A lot of people are telling you that you're wrong, that you need "offsite backup", etc, etc. But that's not what you asked. You've got an external Hard drive already it sounds like. Most banks will rent a safe deposit box for fairly cheap. This will give you that "Fireproof" that you need. You're either going to need a computer with USB 3.0 or external hard drive(s) with ethernet capability.
Didn't you just describe an "offsite backup"? Offsite backup doesn't necessarily mean data backed up live to the cloud. My company's offsite backup strategy is to have tapes picked up by Iron Mountain every week. Weekly backups are stored at a "local" location (about 40 miles from our building), quarterly full dumps are sent to a different facility out of state.
People are telling him he needs "offsite backup" because there's no way to keep his data safe if it's all on-site. Most people aren't willing to drive their data far enough off-site to keep it safe. Putting it your local bank's safe deposit box is only safe until a regional disaster hits both your house and the bank... Iike an earthquake or flood.
There is no "safe" on-site storage option. My brother had an expensive expensive fire-proof gun safe at his house - it weighed over 500 lbs.... After his house literally burnt to the ground (rural area, no one home to alert the first department, by the time the neighbor down the road saw the fire and called the fire department, all they could do was put out the fire in the surrounding brush (and cars)), he couldn't even identify any remaining pieces of the safe or the guns that were in it in the debris.
If you want your data safe, store it off site, preferably in another part of the country so it's not subject to the same local disasters as your house. Mailing snapshots of your data to an out of state relative is probably best.
Because +@domain.com is not easy to figure out? I'd be very surprised if a spammer that acquires your garbage email address can't obtain your actual email address.
What is my "actual" email address? I have around a dozen addresses that are important because I've given to family and friends, (and literally hundreds that I've used to sign up for various newsletters, blog sites, etc). If you just drop the part after the "+" from "hawguy-somesite@mydomain.com" and send an email to "hawguy@mydomain.com", it'll go straight to my spam folder and I'll never see it. The worst a spammer can do is make it seem like his spam is coming from another merchant, so he might change "hawguy-slashdot@mydomain.com to "hawguy-bestbuy@mydomain.com", but there's probably not much reason for him to do so.
Sometimes when looking in my spam email, I've found a number of variations on the username of the email address, usually where the spammer adds a number to the end of the name, like "hawguy-somesite067@mydomain.com". It must be accidental corruption on their side, or maybe they bought a list of a million "valid" email addresses from someone, and got ripped off because the emails were altered to be invalid.
I like the previous poster's scheme of using a wildcard DNS record to put the unique key in the domain name rather than the username.
Agreed. But obviously they have done a cost-benefit analysis and decided against this so far.
I personally like the Google 2-step authentication. Send a temporary code to my phone.
I too like the Google 2-step authentication, but I'm probably screwed if someone steals my phone since they'll have access to my email and SMS verification. I have a 5 digit PIN on the phone, but I don't know how secure a phone PIN is against a determined hacker. If my phone is lost or stolen, hopefully I can send a remote wipe before they hack it.
One thing that I do that inadvertently helps protect me against hack attacks is that I always use a unique email address when I sign up at a site, something like "hawguy+thesitename@mydomain.com" where "thesitename" is the domain name of the site I'm signing up for. I originally did this so I could easily block spam from a particular site, but I think it also helps a bit with security.
So it makes it harder to guess the email address I signed up with (trivial for a human to figure out, but I don't think it's common enough for the automated tools to do it), and if someone steals my "hawguy+thesitename@mydomain.com" email address and password from some site, they can't use that same email address to hack into another site even if I use the same password (and I tend to use the same password for all unimportant sites (like sites that I only use for commenting).
When I spent some time in Japan, I didn't watch any American TV and nothing bad happened. The USA was still there when I came back, and it turns out that I didn't miss any shows worth watching (and most of what I would have watched is on Netflix anyway).
Oh, geez. There's so many unhelpful, sarcastic ways this can be answered. Let's just stick to two:
Yes, but given how many doomsayers there are nowadays, the USA might not be there tomorrow. Or at the very least, November, when the entire country will end if we don't vote correctly. It's frustrating, really.
Even in the USA, I get my news (political or otherwise) through the web, not TV
Okay, now, turn that around, smart guy. How many shows in Japan do you miss?
Well, I'm not really that much in to TV, so there's no shows in Japan that I miss enough to watch to watch from here. When I'm in Japan, I love to watch the morning talk shows where they visit restaurants and try our their specialties, but not enough to go out of my way to watch them.
I like that Amazon lets me pay for the level of redundancy I need - a small bump for multiple availability zone within a single region redundancy, with a larger bump for multiple region redundancy. Not everything I do needs to ride out an East Coast hurricane, but for those things that do, it's really not hard to have a backup site in a different region
Isn't the main argument of the Cloud that you shouldn't worry about infrastructure and that your data will always be accessible if your internet connexion is UP ?
No, that's the vague marketing promise of the cloud but until you read your SLA and dig into the actual details of the service, it's really no promise at all. Many cloud and SaaS offerings have no geographical diversity. I've reviewed a number of services that have "offsite backups" as their disaster recovery plan, and even for those that say they do live data replication offsite, not all of them have servers located offsite, they still need their primary site to come back online before they can restore services. So sure, your data may be safe but you have no way to get to it until they restore their datacenter.
As with all IT services, "buyer beware" applies - make sure you know what you're buying before you buy it.
When I spent some time in Japan, I didn't watch any American TV and nothing bad happened. The USA was still there when I came back, and it turns out that I didn't miss any shows worth watching (and most of what I would have watched is on Netflix anyway).
Few customers have the knowledge and experience to touch the hardware and see if it's "done right". If they had that much expertise in-house, they'd probably just set it up themselves.
Okay, I don't have any data to back this up, do you? I know that in our case, our customers see the cloud buzzword, and so marketing puts it into the proposals, and since it's normally less expensive to outsource this, we end up going that route. It has nothing to do with our own lack of expertise, just that we're too cheap to pay for our own. Think of it like this...I can mow my own lawn, but I choose to pay the kid down the street because my time is more valuable than what it costs to pay him.
I have my company as a datapoint. We have a few dozen servers to colocate -- one or two rack's worth.
Even if we wanted to build a datacenter, we don't have staff on hand that undertstand enough about datacenter design and maintenance to run it. We don't even have a CCIE to provide us with a reliable, scalable internet connection across multiple carriers, nor do we have any cost effective way to bring internet connections from multiple carriers without running them all through the same half mile of conduit. leaving our internet connection subject to a single point of failure.
We could buy a facility more suited to running our critical infrastucture, and hire all of the experts we need to build and manage our datacenter (i.e. we we're not going to have a full-time HVAC engineer to keep our CRAC running, we'd outsource that to a vendor. Likewise we probably aren't going to hire a power engineer to maintain our UPS's, we'd outsource that to a vendor. And we're not going to have a diesel mechanic standing by to maintain our diesel generators, we'd outsource that to a vendor. And so on.)
We'd probably have a network engineer or two on staff and some server admins, but most of our critical infrastructure would be outsourced to vendors. And if we're already outsourcing much of our critical infrastructure support to vendors, is that really much different than outsourcing it all to a Cloud Provider (or colocation center, which is almost the same - Amazon is just one step above a coloc).
The nice thing about using a provider like Amazon is that we can get geographical redundancy for very little additional cost (we don't have to spin up our backup servers until we need to failover so most of our costs for a backup site are for storage at that second site), while if we built our own datacenters, we'd have to build 2 (or more) datacenters ourselves, each housing a rack or two of equipment. That's a lot of capital expenditure for very little gain (if anything) over what Amazon can provide for us.
Our marketing department didn't push us to the cloud, our desire for scalable infrastructure that's far better than we could afford to build ourselves is what pushed us to the cloud.
This is actually one of the major risks with "cloud". When you run your own data centers, you can touch the hardware, talk to the people, and check behind them to make sure things are actually being done right. In the worst case with cloud, you simply trust that "their data is safe", when in fact it might not be at all. In the less bad case, you get a nice contract with SLAs that specify exactly what data being safe means, and what recourse you have if they blow it. This is still not great, because if the past 5 years have taught you nothing else, they should have taught you that YES companies will make bets that end their business if they bet wrong.
I wouldn't say don't use the so-called cloud providers. Just don't naively believe they're doing everything right just because they haven't had a catastrophic failure or screwed up YOUR data yet.
Few customers have the knowledge and experience to touch the hardware and see if it's "done right". If they had that much expertise in-house, they'd probably just set it up themselves.
But for the vast majority of customers that just want someplace to host a few servers, nearly any cloud provider is going to be better than doing it themselves. Category 5 hurricanes are rare, local power failures and ISP outages are much more common and that's what's going to take down most do-it-yourself small-time hosters that run a couple servers in their wiring closet on a UPS, single source ISP connection and maybe a backup generator if they are lucky (but it's not regularly tested or maintained).
I've never heard of a major cloud storage facility that would keep all their servers in one location. They usually have all their data backed up to remote locations, usually far from their main site. We are taking about Amazon, and Google here, not Black Berry RIM. I'm sure their data is safe.
When the cloud storage provider is quite clear about what level of redundancy they provide, you probably shouldn't assume anything -- read what they are providing.
Amazon is quite clear about the distinction between availability zones and regions, and if you're going to host your critical app somewhere, you should probably understand what you're paying for.
I'm sure there are cloud providers that will give you the level of redundancy you're seeking, but probably not at the same price that Amazon charges.
Anyone that hosts their entire web presence at Amazon Virginia (especially after the other outages they've had), or really, in any single Amazon region is getting what they pay for and what they deserve if there's a regional disaster.
It's not hard or expensive to have a cold- or warm- spare site in a different region ready to take over (even if it's a manual cutover), especially since Amazon's new(ish) US-West region in Oregon is the same price as US-East.
I like that Amazon lets me pay for the level of redundancy I need - a small bump for multiple availability zone within a single region redundancy, with a larger bump for multiple region redundancy. Not everything I do needs to ride out an East Coast hurricane, but for those things that do, it's really not hard to have a backup site in a different region.
So what happens when the system goes down? You've suddenly got nobody there who is capable of taking the orders. If it's anything like the coffee shop I had the misfortune of working for the staff aren't allowed to take on each other's roles if there's a problem. I was once threatened with disciplinary action because whilst working alone I was taking orders and serving at the same time (standard practice in the bar trade).
The same thing that happens at the airport when the airline's automated check-in kiosks are down - they muster all of the employees they can to manually take orders meanwhile the queue backs up out the door and everyone is unhappy, but Starbucks saved a few bucks on labor so it looks good to investors even if it annoys customers.
Fortunately, there's still some little competition for coffee, but if Starbucks continues to buy up small chains (like Peets), that may not be the case forever. Small coffee shops aren't always where you want them, Starbucks is everywhere.
Exactly -- when I saw that they'd Sbux scan a barcode, I thought that sounded ok (but what happened toe NFC being the ultimate solution to contactless pay-by-phone!?). But when I saw that Square wanted to be able to track my movements at all times via GPS so my phone can automatically authorize payments at a merchant anytime I walk in the door, that's when I realized that I'm sticking to credit cards.
I don't even mind letting the merchant know that I've walked in their door since they're going to know one way or another, whether I use my phone to pay, my credit card, or my starbucks card, but I don't want to let Square track my every move 24 hours/day and then sell that data to other merchants ("Hey, hawguy walked in your store 23 times over the past month, but he only made a purchase twice. Pay us $$ and we'll pop up an ad on his phone next time he walks in your store"). Or worse, send me promotions based on where I've been so my wife will say "Hey honey, why does Square keep sending you Victoria's Secret promotions when you haven't set foot in that store for years -- who are you buying lingerie for!?!"
the demands of each individual user aren't likely to increase much beyond that.
I think your thinking is way too constrained. If the bandwidth was available, then people could have immersive 3D working environments, and tele-commuting could be far more common. This would result in much less traffic on the roads and a huge reduction in CO2 emissions and oil imports. This is not science fiction. I have used Cisco's "Virtual Meeting Room" and it is pretty good.
You also need to think about things like "Siri", that send audio back to the server for processing, because there isn't enough horsepower in a cellphone. I could see "smart glasses" of the future sending video back to a server. That will require huge bandwidth.
If the bandwidth is available and affordable, the applications will come.
I work in a large multi-building "campus" (well, more of an office park, we have offices in several buildings). It's a 15 - 20 minute walk from one building to the farthest one (depending on who is doing the walking)
We have practically unlimited bandwidth between buildings (and at least a gigabit to remote offices) yet we still make people trudge between buildings for meetings, and teleconferences with remote sites are 720p (or Skype). So bandwidth isn't constraining us from immersive teleconferencing - we'd probably save a dozen man-hours every day in eliminating the need to walk between buildings for meetings which would easily pay for an immersive teleconference system (10 man hours * 250 days/year * $50/hour = $125K/year in labor savings), yet we don't even use the teleconference system we have now for meetings between buildings.
The comparison was actually with the laptop being next to the TV, so that's about as valid as I can get it. I've even found they can see a difference between a 1280x720 TV and my 1600x900 monitor, and that's much less a difference in physical size AND in resolution.
That's about as invalid as you can get. You're comparing a 100+ dpi laptop screen with a 40 or 50 dpi TV screen. Of course people are going to like the sharper screen of the laptop better.
but considering that most people can't tell the difference between 720p and 1080p, I doubt that will ever happen.
Uh, what?
I can see the difference. I can even see the difference between 1080p and 1440p, or 1440p and 2160p. And it's not a slight difference that I could understand people missing. In informal tests, comparing my laptop playing 1080p video to my parent's 720p "HDTV", 100% of those surveyed responded "holy crap that looks better" (margin of error for 95% confidence interval: 9.38%).
You're not "most people" - "most people" haven't even seen 1440p.
And how do you make any sort of fair comparison between a 17" laptop screen and a 32" (or larger?) HDTV? There's no way to fairly compare the two because of the screen size difference.
At normal viewing distances, most people can't see the difference between 720p and 1080p -- you'd need to be within 5 feet of your 40" TV to see the difference. Sure, maybe you have a home theater with a 60" TV and seats 6 feet away, but most people have a TV in the corner of the living room and don't arrange seating for optimal 1080p viewing distance.
Why does he say "it will never be possible to increase bandwidth fast enough to keep up with demand"?
When I want to watch streaming video, I fire up Netflix and watch streaming video. When I want to download a large media file, I find it on bittorrent and download it. The only time I've noticed any internet slowdowns, it's been in my ISP's network, and it's just a transient problem that eventually goes away.
Sure, Netflix has to do some extra work to create a content delivery network to deliver the content near to where I am, but it sounds like the internet is largely keeping up with demand.
Aside from the IPv4->IPv6 transition (we've been a year away from running out of IP addresses for years), is there some impending bandwidth crunch that will kill the internet?
RUNNING machines in plastic bags? I'm not sure you understood the problem domain.
You don't understand the solution.
Put the equipment into plastic bags before loading it on the truck (if it's new equipment, it's probably already wrapped in plastic.
Then when you unload it in the warm datacenter, the moisture condenses on the outside of the bag instead of inside your server.
Once the server is up to room temperature, take it out of the bag, rack it and plug it in, and you're good to go. (Note that in warm humid states, you can have the opposite problem - the cold server is taken from the 65 degree datacenter out to the 95 degree and humid outside air and moisture condenses on it. The plastic bag works here too.)
You don't need to go to Sweden to experience cold temperatures - many datacenters throughout the USA experience temperatures cold enough to cause condensation problems for at least part of the year. The plastic also helps protect equipment that's exposed to moisture that condenses in clouds and falls to the ground (i.e. rain) as it's transferred from the truck to the facility. A problem that Facebook will discover once they open their first datacenter in a rainforest and perhaps they can invent some self deploying canopy that shields the equipment from this mysterious moisture from the sky since they don't seem to like the low-tech plastic bag solution.
I'm all for hating on the telcos, but sometimes "just build more towers" is much, much easier said than done. For instance, it takes three years to get one built in San Francisco. Granted, not every place is as downright insane as San Francisco is, but it's worth mentioning.
Then they had better get started today -- maybe they could spend less money telling me how fabulous their ultra-fast 4G network is (letting me use an entire month's data cap within 15 minutes), and more money on building out that network so I can actually use it.
I live in a busy urban area and have no 4G coverage at all within a block of my house.
I find it convenient when PayPal is a payment option so that I do not have to provide my credit card information to every vendor / seller on the World Wide Web.
The downside being that you DO have to give your card information to PayPal, who weren't exactly the most trustworthy bunch of bastards BEFORE being borged by eBay.
I trust Paypal with my credit card number much more than some random internet merchant. I've had my credit card number replaced 3 times after it was stolen from merchants. Twice initiated by my credit card company, saying that they were replacing my card due to a potential merchant breach, and once after I noticed the fraudulent charges and started receiving spams to an email address used only by a particular merchant.
Any reason not to try to embed little experiments in the weights though?
My guess is density - it's the densest common metal (1.7 times the density of lead), so it takes up less much space than an experiment pod would use. Plus the center of gravity of the tungsten weight is easy to calculate, while an experiment pod's center of gravity could shift if the materials inside move around. 150kg of Tungsten takes up the space of a cube of around 20cm on each side. Even if the experiment pod was a block of solid aluminum, each side of the cube would need to be around 40cm, so would be much bigger.
Stupid discussions like this one arise when perfectly useful expressions are replaced by inferior, more complex expressions merely to make people sound more expert or scientific. In the olden days, they would have said "Live images will be beamed back by the lander".
I don't believe saying it's a "live image" would be any less likely to be challenged. "How can it be live if it happened 15 minutes ago!?"
No one would have been in any doubt what that meant and this bullshit discussion would never have needed to take place, in 'real time' or slightly delayed.
I'm seeing about 100 msec of latency to Slashdot from here, so this conversation is not taking place in "real time".
You draw the line at any signal latency that is too slow to meaningfully respond to in the context that the signal was originally sent from. There's a reason why interrupt handlers in real-time OS's need to finish their job in as few computing cycles as possible.
So in the event of a fully autonomous landing that wasn't designed for any human input or control, is a 15 minute delay "real time" or not? Once the landing sequence starts, there's no going back or adjusting the sequence until the entire 15 minute landing sequence is over, even if a human was orbiting Mars.
An interrupt handler in a Real-Time OS doesn't need to finish its job in as few computing cycles as possible, it only needs to finish within the guaranteed interrupt latency time, which could be a few microseconds or could be 15 minutes. Real time doesn't mean fast, it means it meets hard deadlines.
Slashdot Mirror
A lot of people are telling you that you're wrong, that you need "offsite backup", etc, etc. But that's not what you asked. You've got an external Hard drive already it sounds like. Most banks will rent a safe deposit box for fairly cheap. This will give you that "Fireproof" that you need. You're either going to need a computer with USB 3.0 or external hard drive(s) with ethernet capability.
Didn't you just describe an "offsite backup"? Offsite backup doesn't necessarily mean data backed up live to the cloud. My company's offsite backup strategy is to have tapes picked up by Iron Mountain every week. Weekly backups are stored at a "local" location (about 40 miles from our building), quarterly full dumps are sent to a different facility out of state.
People are telling him he needs "offsite backup" because there's no way to keep his data safe if it's all on-site. Most people aren't willing to drive their data far enough off-site to keep it safe. Putting it your local bank's safe deposit box is only safe until a regional disaster hits both your house and the bank... Iike an earthquake or flood.
There is no "safe" on-site storage option. My brother had an expensive expensive fire-proof gun safe at his house - it weighed over 500 lbs.... After his house literally burnt to the ground (rural area, no one home to alert the first department, by the time the neighbor down the road saw the fire and called the fire department, all they could do was put out the fire in the surrounding brush (and cars)), he couldn't even identify any remaining pieces of the safe or the guns that were in it in the debris.
If you want your data safe, store it off site, preferably in another part of the country so it's not subject to the same local disasters as your house. Mailing snapshots of your data to an out of state relative is probably best.
Because +@domain.com is not easy to figure out? I'd be very surprised if a spammer that acquires your garbage email address can't obtain your actual email address.
What is my "actual" email address? I have around a dozen addresses that are important because I've given to family and friends, (and literally hundreds that I've used to sign up for various newsletters, blog sites, etc). If you just drop the part after the "+" from "hawguy-somesite@mydomain.com" and send an email to "hawguy@mydomain.com", it'll go straight to my spam folder and I'll never see it. The worst a spammer can do is make it seem like his spam is coming from another merchant, so he might change "hawguy-slashdot@mydomain.com to "hawguy-bestbuy@mydomain.com", but there's probably not much reason for him to do so.
Sometimes when looking in my spam email, I've found a number of variations on the username of the email address, usually where the spammer adds a number to the end of the name, like "hawguy-somesite067@mydomain.com". It must be accidental corruption on their side, or maybe they bought a list of a million "valid" email addresses from someone, and got ripped off because the emails were altered to be invalid.
I like the previous poster's scheme of using a wildcard DNS record to put the unique key in the domain name rather than the username.
Agreed. But obviously they have done a cost-benefit analysis and decided against this so far.
I personally like the Google 2-step authentication. Send a temporary code to my phone.
I too like the Google 2-step authentication, but I'm probably screwed if someone steals my phone since they'll have access to my email and SMS verification. I have a 5 digit PIN on the phone, but I don't know how secure a phone PIN is against a determined hacker. If my phone is lost or stolen, hopefully I can send a remote wipe before they hack it.
One thing that I do that inadvertently helps protect me against hack attacks is that I always use a unique email address when I sign up at a site, something like "hawguy+thesitename@mydomain.com" where "thesitename" is the domain name of the site I'm signing up for. I originally did this so I could easily block spam from a particular site, but I think it also helps a bit with security.
So it makes it harder to guess the email address I signed up with (trivial for a human to figure out, but I don't think it's common enough for the automated tools to do it), and if someone steals my "hawguy+thesitename@mydomain.com" email address and password from some site, they can't use that same email address to hack into another site even if I use the same password (and I tend to use the same password for all unimportant sites (like sites that I only use for commenting).
One tower per home is called WIFI. Look it up some time.
I looked it up, but my dictionary said femto-cell.
When I spent some time in Japan, I didn't watch any American TV and nothing bad happened. The USA was still there when I came back, and it turns out that I didn't miss any shows worth watching (and most of what I would have watched is on Netflix anyway).
Oh, geez. There's so many unhelpful, sarcastic ways this can be answered. Let's just stick to two:
Yes, but given how many doomsayers there are nowadays, the USA might not be there tomorrow. Or at the very least, November, when the entire country will end if we don't vote correctly. It's frustrating, really.
Even in the USA, I get my news (political or otherwise) through the web, not TV
Okay, now, turn that around, smart guy. How many shows in Japan do you miss?
Well, I'm not really that much in to TV, so there's no shows in Japan that I miss enough to watch to watch from here. When I'm in Japan, I love to watch the morning talk shows where they visit restaurants and try our their specialties, but not enough to go out of my way to watch them.
I like that Amazon lets me pay for the level of redundancy I need - a small bump for multiple availability zone within a single region redundancy, with a larger bump for multiple region redundancy. Not everything I do needs to ride out an East Coast hurricane, but for those things that do, it's really not hard to have a backup site in a different region
Isn't the main argument of the Cloud that you shouldn't worry about infrastructure and that your data will always be accessible if your internet connexion is UP ?
No, that's the vague marketing promise of the cloud but until you read your SLA and dig into the actual details of the service, it's really no promise at all. Many cloud and SaaS offerings have no geographical diversity. I've reviewed a number of services that have "offsite backups" as their disaster recovery plan, and even for those that say they do live data replication offsite, not all of them have servers located offsite, they still need their primary site to come back online before they can restore services. So sure, your data may be safe but you have no way to get to it until they restore their datacenter.
As with all IT services, "buyer beware" applies - make sure you know what you're buying before you buy it.
When I spent some time in Japan, I didn't watch any American TV and nothing bad happened. The USA was still there when I came back, and it turns out that I didn't miss any shows worth watching (and most of what I would have watched is on Netflix anyway).
Few customers have the knowledge and experience to touch the hardware and see if it's "done right". If they had that much expertise in-house, they'd probably just set it up themselves.
Okay, I don't have any data to back this up, do you? I know that in our case, our customers see the cloud buzzword, and so marketing puts it into the proposals, and since it's normally less expensive to outsource this, we end up going that route. It has nothing to do with our own lack of expertise, just that we're too cheap to pay for our own. Think of it like this...I can mow my own lawn, but I choose to pay the kid down the street because my time is more valuable than what it costs to pay him.
I have my company as a datapoint. We have a few dozen servers to colocate -- one or two rack's worth.
Even if we wanted to build a datacenter, we don't have staff on hand that undertstand enough about datacenter design and maintenance to run it. We don't even have a CCIE to provide us with a reliable, scalable internet connection across multiple carriers, nor do we have any cost effective way to bring internet connections from multiple carriers without running them all through the same half mile of conduit. leaving our internet connection subject to a single point of failure.
We could buy a facility more suited to running our critical infrastucture, and hire all of the experts we need to build and manage our datacenter (i.e. we we're not going to have a full-time HVAC engineer to keep our CRAC running, we'd outsource that to a vendor. Likewise we probably aren't going to hire a power engineer to maintain our UPS's, we'd outsource that to a vendor. And we're not going to have a diesel mechanic standing by to maintain our diesel generators, we'd outsource that to a vendor. And so on.)
We'd probably have a network engineer or two on staff and some server admins, but most of our critical infrastructure would be outsourced to vendors. And if we're already outsourcing much of our critical infrastructure support to vendors, is that really much different than outsourcing it all to a Cloud Provider (or colocation center, which is almost the same - Amazon is just one step above a coloc).
The nice thing about using a provider like Amazon is that we can get geographical redundancy for very little additional cost (we don't have to spin up our backup servers until we need to failover so most of our costs for a backup site are for storage at that second site), while if we built our own datacenters, we'd have to build 2 (or more) datacenters ourselves, each housing a rack or two of equipment. That's a lot of capital expenditure for very little gain (if anything) over what Amazon can provide for us.
Our marketing department didn't push us to the cloud, our desire for scalable infrastructure that's far better than we could afford to build ourselves is what pushed us to the cloud.
This is actually one of the major risks with "cloud". When you run your own data centers, you can touch the hardware, talk to the people, and check behind them to make sure things are actually being done right. In the worst case with cloud, you simply trust that "their data is safe", when in fact it might not be at all. In the less bad case, you get a nice contract with SLAs that specify exactly what data being safe means, and what recourse you have if they blow it. This is still not great, because if the past 5 years have taught you nothing else, they should have taught you that YES companies will make bets that end their business if they bet wrong.
I wouldn't say don't use the so-called cloud providers. Just don't naively believe they're doing everything right just because they haven't had a catastrophic failure or screwed up YOUR data yet.
Few customers have the knowledge and experience to touch the hardware and see if it's "done right". If they had that much expertise in-house, they'd probably just set it up themselves.
But for the vast majority of customers that just want someplace to host a few servers, nearly any cloud provider is going to be better than doing it themselves. Category 5 hurricanes are rare, local power failures and ISP outages are much more common and that's what's going to take down most do-it-yourself small-time hosters that run a couple servers in their wiring closet on a UPS, single source ISP connection and maybe a backup generator if they are lucky (but it's not regularly tested or maintained).
I've never heard of a major cloud storage facility that would keep all their servers in one location. They usually have all their data backed up to remote locations, usually far from their main site. We are taking about Amazon, and Google here, not Black Berry RIM. I'm sure their data is safe.
When the cloud storage provider is quite clear about what level of redundancy they provide, you probably shouldn't assume anything -- read what they are providing.
Amazon is quite clear about the distinction between availability zones and regions, and if you're going to host your critical app somewhere, you should probably understand what you're paying for.
I'm sure there are cloud providers that will give you the level of redundancy you're seeking, but probably not at the same price that Amazon charges.
Anyone that hosts their entire web presence at Amazon Virginia (especially after the other outages they've had), or really, in any single Amazon region is getting what they pay for and what they deserve if there's a regional disaster.
It's not hard or expensive to have a cold- or warm- spare site in a different region ready to take over (even if it's a manual cutover), especially since Amazon's new(ish) US-West region in Oregon is the same price as US-East.
I like that Amazon lets me pay for the level of redundancy I need - a small bump for multiple availability zone within a single region redundancy, with a larger bump for multiple region redundancy. Not everything I do needs to ride out an East Coast hurricane, but for those things that do, it's really not hard to have a backup site in a different region.
So what happens when the system goes down? You've suddenly got nobody there who is capable of taking the orders. If it's anything like the coffee shop I had the misfortune of working for the staff aren't allowed to take on each other's roles if there's a problem. I was once threatened with disciplinary action because whilst working alone I was taking orders and serving at the same time (standard practice in the bar trade).
The same thing that happens at the airport when the airline's automated check-in kiosks are down - they muster all of the employees they can to manually take orders meanwhile the queue backs up out the door and everyone is unhappy, but Starbucks saved a few bucks on labor so it looks good to investors even if it annoys customers.
Fortunately, there's still some little competition for coffee, but if Starbucks continues to buy up small chains (like Peets), that may not be the case forever. Small coffee shops aren't always where you want them, Starbucks is everywhere.
Exactly -- when I saw that they'd Sbux scan a barcode, I thought that sounded ok (but what happened toe NFC being the ultimate solution to contactless pay-by-phone!?). But when I saw that Square wanted to be able to track my movements at all times via GPS so my phone can automatically authorize payments at a merchant anytime I walk in the door, that's when I realized that I'm sticking to credit cards.
I don't even mind letting the merchant know that I've walked in their door since they're going to know one way or another, whether I use my phone to pay, my credit card, or my starbucks card, but I don't want to let Square track my every move 24 hours/day and then sell that data to other merchants ("Hey, hawguy walked in your store 23 times over the past month, but he only made a purchase twice. Pay us $$ and we'll pop up an ad on his phone next time he walks in your store"). Or worse, send me promotions based on where I've been so my wife will say "Hey honey, why does Square keep sending you Victoria's Secret promotions when you haven't set foot in that store for years -- who are you buying lingerie for!?!"
the demands of each individual user aren't likely to increase much beyond that.
I think your thinking is way too constrained. If the bandwidth was available, then people could have immersive 3D working environments, and tele-commuting could be far more common. This would result in much less traffic on the roads and a huge reduction in CO2 emissions and oil imports. This is not science fiction. I have used Cisco's "Virtual Meeting Room" and it is pretty good.
You also need to think about things like "Siri", that send audio back to the server for processing, because there isn't enough horsepower in a cellphone. I could see "smart glasses" of the future sending video back to a server. That will require huge bandwidth.
If the bandwidth is available and affordable, the applications will come.
I work in a large multi-building "campus" (well, more of an office park, we have offices in several buildings). It's a 15 - 20 minute walk from one building to the farthest one (depending on who is doing the walking)
We have practically unlimited bandwidth between buildings (and at least a gigabit to remote offices) yet we still make people trudge between buildings for meetings, and teleconferences with remote sites are 720p (or Skype). So bandwidth isn't constraining us from immersive teleconferencing - we'd probably save a dozen man-hours every day in eliminating the need to walk between buildings for meetings which would easily pay for an immersive teleconference system (10 man hours * 250 days/year * $50/hour = $125K/year in labor savings), yet we don't even use the teleconference system we have now for meetings between buildings.
The comparison was actually with the laptop being next to the TV, so that's about as valid as I can get it. I've even found they can see a difference between a 1280x720 TV and my 1600x900 monitor, and that's much less a difference in physical size AND in resolution.
That's about as invalid as you can get. You're comparing a 100+ dpi laptop screen with a 40 or 50 dpi TV screen. Of course people are going to like the sharper screen of the laptop better.
but considering that most people can't tell the difference between 720p and 1080p, I doubt that will ever happen.
Uh, what?
I can see the difference. I can even see the difference between 1080p and 1440p, or 1440p and 2160p. And it's not a slight difference that I could understand people missing. In informal tests, comparing my laptop playing 1080p video to my parent's 720p "HDTV", 100% of those surveyed responded "holy crap that looks better" (margin of error for 95% confidence interval: 9.38%).
You're not "most people" - "most people" haven't even seen 1440p.
And how do you make any sort of fair comparison between a 17" laptop screen and a 32" (or larger?) HDTV? There's no way to fairly compare the two because of the screen size difference.
At normal viewing distances, most people can't see the difference between 720p and 1080p -- you'd need to be within 5 feet of your 40" TV to see the difference. Sure, maybe you have a home theater with a 60" TV and seats 6 feet away, but most people have a TV in the corner of the living room and don't arrange seating for optimal 1080p viewing distance.
http://carltonbale.com/1080p-does-matter/
Why does he say "it will never be possible to increase bandwidth fast enough to keep up with demand"?
When I want to watch streaming video, I fire up Netflix and watch streaming video. When I want to download a large media file, I find it on bittorrent and download it. The only time I've noticed any internet slowdowns, it's been in my ISP's network, and it's just a transient problem that eventually goes away.
Sure, Netflix has to do some extra work to create a content delivery network to deliver the content near to where I am, but it sounds like the internet is largely keeping up with demand.
Aside from the IPv4->IPv6 transition (we've been a year away from running out of IP addresses for years), is there some impending bandwidth crunch that will kill the internet?
When I first read the summary, I thought they meant Hasbro's Game of Life, which as a child, is much more fun than a cellular automaton.
RUNNING machines in plastic bags? I'm not sure you understood the problem domain.
You don't understand the solution.
Put the equipment into plastic bags before loading it on the truck (if it's new equipment, it's probably already wrapped in plastic.
Then when you unload it in the warm datacenter, the moisture condenses on the outside of the bag instead of inside your server.
Once the server is up to room temperature, take it out of the bag, rack it and plug it in, and you're good to go. (Note that in warm humid states, you can have the opposite problem - the cold server is taken from the 65 degree datacenter out to the 95 degree and humid outside air and moisture condenses on it. The plastic bag works here too.)
You don't need to go to Sweden to experience cold temperatures - many datacenters throughout the USA experience temperatures cold enough to cause condensation problems for at least part of the year. The plastic also helps protect equipment that's exposed to moisture that condenses in clouds and falls to the ground (i.e. rain) as it's transferred from the truck to the facility. A problem that Facebook will discover once they open their first datacenter in a rainforest and perhaps they can invent some self deploying canopy that shields the equipment from this mysterious moisture from the sky since they don't seem to like the low-tech plastic bag solution.
I'm all for hating on the telcos, but sometimes "just build more towers" is much, much easier said than done. For instance, it takes three years to get one built in San Francisco. Granted, not every place is as downright insane as San Francisco is, but it's worth mentioning.
Then they had better get started today -- maybe they could spend less money telling me how fabulous their ultra-fast 4G network is (letting me use an entire month's data cap within 15 minutes), and more money on building out that network so I can actually use it.
I live in a busy urban area and have no 4G coverage at all within a block of my house.
I find it convenient when PayPal is a payment option so that I do not have to provide my credit card information to every vendor / seller on the World Wide Web.
The downside being that you DO have to give your card information to PayPal, who weren't exactly the most trustworthy bunch of bastards BEFORE being borged by eBay.
I trust Paypal with my credit card number much more than some random internet merchant. I've had my credit card number replaced 3 times after it was stolen from merchants. Twice initiated by my credit card company, saying that they were replacing my card due to a potential merchant breach, and once after I noticed the fraudulent charges and started receiving spams to an email address used only by a particular merchant.
Any reason not to try to embed little experiments in the weights though?
My guess is density - it's the densest common metal (1.7 times the density of lead), so it takes up less much space than an experiment pod would use. Plus the center of gravity of the tungsten weight is easy to calculate, while an experiment pod's center of gravity could shift if the materials inside move around. 150kg of Tungsten takes up the space of a cube of around 20cm on each side. Even if the experiment pod was a block of solid aluminum, each side of the cube would need to be around 40cm, so would be much bigger.
Stupid discussions like this one arise when perfectly useful expressions are replaced by inferior, more complex expressions merely to make people sound more expert or scientific. In the olden days, they would have said "Live images will be beamed back by the lander".
I don't believe saying it's a "live image" would be any less likely to be challenged. "How can it be live if it happened 15 minutes ago!?"
No one would have been in any doubt what that meant and this bullshit discussion would never have needed to take place, in 'real time' or slightly delayed.
I'm seeing about 100 msec of latency to Slashdot from here, so this conversation is not taking place in "real time".
You draw the line at any signal latency that is too slow to meaningfully respond to in the context that the signal was originally sent from. There's a reason why interrupt handlers in real-time OS's need to finish their job in as few computing cycles as possible.
So in the event of a fully autonomous landing that wasn't designed for any human input or control, is a 15 minute delay "real time" or not? Once the landing sequence starts, there's no going back or adjusting the sequence until the entire 15 minute landing sequence is over, even if a human was orbiting Mars.
An interrupt handler in a Real-Time OS doesn't need to finish its job in as few computing cycles as possible, it only needs to finish within the guaranteed interrupt latency time, which could be a few microseconds or could be 15 minutes. Real time doesn't mean fast, it means it meets hard deadlines.