For data centers, it is still all about VGA. The reason is KVM systems use it. Even the big network KVMs are generally VGA at heart. The standard in servers is VGA, with one on the back, and one on the front. The back one for wiring to a KVM, the front on for hooking up to a crash cart.
Now this may not sound like a big deal but when you've managed a number of servers it quickly becomes apparent why such a thing can matter. Can Minis do VGA? Well of course, but as you noted you'll need a host of adapters. Then, of course, if there's an issue you get to wonder if it is the system or the adapter, and so on.
There's a lot done in the server world with regards to standards that is the though for not when things go right but when they go wrong. If you've never experienced the problems then you might not appreciate why but if you have, it becomes rather clear.
That a system hasn't failed doesn't mean it can't. If you take a data point of one and believe that's valid, well then you fail at the science and the statistics.
With anything, computers included, the question is NOT if they'll fail, it is when. That system WILL fail. Might be tomorrow, might be 20 years from now. The problem is you don't know when it'll happen. Then suddenly the TCO goes from the cost of the system, up to whatever lost business there is, maybe it costs your job.
Redundancy, backups, etc are things you have not because you think you'll need them, but because you'll be glad you have them if you do. We have experienced very, very few failures in our critical servers. The hardware is solid. However, we have things set up so that we have excess capacity to deal with it. The reason is not because it happens all the time, but because if it does happen, we want to be ready. Same reason we back up to tape, and rotate them out to another building and shit like that.
We've systems older than 8 years still operational. Our pay to print terminals are like 10 years old or more. They are lab systems for a number of incarnations ago. Reason is it is a very easy thing to do, downtime is no biggie, and not time critical, so we throw whatever old junk there as the station. However that those particular ones are still running doesn't mean they'll run forever or that all systems like them will. We've had a number fail and had to get rid of them.
So sorry, but indeed Mac Minis DO die. Yours will, at some point, and you'd better have a recovery plan, otherwise people may be rather displeased with you.
They do whatever they feel is right, and will change direction totally at the drop of a hat. Now in the consumer market, seems to have worked rather well. They've made a shit ton of money. However it means for enterprise that you need to beware. Support can end at any time, products can go away at nay time, etc. It's one of the reasons I'm very anti-Mac in the enterprise because you can't build a support strategy around it because you don't know what will happen. You can have a solution and then have it disappear.
For a cellphone, no big deal. For a server, it is a real issue.
It is $1000. That's the thing here. They aren't all that powerful and they cost a grand. So you can pack 8 of them in to a 5RU shelf, apparently. Ok, that's $8k, presuming no upgrades... Well go have a look at what you get from Dell for $8k. You can get quite a bit of server, including things like ECC RAM and hot swap disks and all that.
I can understand getting a single cheap computer as a server if your needs are low, and thus you aren't going to spend a ton. But when you are talking about tossing a ton of them in a rack, well you have to evaluate what they'd be competing against.
Then you are doing it wrong. The right answer if you want minimal fucking around time is good hardware, with good support, in a failover cluster. You get some VMWare or Hyper-V servers set up. You have them in a failover setup so that if one goes down, the VMs just hop over to another one. Same deal with scaling. Things getting full? Introduce a new system to the cluster, have it rebalance the VMs. Not only that standing up new systems becomes really quick since you can have images bases you just copy, or if it needs to be done from scratch and ISO file you mount. You can do it all at home, and very quickly.
I'm not talking hypothetically here, this is stuff we do. You virtualize things and have enough hardware capacity that a failure doesn't take shit out.
Also you'll find that the server class hardware is way more reliable. In the Dell servers we have, most things that go wrong often are hot swap. PSUs, HDDs, and fans are all hot swap. If one fails, you get a new one sent out, and swap it in while the system is running. No downtime. RAM and CPUs aren't hot swap but generally it can isolate a bad one, and the system can continue running until a downtime is convenient (we've had a bad RAM stick isolated in a server before). You also have total remote monitoring of the hardware with OpenManage and/or iDRAC. I mean it looks at everything: Power usage, temperature, fan speeds, CPU/memory state, disks, firmware, addon cards, the works. You can monitor it all from wherever you like. Have it fire off alters if something goes wrong, get reports on what is and is not up to date, all that shit.
It is amazing how efficient it can make managing things. There is rarely a need to go fiddle around in the server room. Also, with Windows at least, you can have fully automatic patching, or centrally controlled patching. So you don't have to go system to system ordering patches, they can do it themselves or you can use a central console to clear patches to servers (virtual and physical) as you see fit.
If you want to use Mac Minis, ok no problem, but I think you'd find that if you stepped up your system administration game and got some real server hardware/software, it would take was less time for setup and maintenance. When I want to stand up a new virtual server at work it takes just a couple minutes of time, from wherever I am, to get it started. Then the install is largely automated by images I've configured for Windows, or puppet for Linux. The install is fast too, because it is all going from disks.
For that matter, if I have a server that is a prototype for what I want, I just clone it, and then stand that up. VMs are literally as easy as clicking "clone" to make a copy of.
Dell will sell you a 2U (R820) system that can have 4 processors in it up to 2.7GHz and 8 cores each. It can then take 1.5TB of RAM, 16 2.5" drives (magnetic or SSD) 7 PCIe cards and so on. You'd better believe you can stick more than 40 VMs on that sucker, and you can get another one in 5U with 1U to spare.
Or you can go blade server, Dell has options here though IBM has higher density solutions available, if what you want are a lot of systems in a small amount of space.
All of this supports real enterprise stuff like redundant power, ECC RAM, RAID-6 hot-swap drives, central monitoring, failed component isolation, and so on.
Piling a bunch of consumer computers in a rack doesn't really make a lot of sense, particularly ones not designed with good cooling solutions. When you start doing real high density on computers, cooling is a real issue. Servers are made to deal with it, the vent in the front, out the rear so you have have hot/cool zones and they have high speed fans if they need to spin up due to ambient increases. Mac Minis rely by and large on diffusing heat through their cases and a tiny vent at the back, which is not a winning scenario in a dense situation.
You are going to get better power usage in any large scale by bigger systems with virtualization and having them stand up and down as needed. You can do that with real servers that have full lights out management (Dell calls it iDRAC). As load on the servers rise, new servers can be powered on and made read to the cluster as needed. Also all that high end stuff can buy you realtime failover and migration so things can be shifted around as needed.
All this gets rather feasible with the costs you are talking. 8 Mac Mini servers is, minimum, $8000. That gets you 4GB per system, a 2.3 GHz quad per system, and 2 1TB drives. Because I bought one recently I can tell you that you can get a Dell R720xd will run you about $9000ish for 2x 2.6GHz 8 core CPUs, 128GB of RAM (aftermarket) and 6 1TB HDs (which will let you do RAID-6 with a hot spare and have 3TB of space). I'm not seeing what the minis gain you, and I can give you a list of things they don't have.
For all that, it is a similar power profile. The Macs spec in at about 680 watts total, the Dell has redundant 750 watt PSUs though in my testing only pulled about 600 at max load.
All that in less than half the rack space.
So other than "It can run OS-X" what are you getting with a ton o' minis?
That's the only reason. Apple won't allow for OS-X to be virtualized on non-Mac hardware. vSphere would be perfectly capable of handling it, VMWare has their software on Mac, has Mac integration tools, etc but Apple won't allow it. So if you want OS-X in your datacenter, you have to buy a Mac and since there is no Xserve anymore it is a mini or a pro. Well the pros are really expensive, and quite large (like 4U if you got mounting hardware) so Minis it is.
There really isn't a good reason in most cases, but then fanboys have never needed a good reason. We had a case where people asked for it. A department hired some fairly clueless ex-students that have a "web consulting company" to make a site for them. Said students are Maccies. They wanted a Mac server, running Wordpress to develop on. We said you can have Wordpress (though we tried to talk them out of it) on Apache on Linux because that's what our sites run, we aren't buying a Mac server for you.
"If Dell makes a small little machine, you don't know that they'll be making that, in that form factor, six months down the road, or what they're going to do, or how they're going to refresh it,"
Actually, with Dell you have a pretty good idea. They have defined life cycles for their servers, and they are pretty good about maintaining a general class of equipment. This is not the case for their low end consumer stuff necessarily, but the stuff you'd put in a datacenter.
Apple? Shit son, they'll change tack and tell nobody before hand. The Xserve is the best example. Their 1U server, a thing they sold for use in everything including super-computer like clusters. Then, suddenly it is gone. Just can't buy it anymore, no replacement. You need 1U equipment? Fuck you.
Or the Mac Pro, which is on sale, but they let get woefully out of date before updating.
Apple is the ultimate at doing whatever they want new whenever they want it. They are not at all interested in backward compatibility or consistency. They'll stick with a form as long as it suits them and then change.
Now that's fine, I'm not saying it isn't valid, however to act like they are good at stability for datacenters is silly. They are not at all. The next Mac mini could be a totally different form factor, or there could be NO next Mac mini. You don't know and Apple won't release any roadmap.
Heck a funny mini related incident is one of our professors does research with rovers he builds. They use Mac minis as their core controller because he's a Mac guy. They worked fine since they were small, and powered by DC they could hook up to the power supply for everything else. What's that you say? They aren't DC powered? Ahh yes, well a couple generations ago Apple changed it, stuck the PSU inside the unit. Great for consumers, bad for him. He's now stockpiled some older ones to use when they break and is trying to come up with a long term plan.
To me this reads like a Mac zealot trying to justify their use of them as a good thing rather than a well thought out argument for why they are good in the datacenter.
There is all kinds of extra space in a 2.5" SSD. They have a lil' CPU, some flash chips, and that's it more or less. They are quite small. In smaller form factors, then ya space can become an issue but there's plenty in a 2.5" unit.
If you just stop buying something from a company, and say nothing, well then they really don't know why. They may make an incorrect inference as to what your problem was. For example let's say a game comes out with a new DRM that you hate, and also features a new kind of user input that you like. You don't buy it because of the DRM, but you don't speak up. Same with everyone else, they all like the new input, but hate the DRM, but are silent. The game company looks at the abysmal sales and says "Man that new input idea bombed hard, nobody wants that, let's not to that again."
When you don't like a product you very well should make it known why and not buy it.
However I will say he does have some merit in that gamers, PC gamers in particular, seem to be overly whiny and very tribal. What I mean by that is if you are a "good guy" like Valve, you can do no wrong and if you are a "bad guy" like EA you can do no right.
A great example is the unavailability of some EA games on Steam. The reason is that Valve changed their TOS for new games such that if you have DLC, that DLC must be sold through Steam, not your own site. This didn't used to be how it worked, used to be you could sell a game on Steam (and other services) and sell DLC on your own site. EA wasn't ok with that, they wanted the DLC sales. As such there was an impasse and the EA games that have DLC can't be gotten on Steam, though they can be gotten form other DD services like Impulse, Gamefly, Greenman Gaming, and so on.
Now there's not really a bad guy here, both companies have polices they aren't willing to change, and the policies are understandable, though in both cases you can argue against them. Fair enough. However gamers nearly universally decried EA as being greedy assholes that wouldn't let the noble Valve sell their games.
For that matter they got mad at EA for the same shit Valve does: Tying their games to their DD platform. New EA games wish to use Origin, and will make you log in to it. This is true even if you buy them from another DD service. Ok well this is precisely what Valve does with Steamworks. If you buy any Valve title from HL2 on, you have to install and use Steam. Doesn't matter where you get it, retail, other DD service, you are using Steam period. Same deal with 3rd party titles that use Steamworks (like Skyrim).
However when Valve does it, it is A-OK but when EA does it they are OMGWTFEVIL!!!!111.
So I do understand his point. Gamers need to bitch less, and stop being so tribal. Evaluate stuff on its merits, buy or don't, and don't cry all the time. Also, stick to your guns. An informative happening was when Modern Warfare 2 came out. PC gamers were pissed because it had been gimped on the PC. There was a "Boycott MW2" Steam group. Day one of the release? Most of the people in the group were playing MW2. They were willing to whine, but not to put their money where their mouths were.
See with a buyout offer, there's a lot too it. It isn't like when Google says "We'll give you $6 billion," that is final, that if Groupon says "Ok," Google has to hand over a check and it is done. No, rather it is the first step. Next step is the lawyers get together to draft some NDA type stuff, and then Google gets to go over what happens in the company. They get to look at the financials, the operations, all that shit. They get to have a real good look at what they are going to buy. Only if they are then happy, does it go on to a formal offer and then cash changing hands if that is signed.
What Groupon knew would happen is Google would have a look at their books and realize they weren't making money, and had no plan as to how to. Google would then say "Ya, I'm thinking: no," and back out. That would screw over any IPO because people would ask "Well hang on, what did Google see that was wrong?"
They figured, correctly, that if they just went public they could pull a fast one on people and get some cash. It appears they were right.
Any real currency with fluctuations of this magnitude would be said to be in a crisis and there would be massive stabilization efforts going on. If it was actually a currency (it really isn't, it is a speculation and money laundering tool) it would be experiencing the first ever case of hyper-deflation that I'm aware of.
Anyone who thinks this bodes well for Bitcoin's future doesn't understand money very well. This is NOT good. A good story on Bitcoin would be "For the past two years BTC has remained remarkably stable compared to a basket of important world currencies." If it was functioning as a stable store of value, then maybe there would be some validity to using it as a currency.
As it stands, with the massive the huge deflation and massive volatility (it moves up and down a LOT in a day) it is completely unsuitable as a currency.
So sorry, it was all secret. Maybe you think it shouldn't be, but it is. It is all classified by default since the government understand that the diplomats must be able to talk freely among themselves to effectively do their job.
If you have triggers, then you have mental issues that need to be dealt with. Pushing it aside is NOT healthy. If something traumatic happens to you, there will be mental consequences. You need to work through those. That means counselling, confronting, understanding, etc. If you don't, you'll never fully heal. The mind isn't like the body, you can't just leave an injury be and hope it heals on its own. You have to work on it.
So never mind telling others they have to change, if you ever want to truly recover, you have to deal with the mental scars the trauma has left. Just pushing it aside won't do it. IT is still there, lurking, festering.
Christianity grew up. It's not perfect, it still has plenty of crazies in various kinds, but by and large Christianity grew out of the crusades mentality. Islam by and large has not.
Thus I think it is perfectly reasonable for people to criticize Islam in a way they do not criticize Christianity (I'm an atheist by the way). That Christianity was all "kill the unbelievers" 700+ years ago does NOT give Islam license to be that way today. Society can, and should, advance. I would hope that in 700 years people would look back on society today and be glad that their society was even better than it is now.
In very many ways Islam is still largely stuck in the Dark Ages and it needs to stop. We shouldn't give it a pass because Christianity was also in the Dark Ages several centuries ago.
Yes, or rather the bank can (and must under the law). I've had an ATM eat a check. You just call the bank and have them deal with it.
"Can it run out of bit coins to pay out? and you end up losing cash?"
If an ATM runs out of cash, it refuses to process your transaction. No money is debited from your account. Likewise if its in hopper is full it will refuse to accept your deposit.
"How fast can it update the exchange rate?"
Instantly. The ATM itself doesn't store any of that, the bank does. It just communicates with the bank. So it is always current to the rate the bank offers you. Also, this is not an issue with normal ATM operations since you access the same currency your account is denominated in. This only applies internationally, and then the rates shift slowly. BTC shift multiple dollars in a day.
"What happens with a network lost midway though an transfer?"
The transaction doesn't happen. The ATM only does things if the bank says it is ok. So if it loses connectivity before it is complete, the transaction is stopped. You either don't get your money in the case of a withdrawal, or it hands you back your deposit in that case.
"Can it store and transfer at a later time?"
No, ATMs are dumb terminals, after a fashion. While they run all their interface locally, they don't have any account data. They just contact the bank and say "This account number wants to do this, is this ok?" The bank then says yes or no, the transaction happens, everything is updated on the bank's computers.
"What about an power loss?"
Same deal as anything else, you call the bank. They'll have to come and get your card out, also if you had a deposit in the hopper but not processed they'll have to get that out too.
"Hacking? and will be even be crime to hack it? and what will happen in a court if some one sues or they try to change some with breaking a law dealing with any to do with this?"
It is a crime to hack an ATM though it is nearly impossible, since again they are dumb terminals. The crypto between them and the bank is top notch (IBM makes these real cool crypto cards for them). In the event the ATM is attacked and the money stolen, it is an issue for the bank, not you. The risk to the end user is skimming, someone capturing your account information and then using it, same basic deal as credit card fraud and the like.
That the same questions apply doesn't mean the same answers do. Really, there has been a lot of time to think about and work on answers with ATMs. The big reason they work well is that they are just terminals for the bank. They don't store anything, other than the physical cash they dispense. They just transact back to the bank. Also, there's rather a lot of tracking that goes on with regards to bank transactions at many levels. If something happens, there's logging, there's a record, it can almost always be undone.
I understand that money is just a theoretical construct, something that facilitates trade. Trade, people doing things, making things, that is what the economy actually is. Money is just what we use to make that efficient by allowing an arbitrary level of indirection in both time and space for trades. You don't have to barter anymore, you accept money for what you do, and so does everyone else.
This is also why it doesn't matter what money is, it can be anything, gold, salt, teeth, rocks, paper, bits in a computer, whatever (all these have actually been used). It just has to be something everyone agrees on to use.
However that then implies something: Money is only money if you can spend it. If very few people, or nobody, will take your "money" then it isn't. Doesn't matter what claims you make of it, doesn't matter what its value is supposed to be. If you can't spend it, if people won't take it in trade for goods and services, it isn't money.
There's another side to that though: Money is only money if people DO spend it. Since the whole purpose is to facilitate trade, to act as grease for the economy if you like, it only works if people do actually go out and spend it. If people just hoard it, then it is useless. Doesn't matter if it is something that everyone would accept as payment, if it isn't used as payment, if nobody spends it, then it doesn't act as money.
Well now this starts to bring us back to the problem with deflation. It encourages hoarding, the higher the deflation the worse it is. It makes borrowing and lending hard to impossible. It freezes things up because it makes the money stop acting like money. People want to just hold as much as possible for as long as possible.
Now while it might simplistically sound like that's a good thing, more savings right? It isn't when you evaluate it. Again, the only thing real in the economy is trade, people actually doing things. All this higher level theoretical stuff is just to make that work. So if people don't spend money, it means that people are sitting idle, not doing work, it means that the economy does poorly.
Money is not some magical force or substance. It is just a convenient theoretical construct to solve the problems of a barter system. Money that does that well is useful. Something that does not doesn't work as money.
What I find amusing about your carriage-return filled rant is that you love the idea of no inflation it would seem but you seem to then be ok with deflation. No inflation is a good goal, Bitcoins cannot have that by design. They will have deflation, period, because there are a fixed number and the economy is not a fixed size. You either misunderstand how deflation works, how Bitcoin works, or you are smoke screening and really do think deflation is a good thing.
The amusing thing is if you are big on the idea of zero inflation, then fiat currency is currently the best answer. It can have zero inflation. It doesn't, in any actual implementation I am aware of, but it can. The controlling group could carefully balance the supply to make sure that it maps the changes in the overall economy so that there is no inflation. However something with a limited supply? Nope, it'll deflate so long as the economy grows. The economy will get larger, the currency supply will remain fixed, deflation is the inevitable result.
So I reiterate: Go take ECON 200. Stop with the Smartest Motherfucker in the Universe Syndrome(tm) that so many geeks seem to have, thinking you know all the answers. Go learn about some basics of how the economy actually works.
I do IT work and we are all in cubes in a large room, including my boss (he's a tech, not a PHB). It really works better because when someone wanders in needing help, they can more quickly get routed to the person who can actually help them, when there's questions about something we can get them answered quick, and we can chat about ideas.
I find I really like it. It isn't perfect, of course, but overall I'd take it over us all being in individual offices, which we could have, if we wanted (most of us do have an assigned office to use if we need, we just don't).
It will never work as a currency because it has built in deflation, which if you've taken ECON 200 you'll know is a really, really bad thing for an economy.
Nor is it being used as a currency right now. A currency is something people hold, spend, get paid in, etc. Bitcoin is basically used only for three things:
1) Money laundering. Sites that do illegal things, like the Silk Road, use Bitcoin to launder money. They take payment in it but immediately convert that in to an actual currency. They are just using the payment system to launder the money from the client to them.
2) Speculation. Traders play the Bitcoin market to try and make a quick buck. Hence one of the reasons for the extreme volatility in price. If any actual currency had that kind of daily volatility it would be said to be in crisis, yet somehow the BTCtards want to act like it is perfectly ok for Bitcoins to fluctuate like that. It also, of course, makes it even less suitable as a currency for people to hang on to.
3) "Mining." People literally wasting CPU cycles and electricity to generate new Bitcoins. Most because they are bad at math and don't count the total cost of all their stuff and realize that they aren't actually making any money on it.
The inherent deflationary property of it makes it a failure as a currency out of the gate. There may be many other problems it would face on a large scale of usage, nobody has yet to convince me it has a good solution to timing attacks, but it doesn't matter because it fails as a currency.
The people who think deflation is good are people with no understanding of economics past their wallet. They think "Deflation means the money in my wallet is worth more so it is a good thing!" That is not the case, you have to look at the larger economic consequences, and then you discover deflation is very, very bad.
But there's this whole other class of cameras, video cameras, that write a continuous stream to the SD cards. AVCHD cameras are SD based. There your options are to get a card that's fast enough or to have to turn down your detail (or simply not record at all).
Australia is a large country with a lot of nothing. There are like 4 city centres where most people live, a bit of rural population, and then a whole bunch of empty nothing. Not surprising, given the climate and geography really.
The US is probably one of the most rural nations over all. Lots of big cities too, of course, but a substantial amount of population that is spread out over a substantial amount of land.
Then you may need more speed. Your N gets you more like 100mbps effective data rate (test it some time) since the WiFi speeds are displayed raw and there's a lot of overhead. Now that is 100mbps shared among all devices. So, if you connect to your router and it to a wired computer, no problem full bandwidth. However if you connect to another computer on WiFi, oh look, you guys are sharing. Have a bunch of computers on all accessing, that bandwidth starts to get spread thin.
If all you do is one computer to the Internet, then you are fine, for now at least. Otherwise? Yes, more bandwidth is good.
The original article was insightful, it enlightened us to a new evolution taking place. This is just a snarky restatement without any added, and in fact far less, information.
The article straight off says mosquitoes are evolving, and talks about the research as to in the mechanism that is changing.
It seems like almost all dogs can get really protective when they feel the need, even when you wouldn't think so.
We had a really doofy standard poodle. She was an idiot, even on the demeaning scale of dog intelligence. If dogs rode buses, she'd ride the short one. The most loving dog you ever saw. Just wanted to be petted and cuddled and lay on your lap (despite being 90 pounds, she thought she was a lap dog) all day, every day. The least fearsome beast I'd ever met.
Then one day I'm taking the dogs on a walk. A friend of mine is biking by and decides to try and scare me, so he yells as he rides by. The poodle goes in to attach mode. Fangs bared, loud, menacing barks, back down on her legs ready to lunge. She then recognized him and turned in tot eh friendly teddy bear she usually is.
I never thought she had it in her, but she was ready to kill.
I had a roommate who was a union plumber and I really couldn't believe how fucking bad the union was. It really didn't do shit for him. The real telling thing was during the recession. I mean what someone does for you in times of plenty isn't nearly as telling as in times of need. Well basically he sat at home, idle, because they had no work. He got a "raise" during that period because he'd reached the next tier in their ranking system, but there was no work so it didn't do shit.
They had no unemployment assistance or anything, all their training was paid-for education through the University of Phoenix. In fact, he couldn't collect government unemployment insurance because he was still technically employed. They would release him at his request (which he eventually did) and he'd then be eligible but they warned him he'd then be at the back of the line to get a job back when work came back.
They still wanted their union dues though, those were not waived despite lack of work. Oh, and they told him if he left, he couldn't take a non-union job for 2 years. Such preventions aren't legal in this state, but they told him anyhow.
I really can't see anything useful they did for him. I mean I guess on paper he had reasonably high pay per hour, though I don't have a scale of other plumbers to compare it to, but that doesn't do any good if hours worked = 0. Their concern seemed to be to protect the union and collect dues, not to look out for their employees.
Slashdot Mirror
For data centers, it is still all about VGA. The reason is KVM systems use it. Even the big network KVMs are generally VGA at heart. The standard in servers is VGA, with one on the back, and one on the front. The back one for wiring to a KVM, the front on for hooking up to a crash cart.
Now this may not sound like a big deal but when you've managed a number of servers it quickly becomes apparent why such a thing can matter. Can Minis do VGA? Well of course, but as you noted you'll need a host of adapters. Then, of course, if there's an issue you get to wonder if it is the system or the adapter, and so on.
There's a lot done in the server world with regards to standards that is the though for not when things go right but when they go wrong. If you've never experienced the problems then you might not appreciate why but if you have, it becomes rather clear.
That a system hasn't failed doesn't mean it can't. If you take a data point of one and believe that's valid, well then you fail at the science and the statistics.
With anything, computers included, the question is NOT if they'll fail, it is when. That system WILL fail. Might be tomorrow, might be 20 years from now. The problem is you don't know when it'll happen. Then suddenly the TCO goes from the cost of the system, up to whatever lost business there is, maybe it costs your job.
Redundancy, backups, etc are things you have not because you think you'll need them, but because you'll be glad you have them if you do. We have experienced very, very few failures in our critical servers. The hardware is solid. However, we have things set up so that we have excess capacity to deal with it. The reason is not because it happens all the time, but because if it does happen, we want to be ready. Same reason we back up to tape, and rotate them out to another building and shit like that.
We've systems older than 8 years still operational. Our pay to print terminals are like 10 years old or more. They are lab systems for a number of incarnations ago. Reason is it is a very easy thing to do, downtime is no biggie, and not time critical, so we throw whatever old junk there as the station. However that those particular ones are still running doesn't mean they'll run forever or that all systems like them will. We've had a number fail and had to get rid of them.
So sorry, but indeed Mac Minis DO die. Yours will, at some point, and you'd better have a recovery plan, otherwise people may be rather displeased with you.
They do whatever they feel is right, and will change direction totally at the drop of a hat. Now in the consumer market, seems to have worked rather well. They've made a shit ton of money. However it means for enterprise that you need to beware. Support can end at any time, products can go away at nay time, etc. It's one of the reasons I'm very anti-Mac in the enterprise because you can't build a support strategy around it because you don't know what will happen. You can have a solution and then have it disappear.
For a cellphone, no big deal. For a server, it is a real issue.
It is $1000. That's the thing here. They aren't all that powerful and they cost a grand. So you can pack 8 of them in to a 5RU shelf, apparently. Ok, that's $8k, presuming no upgrades... Well go have a look at what you get from Dell for $8k. You can get quite a bit of server, including things like ECC RAM and hot swap disks and all that.
I can understand getting a single cheap computer as a server if your needs are low, and thus you aren't going to spend a ton. But when you are talking about tossing a ton of them in a rack, well you have to evaluate what they'd be competing against.
Then you are doing it wrong. The right answer if you want minimal fucking around time is good hardware, with good support, in a failover cluster. You get some VMWare or Hyper-V servers set up. You have them in a failover setup so that if one goes down, the VMs just hop over to another one. Same deal with scaling. Things getting full? Introduce a new system to the cluster, have it rebalance the VMs. Not only that standing up new systems becomes really quick since you can have images bases you just copy, or if it needs to be done from scratch and ISO file you mount. You can do it all at home, and very quickly.
I'm not talking hypothetically here, this is stuff we do. You virtualize things and have enough hardware capacity that a failure doesn't take shit out.
Also you'll find that the server class hardware is way more reliable. In the Dell servers we have, most things that go wrong often are hot swap. PSUs, HDDs, and fans are all hot swap. If one fails, you get a new one sent out, and swap it in while the system is running. No downtime. RAM and CPUs aren't hot swap but generally it can isolate a bad one, and the system can continue running until a downtime is convenient (we've had a bad RAM stick isolated in a server before). You also have total remote monitoring of the hardware with OpenManage and/or iDRAC. I mean it looks at everything: Power usage, temperature, fan speeds, CPU/memory state, disks, firmware, addon cards, the works. You can monitor it all from wherever you like. Have it fire off alters if something goes wrong, get reports on what is and is not up to date, all that shit.
It is amazing how efficient it can make managing things. There is rarely a need to go fiddle around in the server room. Also, with Windows at least, you can have fully automatic patching, or centrally controlled patching. So you don't have to go system to system ordering patches, they can do it themselves or you can use a central console to clear patches to servers (virtual and physical) as you see fit.
If you want to use Mac Minis, ok no problem, but I think you'd find that if you stepped up your system administration game and got some real server hardware/software, it would take was less time for setup and maintenance. When I want to stand up a new virtual server at work it takes just a couple minutes of time, from wherever I am, to get it started. Then the install is largely automated by images I've configured for Windows, or puppet for Linux. The install is fast too, because it is all going from disks.
For that matter, if I have a server that is a prototype for what I want, I just clone it, and then stand that up. VMs are literally as easy as clicking "clone" to make a copy of.
Dell will sell you a 2U (R820) system that can have 4 processors in it up to 2.7GHz and 8 cores each. It can then take 1.5TB of RAM, 16 2.5" drives (magnetic or SSD) 7 PCIe cards and so on. You'd better believe you can stick more than 40 VMs on that sucker, and you can get another one in 5U with 1U to spare.
Or you can go blade server, Dell has options here though IBM has higher density solutions available, if what you want are a lot of systems in a small amount of space.
All of this supports real enterprise stuff like redundant power, ECC RAM, RAID-6 hot-swap drives, central monitoring, failed component isolation, and so on.
Piling a bunch of consumer computers in a rack doesn't really make a lot of sense, particularly ones not designed with good cooling solutions. When you start doing real high density on computers, cooling is a real issue. Servers are made to deal with it, the vent in the front, out the rear so you have have hot/cool zones and they have high speed fans if they need to spin up due to ambient increases. Mac Minis rely by and large on diffusing heat through their cases and a tiny vent at the back, which is not a winning scenario in a dense situation.
You are going to get better power usage in any large scale by bigger systems with virtualization and having them stand up and down as needed. You can do that with real servers that have full lights out management (Dell calls it iDRAC). As load on the servers rise, new servers can be powered on and made read to the cluster as needed. Also all that high end stuff can buy you realtime failover and migration so things can be shifted around as needed.
All this gets rather feasible with the costs you are talking. 8 Mac Mini servers is, minimum, $8000. That gets you 4GB per system, a 2.3 GHz quad per system, and 2 1TB drives. Because I bought one recently I can tell you that you can get a Dell R720xd will run you about $9000ish for 2x 2.6GHz 8 core CPUs, 128GB of RAM (aftermarket) and 6 1TB HDs (which will let you do RAID-6 with a hot spare and have 3TB of space). I'm not seeing what the minis gain you, and I can give you a list of things they don't have.
For all that, it is a similar power profile. The Macs spec in at about 680 watts total, the Dell has redundant 750 watt PSUs though in my testing only pulled about 600 at max load.
All that in less than half the rack space.
So other than "It can run OS-X" what are you getting with a ton o' minis?
That's the only reason. Apple won't allow for OS-X to be virtualized on non-Mac hardware. vSphere would be perfectly capable of handling it, VMWare has their software on Mac, has Mac integration tools, etc but Apple won't allow it. So if you want OS-X in your datacenter, you have to buy a Mac and since there is no Xserve anymore it is a mini or a pro. Well the pros are really expensive, and quite large (like 4U if you got mounting hardware) so Minis it is.
There really isn't a good reason in most cases, but then fanboys have never needed a good reason. We had a case where people asked for it. A department hired some fairly clueless ex-students that have a "web consulting company" to make a site for them. Said students are Maccies. They wanted a Mac server, running Wordpress to develop on. We said you can have Wordpress (though we tried to talk them out of it) on Apache on Linux because that's what our sites run, we aren't buying a Mac server for you.
"If Dell makes a small little machine, you don't know that they'll be making that, in that form factor, six months down the road, or what they're going to do, or how they're going to refresh it,"
Actually, with Dell you have a pretty good idea. They have defined life cycles for their servers, and they are pretty good about maintaining a general class of equipment. This is not the case for their low end consumer stuff necessarily, but the stuff you'd put in a datacenter.
Apple? Shit son, they'll change tack and tell nobody before hand. The Xserve is the best example. Their 1U server, a thing they sold for use in everything including super-computer like clusters. Then, suddenly it is gone. Just can't buy it anymore, no replacement. You need 1U equipment? Fuck you.
Or the Mac Pro, which is on sale, but they let get woefully out of date before updating.
Apple is the ultimate at doing whatever they want new whenever they want it. They are not at all interested in backward compatibility or consistency. They'll stick with a form as long as it suits them and then change.
Now that's fine, I'm not saying it isn't valid, however to act like they are good at stability for datacenters is silly. They are not at all. The next Mac mini could be a totally different form factor, or there could be NO next Mac mini. You don't know and Apple won't release any roadmap.
Heck a funny mini related incident is one of our professors does research with rovers he builds. They use Mac minis as their core controller because he's a Mac guy. They worked fine since they were small, and powered by DC they could hook up to the power supply for everything else. What's that you say? They aren't DC powered? Ahh yes, well a couple generations ago Apple changed it, stuck the PSU inside the unit. Great for consumers, bad for him. He's now stockpiled some older ones to use when they break and is trying to come up with a long term plan.
To me this reads like a Mac zealot trying to justify their use of them as a good thing rather than a well thought out argument for why they are good in the datacenter.
There is all kinds of extra space in a 2.5" SSD. They have a lil' CPU, some flash chips, and that's it more or less. They are quite small. In smaller form factors, then ya space can become an issue but there's plenty in a 2.5" unit.
If you just stop buying something from a company, and say nothing, well then they really don't know why. They may make an incorrect inference as to what your problem was. For example let's say a game comes out with a new DRM that you hate, and also features a new kind of user input that you like. You don't buy it because of the DRM, but you don't speak up. Same with everyone else, they all like the new input, but hate the DRM, but are silent. The game company looks at the abysmal sales and says "Man that new input idea bombed hard, nobody wants that, let's not to that again."
When you don't like a product you very well should make it known why and not buy it.
However I will say he does have some merit in that gamers, PC gamers in particular, seem to be overly whiny and very tribal. What I mean by that is if you are a "good guy" like Valve, you can do no wrong and if you are a "bad guy" like EA you can do no right.
A great example is the unavailability of some EA games on Steam. The reason is that Valve changed their TOS for new games such that if you have DLC, that DLC must be sold through Steam, not your own site. This didn't used to be how it worked, used to be you could sell a game on Steam (and other services) and sell DLC on your own site. EA wasn't ok with that, they wanted the DLC sales. As such there was an impasse and the EA games that have DLC can't be gotten on Steam, though they can be gotten form other DD services like Impulse, Gamefly, Greenman Gaming, and so on.
Now there's not really a bad guy here, both companies have polices they aren't willing to change, and the policies are understandable, though in both cases you can argue against them. Fair enough. However gamers nearly universally decried EA as being greedy assholes that wouldn't let the noble Valve sell their games.
For that matter they got mad at EA for the same shit Valve does: Tying their games to their DD platform. New EA games wish to use Origin, and will make you log in to it. This is true even if you buy them from another DD service. Ok well this is precisely what Valve does with Steamworks. If you buy any Valve title from HL2 on, you have to install and use Steam. Doesn't matter where you get it, retail, other DD service, you are using Steam period. Same deal with 3rd party titles that use Steamworks (like Skyrim).
However when Valve does it, it is A-OK but when EA does it they are OMGWTFEVIL!!!!111.
So I do understand his point. Gamers need to bitch less, and stop being so tribal. Evaluate stuff on its merits, buy or don't, and don't cry all the time. Also, stick to your guns. An informative happening was when Modern Warfare 2 came out. PC gamers were pissed because it had been gimped on the PC. There was a "Boycott MW2" Steam group. Day one of the release? Most of the people in the group were playing MW2. They were willing to whine, but not to put their money where their mouths were.
See with a buyout offer, there's a lot too it. It isn't like when Google says "We'll give you $6 billion," that is final, that if Groupon says "Ok," Google has to hand over a check and it is done. No, rather it is the first step. Next step is the lawyers get together to draft some NDA type stuff, and then Google gets to go over what happens in the company. They get to look at the financials, the operations, all that shit. They get to have a real good look at what they are going to buy. Only if they are then happy, does it go on to a formal offer and then cash changing hands if that is signed.
What Groupon knew would happen is Google would have a look at their books and realize they weren't making money, and had no plan as to how to. Google would then say "Ya, I'm thinking: no," and back out. That would screw over any IPO because people would ask "Well hang on, what did Google see that was wrong?"
They figured, correctly, that if they just went public they could pull a fast one on people and get some cash. It appears they were right.
Any real currency with fluctuations of this magnitude would be said to be in a crisis and there would be massive stabilization efforts going on. If it was actually a currency (it really isn't, it is a speculation and money laundering tool) it would be experiencing the first ever case of hyper-deflation that I'm aware of.
Anyone who thinks this bodes well for Bitcoin's future doesn't understand money very well. This is NOT good. A good story on Bitcoin would be "For the past two years BTC has remained remarkably stable compared to a basket of important world currencies." If it was functioning as a stable store of value, then maybe there would be some validity to using it as a currency.
As it stands, with the massive the huge deflation and massive volatility (it moves up and down a LOT in a day) it is completely unsuitable as a currency.
So sorry, it was all secret. Maybe you think it shouldn't be, but it is. It is all classified by default since the government understand that the diplomats must be able to talk freely among themselves to effectively do their job.
If you have triggers, then you have mental issues that need to be dealt with. Pushing it aside is NOT healthy. If something traumatic happens to you, there will be mental consequences. You need to work through those. That means counselling, confronting, understanding, etc. If you don't, you'll never fully heal. The mind isn't like the body, you can't just leave an injury be and hope it heals on its own. You have to work on it.
So never mind telling others they have to change, if you ever want to truly recover, you have to deal with the mental scars the trauma has left. Just pushing it aside won't do it. IT is still there, lurking, festering.
But ya
Christianity grew up. It's not perfect, it still has plenty of crazies in various kinds, but by and large Christianity grew out of the crusades mentality. Islam by and large has not.
Thus I think it is perfectly reasonable for people to criticize Islam in a way they do not criticize Christianity (I'm an atheist by the way). That Christianity was all "kill the unbelievers" 700+ years ago does NOT give Islam license to be that way today. Society can, and should, advance. I would hope that in 700 years people would look back on society today and be glad that their society was even better than it is now.
In very many ways Islam is still largely stuck in the Dark Ages and it needs to stop. We shouldn't give it a pass because Christianity was also in the Dark Ages several centuries ago.
"Can it refund your cash if there is any fault?"
Yes, or rather the bank can (and must under the law). I've had an ATM eat a check. You just call the bank and have them deal with it.
"Can it run out of bit coins to pay out? and you end up losing cash?"
If an ATM runs out of cash, it refuses to process your transaction. No money is debited from your account. Likewise if its in hopper is full it will refuse to accept your deposit.
"How fast can it update the exchange rate?"
Instantly. The ATM itself doesn't store any of that, the bank does. It just communicates with the bank. So it is always current to the rate the bank offers you. Also, this is not an issue with normal ATM operations since you access the same currency your account is denominated in. This only applies internationally, and then the rates shift slowly. BTC shift multiple dollars in a day.
"What happens with a network lost midway though an transfer?"
The transaction doesn't happen. The ATM only does things if the bank says it is ok. So if it loses connectivity before it is complete, the transaction is stopped. You either don't get your money in the case of a withdrawal, or it hands you back your deposit in that case.
"Can it store and transfer at a later time?"
No, ATMs are dumb terminals, after a fashion. While they run all their interface locally, they don't have any account data. They just contact the bank and say "This account number wants to do this, is this ok?" The bank then says yes or no, the transaction happens, everything is updated on the bank's computers.
"What about an power loss?"
Same deal as anything else, you call the bank. They'll have to come and get your card out, also if you had a deposit in the hopper but not processed they'll have to get that out too.
"Hacking? and will be even be crime to hack it? and what will happen in a court if some one sues or they try to change some with breaking a law dealing with any to do with this?"
It is a crime to hack an ATM though it is nearly impossible, since again they are dumb terminals. The crypto between them and the bank is top notch (IBM makes these real cool crypto cards for them). In the event the ATM is attacked and the money stolen, it is an issue for the bank, not you. The risk to the end user is skimming, someone capturing your account information and then using it, same basic deal as credit card fraud and the like.
That the same questions apply doesn't mean the same answers do. Really, there has been a lot of time to think about and work on answers with ATMs. The big reason they work well is that they are just terminals for the bank. They don't store anything, other than the physical cash they dispense. They just transact back to the bank. Also, there's rather a lot of tracking that goes on with regards to bank transactions at many levels. If something happens, there's logging, there's a record, it can almost always be undone.
I understand that money is just a theoretical construct, something that facilitates trade. Trade, people doing things, making things, that is what the economy actually is. Money is just what we use to make that efficient by allowing an arbitrary level of indirection in both time and space for trades. You don't have to barter anymore, you accept money for what you do, and so does everyone else.
This is also why it doesn't matter what money is, it can be anything, gold, salt, teeth, rocks, paper, bits in a computer, whatever (all these have actually been used). It just has to be something everyone agrees on to use.
However that then implies something: Money is only money if you can spend it. If very few people, or nobody, will take your "money" then it isn't. Doesn't matter what claims you make of it, doesn't matter what its value is supposed to be. If you can't spend it, if people won't take it in trade for goods and services, it isn't money.
There's another side to that though: Money is only money if people DO spend it. Since the whole purpose is to facilitate trade, to act as grease for the economy if you like, it only works if people do actually go out and spend it. If people just hoard it, then it is useless. Doesn't matter if it is something that everyone would accept as payment, if it isn't used as payment, if nobody spends it, then it doesn't act as money.
Well now this starts to bring us back to the problem with deflation. It encourages hoarding, the higher the deflation the worse it is. It makes borrowing and lending hard to impossible. It freezes things up because it makes the money stop acting like money. People want to just hold as much as possible for as long as possible.
Now while it might simplistically sound like that's a good thing, more savings right? It isn't when you evaluate it. Again, the only thing real in the economy is trade, people actually doing things. All this higher level theoretical stuff is just to make that work. So if people don't spend money, it means that people are sitting idle, not doing work, it means that the economy does poorly.
Money is not some magical force or substance. It is just a convenient theoretical construct to solve the problems of a barter system. Money that does that well is useful. Something that does not doesn't work as money.
What I find amusing about your carriage-return filled rant is that you love the idea of no inflation it would seem but you seem to then be ok with deflation. No inflation is a good goal, Bitcoins cannot have that by design. They will have deflation, period, because there are a fixed number and the economy is not a fixed size. You either misunderstand how deflation works, how Bitcoin works, or you are smoke screening and really do think deflation is a good thing.
The amusing thing is if you are big on the idea of zero inflation, then fiat currency is currently the best answer. It can have zero inflation. It doesn't, in any actual implementation I am aware of, but it can. The controlling group could carefully balance the supply to make sure that it maps the changes in the overall economy so that there is no inflation. However something with a limited supply? Nope, it'll deflate so long as the economy grows. The economy will get larger, the currency supply will remain fixed, deflation is the inevitable result.
So I reiterate: Go take ECON 200. Stop with the Smartest Motherfucker in the Universe Syndrome(tm) that so many geeks seem to have, thinking you know all the answers. Go learn about some basics of how the economy actually works.
I do IT work and we are all in cubes in a large room, including my boss (he's a tech, not a PHB). It really works better because when someone wanders in needing help, they can more quickly get routed to the person who can actually help them, when there's questions about something we can get them answered quick, and we can chat about ideas.
I find I really like it. It isn't perfect, of course, but overall I'd take it over us all being in individual offices, which we could have, if we wanted (most of us do have an assigned office to use if we need, we just don't).
It will never work as a currency because it has built in deflation, which if you've taken ECON 200 you'll know is a really, really bad thing for an economy.
Nor is it being used as a currency right now. A currency is something people hold, spend, get paid in, etc. Bitcoin is basically used only for three things:
1) Money laundering. Sites that do illegal things, like the Silk Road, use Bitcoin to launder money. They take payment in it but immediately convert that in to an actual currency. They are just using the payment system to launder the money from the client to them.
2) Speculation. Traders play the Bitcoin market to try and make a quick buck. Hence one of the reasons for the extreme volatility in price. If any actual currency had that kind of daily volatility it would be said to be in crisis, yet somehow the BTCtards want to act like it is perfectly ok for Bitcoins to fluctuate like that. It also, of course, makes it even less suitable as a currency for people to hang on to.
3) "Mining." People literally wasting CPU cycles and electricity to generate new Bitcoins. Most because they are bad at math and don't count the total cost of all their stuff and realize that they aren't actually making any money on it.
The inherent deflationary property of it makes it a failure as a currency out of the gate. There may be many other problems it would face on a large scale of usage, nobody has yet to convince me it has a good solution to timing attacks, but it doesn't matter because it fails as a currency.
The people who think deflation is good are people with no understanding of economics past their wallet. They think "Deflation means the money in my wallet is worth more so it is a good thing!" That is not the case, you have to look at the larger economic consequences, and then you discover deflation is very, very bad.
But there's this whole other class of cameras, video cameras, that write a continuous stream to the SD cards. AVCHD cameras are SD based. There your options are to get a card that's fast enough or to have to turn down your detail (or simply not record at all).
Australia is a large country with a lot of nothing. There are like 4 city centres where most people live, a bit of rural population, and then a whole bunch of empty nothing. Not surprising, given the climate and geography really.
The US is probably one of the most rural nations over all. Lots of big cities too, of course, but a substantial amount of population that is spread out over a substantial amount of land.
Then you may need more speed. Your N gets you more like 100mbps effective data rate (test it some time) since the WiFi speeds are displayed raw and there's a lot of overhead. Now that is 100mbps shared among all devices. So, if you connect to your router and it to a wired computer, no problem full bandwidth. However if you connect to another computer on WiFi, oh look, you guys are sharing. Have a bunch of computers on all accessing, that bandwidth starts to get spread thin.
If all you do is one computer to the Internet, then you are fine, for now at least. Otherwise? Yes, more bandwidth is good.
The original article was insightful, it enlightened us to a new evolution taking place. This is just a snarky restatement without any added, and in fact far less, information.
The article straight off says mosquitoes are evolving, and talks about the research as to in the mechanism that is changing.
It seems like almost all dogs can get really protective when they feel the need, even when you wouldn't think so.
We had a really doofy standard poodle. She was an idiot, even on the demeaning scale of dog intelligence. If dogs rode buses, she'd ride the short one. The most loving dog you ever saw. Just wanted to be petted and cuddled and lay on your lap (despite being 90 pounds, she thought she was a lap dog) all day, every day. The least fearsome beast I'd ever met.
Then one day I'm taking the dogs on a walk. A friend of mine is biking by and decides to try and scare me, so he yells as he rides by. The poodle goes in to attach mode. Fangs bared, loud, menacing barks, back down on her legs ready to lunge. She then recognized him and turned in tot eh friendly teddy bear she usually is.
I never thought she had it in her, but she was ready to kill.
I had a roommate who was a union plumber and I really couldn't believe how fucking bad the union was. It really didn't do shit for him. The real telling thing was during the recession. I mean what someone does for you in times of plenty isn't nearly as telling as in times of need. Well basically he sat at home, idle, because they had no work. He got a "raise" during that period because he'd reached the next tier in their ranking system, but there was no work so it didn't do shit.
They had no unemployment assistance or anything, all their training was paid-for education through the University of Phoenix. In fact, he couldn't collect government unemployment insurance because he was still technically employed. They would release him at his request (which he eventually did) and he'd then be eligible but they warned him he'd then be at the back of the line to get a job back when work came back.
They still wanted their union dues though, those were not waived despite lack of work. Oh, and they told him if he left, he couldn't take a non-union job for 2 years. Such preventions aren't legal in this state, but they told him anyhow.
I really can't see anything useful they did for him. I mean I guess on paper he had reasonably high pay per hour, though I don't have a scale of other plumbers to compare it to, but that doesn't do any good if hours worked = 0. Their concern seemed to be to protect the union and collect dues, not to look out for their employees.