Perhaps it would be a good idea to start by defining what exactly "cloud computing" means?
Because looking at the Wikipedia article I see only a brief mention of reliability through redundancy: Reliability is improved if multiple redundant sites are used, which makes well-designed cloud computing suitable for business continuity and disaster recovery.
As CTO of a data hosting "cloud services" provider myself, I'm proud of our track record for reliability and redundancy. All our systems are backed up offsite to multiple locations every 24 hours. At all times we provide off-site hosting facilities at about 50% of the capacity of our primary production cluster for disaster recovery scenarios. All our systems are redundant on site; the loss of any single system can be healed within an hour or so. Our system uptime averages approaches 4 nines over years in an industry where 2 nines is considered exceptional.
Having systems replicate in near-real-time to multiple locations and autonomically route around large network outages on extremely short notice is an extremely tough and expensive thing to do. It sounds simple enough, but the devil is in the details. The number of things that can go wrong is simply staggering, and trying to account for every possible thing that can go wrong is simply infeasible to proactively account for.
What do you do when the problem is due to a router outage offsite? What if 70% of routes work, and 30% don't due to a BGP burp? What if latency is 50ms? 100ms? 200ms? At what point is it "down" ? What if network conditions are excellent but there's a problem with DNS? What if a power surge generated by your UPS (yes, it happens!) takes out 20% of your production cluster? Do you heal or switch to hot backup? What if it takes out 30% of your production cluster? 40% 60%?
People think of "online" is a boolean yes/no question, but it's not. We calculate our uptimes based on expected end-user experience, but rarely is our production cluster actually working at 100%. Nearly always, some system or other is in need of attention and it doesn't constitute an emergency because there's still additional redundancy in the system.
Hell, even working for low wage, a person is expensive. Thus the most effort should be put in having them do the least effort.
Yeah, but here's where it gets weird: Software is highly valuable because it allows a single person to do the same task a very large number of times.
Because of this, it's not like stacking logs or gluing tile. If you have sufficient leverage, you can spend a stupid amount of money getting everything just right and profit immensely from it. (See: Apple)
On the other hand, most software isn't leveraged nearly as much. Most software is "commodity" and doesn't represent a huge leverage of anybody's skill. And that's why, in some environments, egging out the last ms of performance out of an algorithm is critical, and in others it's a total waste of money.
The important part is to determine just how important the possible performance increase actually might be.
I sooooo wish this were true! The problem is in concentrated wealth.
If I (a "little people") crash a car, the most anybody could get out of me would be my life savings, which (at 40) adds up to a few hundred thousand. Enough for an ambulance chaser and a douchebag to make my life suck, but not enough to bring out the big guns.
But when the "driver" of a car is a software company with millions of installs, any crash at all is enough for said ambulance chaser and douchebag to go for the jugular for millions. Ambulance chaser might get the backing of the big, tall, class-action corporate lawyer type of legal firm and turn this into a real law suit!
Even though it's safer. It's not what's sensible, it's what's legal.
It's pretty easy to put together a basic security system. Require an identity token of some sort, and require proof of knowledge of a secret, and you have the makings of a security system!
Security is not a boolean. Security is a variable, ranging from non at all to mild, moderate, to extremely secure.
Little things can greatly add greatly to real security (such as free permits for concealed weapons and password strength requirements), and big, obvious, "secure" things can easily be nothing more than theater. (EG: the TSA goons at the airports)
To be truly secure at the high end is surprisingly difficult. As the value of the prize increases in value, the number of potentially useful attacks increases exponentially. A dollar-store lock will reasonably protect a $50 used bike in most areas, but at $500, the lock has to be able to reasonably defend itself from something like a grinder. At $5,000, blow torches become reasonable, and at $50,000, plastic explosives are a fair bet.
See how much more difficult it gets to defend concentrated wealth? It's *hard* to do it right!
Windows server looked after by a good sysadmin == secure. Mac server looked after by bad sysadmin == insecure.
The sad part is that much (most?) of being a good sysadmin consists of ensuring that you install security updates regularly. I've been close enough to embarrassing hacks on several servers to know what happened, and all (but one!) have been hacked as a result of a poor update policy. (The last one was due to a weak root password + passwordAuthentication enabled on ssh)
For all my own systems, I demand a strong, default-deny firewall, and (most importantly!) regular, frequent updates. Tools like yum make this easy.
Yes, things like strictly firewalling your systems, moving services only used internally to nonstandard ports, disabling (or never installing) unneeded services, etc. are good ideas and I strongly recommend them all.
But first and foremost, make sure your system is up to date!
That's what a Smart Grid does -- people think it will let them pay spot prices for electricity, no no no, it's about the power company collecting Google-style metrics of power consumption on an appliance-by-appliance, outlet-by-outlet basis, and then giving you 10% off your bill if you consent to having a remote power cutoff installed on your washing machine, air conditioner, and car charger.
All TFA says is that it's easily possible to build a dynamic system that's unstable. Duh. In aviation, the attribute that you are looking for is called "positive dynamic stability" - which simply means that when things get interrupted or jarred, that the system actively works to mitigate the change and stabilize over time.
This is also something I achieve with a self-healing load-balanced computing cluster. Things happen, loads spike, etc. The system should gradually self-stabilize.
This is a problem as old as the presence dynamic systems.
The only problem with drones is knowing that they're there. They're inevitable, and why I watch the rollout of ADS-B with interest.
That's why I ALWAYS am on approach control in Echo airspace and up. Here in North Cali, Beale AFB has "temporary" flight restrictions going at least half the time for UAVs in the area, and the only real restriction is that you have to have a VFR flight following or be under IFR.
Having learned on the "steam guage panel" and now flying with a Garmin 496, I can declare with confidence that GPS technologies combined with good visual display and information technologies make a night vs day difference in flying!
I suspect that drones will simply be required to be ADS-B compliant, and that pretty much everybody will have the receivers installed on their aircraft.
If such a transmitter was located on you house, unless you had an antenna that ensured the radiation was not directed inward the building would be quite uninhabitable.
99.9998% of the power emitted by my pocket cell phone is wasted. Only the minute fraction of those waves that happen to coincide with the line between myself and the cell tower are actually converted into anything useful. Sadly, the 99.998% ratio is probably optimistic, it's likely to be considerably worse than that!
So here we are, covering the Earth with radio-emitting devices by the billions, (with a "B") with emissions ranging from a few milliwatts per device up into the millions. Thanks to the inverse-square law, the milli-watt cordless phone on your desk may contribute far more power than the 5 million watt TV transmitter a few miles away, but in all cases, there is a considerable amount of radio flux virtually ANYWHERE people are likely to reside.
Using ANY of it will cost (on average) less than 0.001% of available signal. This is about as green as it gets!
Remember that movie a few years back, "Zoro" with Antonio Banderas? (If not, it's probably up on Netflix Instant Play or a torrent someplace)
The movie was all about the "passing of the guard" - a new, younger man taking the role of "Zoro", the anonymous masked crime fighter of the previous generation. It's a good movie, so I recommend it highly. But it also does a passable job of showing the difference between an identity and an idea.
I'm guessing that there are, in fact, a half dozen or more actual people who have had the identity of "Topiary". They may have shared the pseudo-identity concurrently, so, who did what?
I'm getting out the popcorn and getting ready to watch the show!
Sure, some solutions are faster than MySQL out of the box by skipping much of the language parsing and stuff that any SQL solution has to do. But that's not to say that they are actually more efficient at key retrieval.
"Always have a proper backup! You can learn more about backups here""
Sorry, but this is the height of naive. CentOS casts a very *VERY* long shadow!
Do you want a nice, supported, "Enterprise" Linux but don't have much budget to spend? Guess what: CentOS is almost your only choice. SL is nice, but it's not binary compatible with RHEL. Mix and match a few packages with a few "EL5/6" repos and you very quickly will run into binary hell.
Other than RHEL, what "Enterprise" options are available? What you need is something that is stable, conservative in anything that changes the environment un-necessarily, and is likely to support 3rd party software from vendors.
What else is there? There's Oracle's re-branded RHEL, there's RHEL itself, and there's CentOS. Debian is almost good enough.
While there are plenty of Linux distros, there are surprisingly few other options available at the commercial/enterprise tier.
OP has very good advice. Take it. Understanding what negative and positive pressure can be very important in circumstances like this.
Ever wonder why dust collects so badly inside your PC? It's a negative pressure environment - the main power fan at the back blows hot air OUT of the computer, causing air to rush into every crack and orifice in your case, making your expensive electronics into a poor quality air filter. The dust collected is a byproduct of this fact.
I once was called in to deal with a computer in a very dusty environment. (they raised pets) Their computer required extremely frequent cleaning and despite this, they had numerous hardware failures. The CD ROM drive was pretty much always useless no matter how often it was replaced. Using a medium-sized box, a cheap 8" fan, and HEPA air filters and lots of duct tape, I made a large, low-pressure air filter that blew large amounts of HEPA purified air into the computer, creating a positive pressure inside the case of clean, filtered air. A year later, the computer had only traces of dust and was working perfectly, including the CD ROM drive!
In PPs example, you want to create a negative pressure environment to keep dust from getting OUT.
When a traffic light is broken, you are supposed to treat it as a stop sign. If it won't sense your bike, isn't it effectively broken? I've no idea how that would hold up in traffic court, but as an avid cyclist, I've acted on this for years without incident.
Especially this early in the life cycle of this "cloud" crap. Any expectation of not loosing your data if you don't keep a backup yourself is entirely your own fault.
Early? Come again? I've been providing "cloud computing" services for about a decade. Is 10 years early?
Of course, the names have changed. There was "Service Oriented Architecture" and before that it was "Software as a Service" and then before that "managed application hosting" and various other names for it over time. Name doesn't matter, it's still the same thing.
Backups, of the redundant and routinely verified type, have always been an integral of the solutions I've worked to provide. I have an archival record of nightly snapshots going back for *years*.
I'm going to make a dare: Go talk to "little people" whose livelyhood rests squarely on the state of their computers. I'm talking about small office clerks, college professors, gas station managers, single-practice lawyers, authors... even most schools and school district offices. Get them to describe how they back up their stuff (of course they do backups!) and then ask them to SHOW you their most recent backup. Excepting large organizations and/or wealthy firms, if you have a large enough sample size, you'll rarely find more than 10% or so of everybody has backups of their critical data made within the past week.
Reputable cloud providers make a concerted effort to provide reasonable backups of their data - it is their business, after all - and just like a mechanic is more likely to do a decent job wrenching a car than an average Joe, a professional hosting company is more likely to do a decent job hosting and the average Joe.
Of course, there tend to be lots of hosting professional types around/., so it's probably true that even a professional firm wouldn't out perform you.
The latency for a radio link of 250 miles is in the order of a few milliseconds, and 250 miles is enough range to make finding a semi-mobile "home base" somewhere between tough and impossible.
As a pilot, I routinely hear radio calls 250 miles away when flying 10,000 feet; having a radio-relay essentially circling near the home base at 10,000 feet to support drone activity for a 500 mile circumference is a small price to pay. And the cost of having two such relays circling at 10,000 every 250 miles to get another 250 miles radius is also rather small.
Do you think pilots are going to notice a 5 ms latency at 500 miles? (me neither)
I like the cloud for some things. But i also like it if a device which has more memory than i need for all my personal documents (including 10000 Photos) is used wise enough not to require 24x7 online access.
That's a matter of personal preference.
capped data is the expression of a physical reality vs. a marketing tool used to push users quickly into freshly build networks without investing in the sw and forcing them to new phones.
Capped data is a joke. It's a movement towards charging per-unit prices for a service that has no meaningful per-unit cost. Sure, it costs money to build a network, blah blah blah. But there is no fixed cost for moving data around. A Gbit switch costs about as much as a 100 Mbit switch did a few years back, and moves 100x as much data in a unit of time as the 100 Mbit one. It uses about the same amount of electricity, regardless of how much data is being moved.
Where did that per-unit cost go?
Because of this, I figure it's only a matter of time before this whole "cap the user" nonsense goes away.
Back when socialism wasn't a four letter word, it was commonplace for the US govt. to fund development in areas that were deemed strategic. Boeing benefit from billions in exploratory funding, as did Lockheed, Honeywell, and too many other companies to name.
Cisco, for example, exists because of federal research dollars into Darpa Net.
The US would do well to remember how we used to do it before China learned from us!
It may well be that MS is making this decision for self-fulfilling reasons, EG to protect Silverlight in the marketplace. But with IE continuing to lose market share year after year (from its high of about 90%, it's under half nowadays with nary an uptick in sight) one has to ask if they can afford to, once again, be "the big guy who couldn't".
For the past two years, we've simply told our clients that, to use our system, they had to be running Firefox or Chrome, and that we didn't support IE - it simply couldn't do what we needed and we found that having the features is more important to our clients than having compatibility. They *will* switch if they need to, if you provide features they need.
Finally, with IE 9, we may consider supporting it this upcoming fiscal year. Now, in this market place, if I developed software that needed or used 3D effects in a browser, I wouldn't hesitate to drop IE support for even a second. Microsoft doesn't control the game, anymore. This may be their version of IBM's PS/2 Micro-channel debacle.
(For those who don't remember, IBM created the "PC-compatible" marketplace and thought they ruled the roost. They decided to come out with an incompatible schema for hardware called the Microchannel bus which offered numerous technical advantages over the industry standard ISA bus, which failed miserably because nobody else wanted to license the tech)
If you're choosing the language to write your app based on how it performs, you are likely the one making bad technical decisions.
Depending on the job, there are numerous things to consider, and for some things, performance is one of them. For example, if you want to write a graphics intense video game, performance is most definitely a primary factor.
Ok, it sounds like a suit-bot saying, but it actually has a real, specific meaning.
Companies should figure out what it is that makes their company unique and specific, and spend money making sure that core competency is maintained. Anything else is a candidate for outsourcing.
On a personal level, we do it all the time! I am, at best, a modest mechanic - anything much more complicated than a battery or alternator stymies me - so I have my car worked on by my mechanic. I'm not much of a doctor, anything beyond "diabetes" or "hematoma" makes me dizzy, so when there are medical questions, I consult my doctor.
But I don't hire people to fix my computer - that falls well within my lines as a tech professional, so my routers are set up sensibly and my personal computer and important data is all redundantly backed up.
If a company makes computers, it's a dumb move to outsource making computers, because that's giving up their core competence and their company ceases to have a reason for existing!
But a car manufacturing company could do well reducing costs by outsourcing their email to a 3rd party vendor, so long as adequate SLAs exist.
Ask yourself: how many average, ordinary people back up their own computer? Do you really want that average, ordinary guy (who never backs up) in charge of keeping your medical records safe and backed up?
Even with all the risks that outsourcing represents, in many (most?) cases, the data is safer with a 3rd party.
Google provides automatic redirection to the unwashed masses based on geolocation and adds a very simple way to get around it for those who don't want it. Just Google it.
Slashdot Mirror
Perhaps it would be a good idea to start by defining what exactly "cloud computing" means?
Because looking at the Wikipedia article I see only a brief mention of reliability through redundancy: Reliability is improved if multiple redundant sites are used, which makes well-designed cloud computing suitable for business continuity and disaster recovery.
As CTO of a data hosting "cloud services" provider myself, I'm proud of our track record for reliability and redundancy. All our systems are backed up offsite to multiple locations every 24 hours. At all times we provide off-site hosting facilities at about 50% of the capacity of our primary production cluster for disaster recovery scenarios. All our systems are redundant on site; the loss of any single system can be healed within an hour or so. Our system uptime averages approaches 4 nines over years in an industry where 2 nines is considered exceptional.
Having systems replicate in near-real-time to multiple locations and autonomically route around large network outages on extremely short notice is an extremely tough and expensive thing to do. It sounds simple enough, but the devil is in the details. The number of things that can go wrong is simply staggering, and trying to account for every possible thing that can go wrong is simply infeasible to proactively account for.
What do you do when the problem is due to a router outage offsite? What if 70% of routes work, and 30% don't due to a BGP burp? What if latency is 50ms? 100ms? 200ms? At what point is it "down" ? What if network conditions are excellent but there's a problem with DNS? What if a power surge generated by your UPS (yes, it happens!) takes out 20% of your production cluster? Do you heal or switch to hot backup? What if it takes out 30% of your production cluster? 40% 60%?
People think of "online" is a boolean yes/no question, but it's not. We calculate our uptimes based on expected end-user experience, but rarely is our production cluster actually working at 100%. Nearly always, some system or other is in need of attention and it doesn't constitute an emergency because there's still additional redundancy in the system.
Hell, even working for low wage, a person is expensive. Thus the most effort should be put in having them do the least effort.
Yeah, but here's where it gets weird: Software is highly valuable because it allows a single person to do the same task a very large number of times.
Because of this, it's not like stacking logs or gluing tile. If you have sufficient leverage, you can spend a stupid amount of money getting everything just right and profit immensely from it. (See: Apple)
On the other hand, most software isn't leveraged nearly as much. Most software is "commodity" and doesn't represent a huge leverage of anybody's skill. And that's why, in some environments, egging out the last ms of performance out of an algorithm is critical, and in others it's a total waste of money.
The important part is to determine just how important the possible performance increase actually might be.
I sooooo wish this were true! The problem is in concentrated wealth.
If I (a "little people") crash a car, the most anybody could get out of me would be my life savings, which (at 40) adds up to a few hundred thousand. Enough for an ambulance chaser and a douchebag to make my life suck, but not enough to bring out the big guns.
But when the "driver" of a car is a software company with millions of installs, any crash at all is enough for said ambulance chaser and douchebag to go for the jugular for millions. Ambulance chaser might get the backing of the big, tall, class-action corporate lawyer type of legal firm and turn this into a real law suit!
Even though it's safer. It's not what's sensible, it's what's legal.
A true geek has no life to speak of on Friday, anyway!
It's pretty easy to put together a basic security system. Require an identity token of some sort, and require proof of knowledge of a secret, and you have the makings of a security system!
Security is not a boolean. Security is a variable, ranging from non at all to mild, moderate, to extremely secure.
Little things can greatly add greatly to real security (such as free permits for concealed weapons and password strength requirements), and big, obvious, "secure" things can easily be nothing more than theater. (EG: the TSA goons at the airports)
To be truly secure at the high end is surprisingly difficult. As the value of the prize increases in value, the number of potentially useful attacks increases exponentially. A dollar-store lock will reasonably protect a $50 used bike in most areas, but at $500, the lock has to be able to reasonably defend itself from something like a grinder. At $5,000, blow torches become reasonable, and at $50,000, plastic explosives are a fair bet.
See how much more difficult it gets to defend concentrated wealth? It's *hard* to do it right!
Windows server looked after by a good sysadmin == secure.
Mac server looked after by bad sysadmin == insecure.
The sad part is that much (most?) of being a good sysadmin consists of ensuring that you install security updates regularly. I've been close enough to embarrassing hacks on several servers to know what happened, and all (but one!) have been hacked as a result of a poor update policy. (The last one was due to a weak root password + passwordAuthentication enabled on ssh)
For all my own systems, I demand a strong, default-deny firewall, and (most importantly!) regular, frequent updates. Tools like yum make this easy.
Yes, things like strictly firewalling your systems, moving services only used internally to nonstandard ports, disabling (or never installing) unneeded services, etc. are good ideas and I strongly recommend them all.
But first and foremost, make sure your system is up to date!
That's what a Smart Grid does -- people think it will let them pay spot prices for electricity, no no no, it's about the power company collecting Google-style metrics of power consumption on an appliance-by-appliance, outlet-by-outlet basis, and then giving you 10% off your bill if you consent to having a remote power cutoff installed on your washing machine, air conditioner, and car charger.
I'm not aware of there being a definition for what is a "smart grid". It's still evolving, and many people have different ideas about what "smart" grid should be.
All TFA says is that it's easily possible to build a dynamic system that's unstable. Duh. In aviation, the attribute that you are looking for is called "positive dynamic stability" - which simply means that when things get interrupted or jarred, that the system actively works to mitigate the change and stabilize over time.
This is also something I achieve with a self-healing load-balanced computing cluster. Things happen, loads spike, etc. The system should gradually self-stabilize.
This is a problem as old as the presence dynamic systems.
The only problem with drones is knowing that they're there. They're inevitable, and why I watch the rollout of ADS-B with interest.
That's why I ALWAYS am on approach control in Echo airspace and up. Here in North Cali, Beale AFB has "temporary" flight restrictions going at least half the time for UAVs in the area, and the only real restriction is that you have to have a VFR flight following or be under IFR.
Having learned on the "steam guage panel" and now flying with a Garmin 496, I can declare with confidence that GPS technologies combined with good visual display and information technologies make a night vs day difference in flying!
I suspect that drones will simply be required to be ADS-B compliant, and that pretty much everybody will have the receivers installed on their aircraft.
If such a transmitter was located on you house, unless you had an antenna that ensured the radiation was not directed inward the building would be quite uninhabitable.
-citation needed
This guy is either rich, or she has Asperger's. Else, how else could the conversation go?
He: "I got it! The perfect wedding! We get married by a computer!"
She: "Should we go with blue or white?"
He: "But, the computer would marry us!"
She: "Should we go with lacy, or silky?"
He: "The computer should run Linux!"
She: "I think I like lacy more than silky..."
He: "I could even have it running LISP!"
She: "You want to have a preacher with a lisp? What the HELL are you thinking!?"
99.9998% of the power emitted by my pocket cell phone is wasted. Only the minute fraction of those waves that happen to coincide with the line between myself and the cell tower are actually converted into anything useful. Sadly, the 99.998% ratio is probably optimistic, it's likely to be considerably worse than that!
So here we are, covering the Earth with radio-emitting devices by the billions, (with a "B") with emissions ranging from a few milliwatts per device up into the millions. Thanks to the inverse-square law, the milli-watt cordless phone on your desk may contribute far more power than the 5 million watt TV transmitter a few miles away, but in all cases, there is a considerable amount of radio flux virtually ANYWHERE people are likely to reside.
Using ANY of it will cost (on average) less than 0.001% of available signal. This is about as green as it gets!
Remember that movie a few years back, "Zoro" with Antonio Banderas? (If not, it's probably up on Netflix Instant Play or a torrent someplace)
The movie was all about the "passing of the guard" - a new, younger man taking the role of "Zoro", the anonymous masked crime fighter of the previous generation. It's a good movie, so I recommend it highly. But it also does a passable job of showing the difference between an identity and an idea.
I'm guessing that there are, in fact, a half dozen or more actual people who have had the identity of "Topiary". They may have shared the pseudo-identity concurrently, so, who did what?
I'm getting out the popcorn and getting ready to watch the show!
Sure, some solutions are faster than MySQL out of the box by skipping much of the language parsing and stuff that any SQL solution has to do. But that's not to say that they are actually more efficient at key retrieval.
For example, one developer found that the best no-sql solution was.... MySQL, which excels at simple key retrieval. He was able to best MemCached by a factor of almost 2.
Use the right tool for the job.
"Always have a proper backup! You can learn more about backups here""
Sorry, but this is the height of naive. CentOS casts a very *VERY* long shadow!
Do you want a nice, supported, "Enterprise" Linux but don't have much budget to spend? Guess what: CentOS is almost your only choice. SL is nice, but it's not binary compatible with RHEL. Mix and match a few packages with a few "EL5/6" repos and you very quickly will run into binary hell.
Other than RHEL, what "Enterprise" options are available? What you need is something that is stable, conservative in anything that changes the environment un-necessarily, and is likely to support 3rd party software from vendors.
What else is there? There's Oracle's re-branded RHEL, there's RHEL itself, and there's CentOS. Debian is almost good enough.
While there are plenty of Linux distros, there are surprisingly few other options available at the commercial/enterprise tier.
OP has very good advice. Take it. Understanding what negative and positive pressure can be very important in circumstances like this.
Ever wonder why dust collects so badly inside your PC? It's a negative pressure environment - the main power fan at the back blows hot air OUT of the computer, causing air to rush into every crack and orifice in your case, making your expensive electronics into a poor quality air filter. The dust collected is a byproduct of this fact.
I once was called in to deal with a computer in a very dusty environment. (they raised pets) Their computer required extremely frequent cleaning and despite this, they had numerous hardware failures. The CD ROM drive was pretty much always useless no matter how often it was replaced. Using a medium-sized box, a cheap 8" fan, and HEPA air filters and lots of duct tape, I made a large, low-pressure air filter that blew large amounts of HEPA purified air into the computer, creating a positive pressure inside the case of clean, filtered air. A year later, the computer had only traces of dust and was working perfectly, including the CD ROM drive!
In PPs example, you want to create a negative pressure environment to keep dust from getting OUT.
When a traffic light is broken, you are supposed to treat it as a stop sign. If it won't sense your bike, isn't it effectively broken? I've no idea how that would hold up in traffic court, but as an avid cyclist, I've acted on this for years without incident.
Especially this early in the life cycle of this "cloud" crap. Any expectation of not loosing your data if you don't keep a backup yourself is entirely your own fault.
Early? Come again? I've been providing "cloud computing" services for about a decade. Is 10 years early?
Of course, the names have changed. There was "Service Oriented Architecture" and before that it was "Software as a Service" and then before that "managed application hosting" and various other names for it over time. Name doesn't matter, it's still the same thing.
Backups, of the redundant and routinely verified type, have always been an integral of the solutions I've worked to provide. I have an archival record of nightly snapshots going back for *years*.
I'm going to make a dare: Go talk to "little people" whose livelyhood rests squarely on the state of their computers. I'm talking about small office clerks, college professors, gas station managers, single-practice lawyers, authors... even most schools and school district offices. Get them to describe how they back up their stuff (of course they do backups!) and then ask them to SHOW you their most recent backup. Excepting large organizations and/or wealthy firms, if you have a large enough sample size, you'll rarely find more than 10% or so of everybody has backups of their critical data made within the past week.
Reputable cloud providers make a concerted effort to provide reasonable backups of their data - it is their business, after all - and just like a mechanic is more likely to do a decent job wrenching a car than an average Joe, a professional hosting company is more likely to do a decent job hosting and the average Joe.
Of course, there tend to be lots of hosting professional types around /., so it's probably true that even a professional firm wouldn't out perform you.
Carriers? That's just silly.
The latency for a radio link of 250 miles is in the order of a few milliseconds, and 250 miles is enough range to make finding a semi-mobile "home base" somewhere between tough and impossible.
As a pilot, I routinely hear radio calls 250 miles away when flying 10,000 feet; having a radio-relay essentially circling near the home base at 10,000 feet to support drone activity for a 500 mile circumference is a small price to pay. And the cost of having two such relays circling at 10,000 every 250 miles to get another 250 miles radius is also rather small.
Do you think pilots are going to notice a 5 ms latency at 500 miles? (me neither)
Still missing the point! Yes, it costs money for equipment and upgrades. Money that I pay per month, amortized over the useful life of the equipment.
None of which changes the point - there is no meaningful cost per unit of data transferred!
I like the cloud for some things. But i also like it if a device which has more memory than i need for all my personal documents (including 10000 Photos) is used wise enough not to require 24x7 online access.
That's a matter of personal preference.
capped data is the expression of a physical reality vs. a marketing tool used to push users quickly into freshly build networks without investing in the sw and forcing them to new phones.
Capped data is a joke. It's a movement towards charging per-unit prices for a service that has no meaningful per-unit cost. Sure, it costs money to build a network, blah blah blah. But there is no fixed cost for moving data around. A Gbit switch costs about as much as a 100 Mbit switch did a few years back, and moves 100x as much data in a unit of time as the 100 Mbit one. It uses about the same amount of electricity, regardless of how much data is being moved.
Where did that per-unit cost go?
Because of this, I figure it's only a matter of time before this whole "cap the user" nonsense goes away.
Back when socialism wasn't a four letter word, it was commonplace for the US govt. to fund development in areas that were deemed strategic. Boeing benefit from billions in exploratory funding, as did Lockheed, Honeywell, and too many other companies to name.
Cisco, for example, exists because of federal research dollars into Darpa Net.
The US would do well to remember how we used to do it before China learned from us!
It may well be that MS is making this decision for self-fulfilling reasons, EG to protect Silverlight in the marketplace. But with IE continuing to lose market share year after year (from its high of about 90%, it's under half nowadays with nary an uptick in sight) one has to ask if they can afford to, once again, be "the big guy who couldn't".
For the past two years, we've simply told our clients that, to use our system, they had to be running Firefox or Chrome, and that we didn't support IE - it simply couldn't do what we needed and we found that having the features is more important to our clients than having compatibility. They *will* switch if they need to, if you provide features they need.
Finally, with IE 9, we may consider supporting it this upcoming fiscal year. Now, in this market place, if I developed software that needed or used 3D effects in a browser, I wouldn't hesitate to drop IE support for even a second. Microsoft doesn't control the game, anymore. This may be their version of IBM's PS/2 Micro-channel debacle.
(For those who don't remember, IBM created the "PC-compatible" marketplace and thought they ruled the roost. They decided to come out with an incompatible schema for hardware called the Microchannel bus which offered numerous technical advantages over the industry standard ISA bus, which failed miserably because nobody else wanted to license the tech)
If you're choosing the language to write your app based on how it performs, you are likely the one making bad technical decisions.
Depending on the job, there are numerous things to consider, and for some things, performance is one of them. For example, if you want to write a graphics intense video game, performance is most definitely a primary factor.
Ok, it sounds like a suit-bot saying, but it actually has a real, specific meaning.
Companies should figure out what it is that makes their company unique and specific, and spend money making sure that core competency is maintained. Anything else is a candidate for outsourcing.
On a personal level, we do it all the time! I am, at best, a modest mechanic - anything much more complicated than a battery or alternator stymies me - so I have my car worked on by my mechanic. I'm not much of a doctor, anything beyond "diabetes" or "hematoma" makes me dizzy, so when there are medical questions, I consult my doctor.
But I don't hire people to fix my computer - that falls well within my lines as a tech professional, so my routers are set up sensibly and my personal computer and important data is all redundantly backed up.
If a company makes computers, it's a dumb move to outsource making computers, because that's giving up their core competence and their company ceases to have a reason for existing!
But a car manufacturing company could do well reducing costs by outsourcing their email to a 3rd party vendor, so long as adequate SLAs exist.
Ask yourself: how many average, ordinary people back up their own computer? Do you really want that average, ordinary guy (who never backs up) in charge of keeping your medical records safe and backed up?
Even with all the risks that outsourcing represents, in many (most?) cases, the data is safer with a 3rd party.
Google provides automatic redirection to the unwashed masses based on geolocation and adds a very simple way to get around it for those who don't want it. Just Google it.
I can't, you insensitive clod - it's in SPANISH!