Like renting a truck or something, where you just need to do something once.
They don't care. Nobody in charge of anything cares because they are only concerned about next quarter and what comes after.
I actually have a theory that the short-horizon world we live in is some weird byproduct of the baby boomer generation. They're all zeroing in on retirement and are just looking to maximize their retirement income and they literally unconcerned about stuff past about age 80.
I think it will take about another 10 years, when the boomers are nearly all retired or dead, for anything meaningful to change.
...will turn out to be tricking its customer base into renting rather than owning its software. He bought off the fiefdoms by picking winners and turning them into rent-receiving franchises.
As long as few viable alternatives exist for Office and Exchange and Windows remains their nearly exclusive platform and all turn into a rent-seeking business, Microsoft will continue to make a lot of money.
The corollary to the superstar programmer is that more productive engineers are allowed to take bigger risks in code deployment,allowing them to skip some rote development that a "less productive" engineer may be required to go through.
When it works out, the more productive engineer is allowed to continue risk taking and becomes more productive.
I think there's also an economics argument.
Curative drugs that eliminate disease or produce significant remissions or reductions in illness aren't recurring growth income. You can take them once or very infrequently. Drugs which treat symptoms or only provide short-term gains have to be taken regularly and these drugs have greater per patient consumption and thus greater revenue for drug makers.
Drug makers this have a perverse incentive to merely maintain static illness in patients because it's the most profitable economic model.
Maybe it's their methods -- throw a ton of money at the top 25% of MBA candidates for internships and see what sticks. You gain a database of these people and presumably a fair percentage actually take the internship, so you get to test drive them, too.
Do this for 5 years and you have a pretty good idea of who's worth a damn in senior leadership positions for the next 20 years.
I'm willing to buy the argument that they were more easily exploitable because of their domestic Russian base -- that means vulnerable humans who can turned through the usual apparatus of spycraft and domestic security services, as well as increased general vulnerability because of their geographic location.
That being said, I think any software producer whose products are expected to run at "ring zero" of security should be thought of as vulnerable, regardless of where they are based. I'm sure the intelligence services and security services long ago made the conceptual leap that these were vulnerable targets that would give them direct entry into high value targets due to the nature of their functional security requirements.
I think the chain of trust anymore is pretty much broken and it's not really very paranoid to consider anything secure.
Don't get me wrong, the red man has taken it up the ass. Once the sole proprietor of a mighty hemisphere, he's now hawking gambling, cut rate cigarettes and and has-been rock bands after the white man fucked him over but good.
But this sovereign tribe thing is BS. You want to hold the patent? Fine, then it's only valid on the reservation. Unless the patent violation occurs on tribal land, you're outside your sovereign territory and authority.
If you want to wave your sovereignty around, do something brave. Legalize pot and tell the state legislature to fuck off.
I like aspects of its user interface, but it still can't find places by name for shit. I can put place names into Google Maps and it always finds what I'm looking for. Most of the time Apple Maps doesn't find anything or it finds something totally wrong, far away.
My only suggestion would be that the threshold for posts "disappearing" should be higher -- maybe -20 or something, or some obvious way to restore them for users who want to see them or some kind of threaded sort that ignores votes.
I hear this complaint about Reddit, but I don't really see it much.
The only place I see it is in a sub for my own local city. My city is admittedly very, very liberal, but this sub is overrun with bike advocates and antifa supporters. I've been downvoted into oblivion for having normal sentiments any property owner would have (like objecting to homeless people camping and peeing nearby), but nobody has banned me for it.
Most of the subs I visit, though, have very little political content. Usually you get downvoted just for shitposting or being way off topic, which is mostly fair.
I think Reddit's biggest problem isn't censorship but just high levels of shitposting and relentless memeposting, but mostly this is a "major" subreddit problem, not a niche subreddit problem. It's kind of no different than USENET was.
Now you're talking value!!
Won't cheap panels combined with better/cheaper battery storage lead to a trend of just breaking grid ties and running standalone?
I think it's already somewhat practical now if you're willing to accept some limits on max current output, especially at night.