... and the replication systems are typically not worth much more than a dime, sadly.
We have a pretty beefy set up; 4x 16 Core Xeon DB servers with 128 GB of RAM each and Enterprise SSDs, serving hundreds of instances of like-schema databases, one per (organizational) customer, serving an aggregate peak of about 1,000 queries/second in a mixed read/write load.
And we've never been able to get replication to work reliably, ever. In every case we've ever tried, we've seen a net reduction in reliability. Every single time. Not that we've stopped trying, it has just never reached "just works" territory.
Venus has a mass comparable to Earth's. Mars' mass is 1/10th Earth's. While it's nice to put hard numbers to it, the reason why Mars' atmosphere got blown away isn't hard to understand.
I've been dealing with metabolic syndrome for years, and so far, my blood sugar remains in normal range, weight, cholesterol, etc. is normal, though I do still take some pills to reduce hypertension. I started with The Diabetes Diet by Dr. Bernstein which laid out the relationship between sugar, blood sugar, and diabetes decades ago. Bernstein is literally the guy who changed the treatment of diabetes in the 1970s and at least doubled the life expectancy of diabetics.
If I keep my diet to simple meats and vegetables, I feel far better, sustaining much higher energy and work performance levels, even as my blood sugars stay down (A1C of 6.0) and "all the numbers get better".
Starch, simple sugars and saturated fats are just death. Just stay away. Granted, that means that you can't eat at least half of what the grocery store sells, but are those deep fried starch crackers really all that great?
So let's do away with this "post scarcity" nonsense.
For all intents and purposes, we already live in a "post scarcity" world. Even homeless bums and mentally ill in the first world do not starve to death. I spend less than 30 minutes per day earning the food I eat - everything else pays for other stuff. And yet, scarcities exist, and money-based economy is still chugging away. I want the latest indie song. My clients need me to write a program to help solve a regulatory issue. Etc.
Depending on the corporate structure, you doom your career with the company if you ask for such orders in writing.
Note at all:
From: PropellerHead@vwtech.com To: SupremeManager@vwtech.com Subject: Smog modification clarification
Mr Smegbert,
I just wanted to clarify your verbal request to disable the backfeed loop on the emissions detector; did you want that to happen all the time or only automatically when the engine was not in drive? The possibility exists also to make this a button that the driver could push.
Thanks.
Smarty McSmartpants Sr Propeller Head Engineer guy
5 years ago, it seemed that BTRFS was rapidly getting there, and its inclusion into the kernel made it feel like a rather sure bet!
(crickets)
5 years later, BTRFS is still "rapidly" getting there. I've tried it numerous times and had horrible data loss events literally every single time, and this as recently as a month ago.
Meanwhile, we're using ZFS on Linux in a complex production environment in a worst-case mixed read/write use case and it's been absolutely rock solid bullet proof, demonstrably more stable than EXT4. Yes. More stable than EXT4. And this while bring so many incredible features to the administration table! Until you've lived with snapshots, replication, clones, pools, zvols, extendable pools, and dynamic resource allocation, it's like trying to explain Monet to a blind person.
I sincerely hope that ZFS finally becomes a first class citizen in the Linux community.
Regardless of whether QNX is superior or not technically, it no longer matters. They've lost because people want to use what is popular (and has apps), and Android and iOS are it.
I'm just glad that, a few years ago, when Windows/OSX ruled the roost, that the hairy hippies didn't say this about Linux. We can crow now, that Linux is installed on more devices than any other kernel or O/S, but Linux wasn't always such a sure bet.
Diversity is good. I welcome it. I'm hoping they digest the Android ecosystem and learn to use it to strengthen QNX.
I have a precision M3800 and love it! Lightweight, decent battery life, gorgeous 4K screen, wickedly fast i7 processor, dual HD ports, (one mSATA) HDMI support...
All of which makes it a beautiful laptop, but add to that native Linux support... I'm a Fedora fan so I bought with windows and dual boot. It "just works" with a Fedora install.
Your confidence is proof of your inexperience. Data... dies. Sorry, that's just the truth.
If you've ever tried to do a data recovery on years-old data, whether it's audio tape, film, HDD, flash, CD/DVD rips, whatever. They all have an error rate that increases over time.
The only way to preserve data long term is to actively manage it. Keep redundant copies. Use error correcting code to identify data errors and correct them. Media must be periodically re-read and written to ensure "freshness". Non-digital data must be redundantly copied in line with its utility, analog data should be digitized to minimize generation loss.
We maintain a large ZFS file store. We scrub everything weekly, a process that does all the above to proactively identify small data errors and fix them before they become big, unrecoverable ones. We store all our data in redundant storage pools, and replicate constantly.
This, or a process like this, is what's required to keep data squeaky clean, secure, and accurate.
What you're describing for "unlimited" is what would be termed in a data center "unmetered". If I buy a 100 Mbit unmetered pipe, I can do exactly as you say, max out the 100 Mbit pipe 24x7 as I please.
What customers really want, most likely, is something like a "burstable" connection with reasonable limits. Let's say I buy a 100 Mbit "burstable" connection with a 10 Mbit commit. That means I can use up to 100 Mbits at any moment, but if the average is over 10 Mbit I pay more. (It's actually not average, it's 95th percentile, but we'll call it "average" for this conversation)
So there are limits! Fine. I'd happily go for an agreement that
1) states an average data rate,
2) Allows me to burst up to 4x or 5x that rate,
3) Throttles later in the month to maintain the average data rate or less.
4) As technology advances so that bits are cheaper/faster to send my average data rate climbs, or monthly price drops
I think the problem isn't with 1, 2, or 3, but with #4 It's much cheaper to send a GB of data now than it was 3-5 years ago. Why hasn't my usage cap gone up, or my monthly price dropped? Until that question is answered, all we're dealing with are lies and spin.
I'm advising everyone to install Linux from now on, this crap is not worth it, not even for free.
If you're this late in the game and *finally* saying this, well, welcome to the club!
I switched almost 20 years ago to Linux, when my Windows 98 computer emailed a word file of customer names and (private) contact info with a virus. Realizing the risk of staying with an insecure platform, I jumped to using Linux for my workstation full time.
I've never looked back.
RedHat Linux became Fedora/RHEL/CentOS but picking the "main" commercial distro at the time has paid enormous dividends over the years! In the intervening years, I went from newbie to experienced software developer, with pay scale to match. Security has been excellent; the constant plague of malware and virus updates are a long distant memory.
This while serving thousands of users at hundreds of clients 24x7.
Yes, I still Windows - for games. And that is dwindling.
If they had local wifi and a small server with last year's hit movies on demand for $1, they'd have a plane full of happy customers. To watch a movie, you wouldn't get access to the Internet - that would still cost $5 just like now, since the $5 customers are corporate.
No, it isn't. It's an attempt at a shunk down, big-box PC. You know, the boring beige boxes that nobody buys any more? I see no way that this saves money over time. The branding is in software, which this doesn't fix. See: Cyanogenmod which works with many already existing phones. It's highly impractical, expensive, and architecturally prone to failure, as you have a mobile, device commonly subjected to strong impacts, which is exactly when you don't want removable, (flimsy) interlocking pieces.
> I'm not going to buy a phone until I can get something like this, and I don't really care if it's made by google or someone else.
There's no hard, fast answer, although it would probably be popular around here to assume that the right place is with the Tech dept. This is certainly supportable; I've seen plenty of clueless administrators blinded by blinking lights and flashy fluff make architecturally very poor choices!
At work, we are a vertical stack cloud-based software vendor. We work with hundreds of clients and deliver a very excellent product that saves our clients $$$. Several times now, I've seen IT departments that have ballooned into inefficient "candy stores" for developers who are mostly intent on increasing their take of the organization's $$. It mostly happens because the managers at our client organizations aren't techies in any sense of the word, so they take whatever techno mumbo jumbo blurted out by the techies as gospel.
When the powers that be at the organization bring us in, and ask the tech department, they are almost universally ice cold to the idea of working with us, as their job is potentially on the line. Change = BAD! And so we see a fight while the corrupt IT department and the management duke it out. We've lost a few, we've won most. In any case, we often come in as little as 1/5 the cost of the bloated, internal IT department's offerings, while offering better service, better security, and strongly worded privacy and availability clauses.
So there isn't a right answer, you know? Some CxOs are clueless or corrupt. Some IT departments are similarly incompetent or corrupt. It all really comes down to "people are people".
There is only a difference of semantics between the following two statements:
1) Conserving resources during an emergency by strongly discouraging the waste of a suddenly valuable commodity.
2) Taking advantage of an emergency by gouging customers in need of a suddenly valuable commodity.
There is literally no difference in practice between the two, the difference is intent of the seller, the actions could, quite literally be exactly the same. If we can use greed to make a bad situation better, shouldn't we?
I note that you are talking about the sound of the plane(s) at approach speeds, not the hypersonic speeds for which the Concorde is unique.
The issue wasn't that the Concorde was loud during take offs and landings, the issue was that the Concorde was ridiculously noisy at altitude, flying at 3,000 MPH or better.
You never experienced the Concorde at full speed. And that's why the Concorde wasn't economically successful.
We had a problem that ext* just couldn't handle. We have a medium sized filesystem with about 250 million data files that we needed to back up. Every day. Rsync completely failed at the job, taking between 1 and 2 days to do the job.
Desperate to find a solution, we tried ZFS and snapshot replication. Our time to replicate to DR, dropped from days to a few hours, backup storage requirements dropped through the floor, and server load dropped at the same time! This is on a reasonably priced set of systems, Xeon-based intel systems with just 32 GB of RAM and 6x 4 TB drives.
ZFS is pretty decent, and has proven to be more reliable for our use than ext*. However, its licensing presents a developmental pit fall. On Linux, it won't ever be a "first rate citizen" even though the ZoL project has done a great job making it very available. ZFS also has a number of pretty terrible problems:
1) You can't remove a vdev from a ZFS pool without destroying the pool.
2) You can't upgrade a vdev's redundancy level once you've added it to a pool.
This means that, if you're careful, ZFS is wonderful. But it's easy to make a mistake that you can't easily back out of. See the section hating your data to see what I mean.
BTRFS has been "only a few years away now" for quite a few years now. I'm not convinced it will ever reach production ready status. Apparently it has some architectural problems that have been criticized pretty soundly. I'm no longer convinced about the future inevitability of BTRFS.
I sincerely hope that BCacheFS really delivers on these promises, I'd love it!
There have been a *lot* of smart people in the history of Computer Science over a very *long* period of time, and the best of the best of their innovations we now call "classic solutions". Solutions like SQL, POSIX, etc.
It has become popular to decide that such solutions are "antiquated" in the face of some new "great thing".
Remember NoSQL? Well, yeah. There actually *is* a very small set of problems for relating data not best served by SQL. But even those cases often collapse into something best left to the tried and true "stodgy" technology of SQL and a competent admin.
Object storage is another way of saying "REST API roolz doodz!" and while REST is a fantastic technology for integrating disparate product stacks, it's hardly a replacement for a proper, local, filesystem.
Give it some time! We need to let the AI mature like a fine wine, and filter down into consumer devices.
The thing is, that consumer devices don't, themselves, *need* to have AI in them at all.
Try using Google Maps on your phone without an Internet connection. It's dead, Jim! Try using Siri without an Internet connection. Nope. Try using voice-to-text on your phone without a network connection. Bzzzzt!
AI doesn't need to be on your phone to be useful. As AI is developed, it'll be hosted in massive server farms (a la Watson) and time sliced for consumers. And even though we think AI will turn up in "high end" uses before it becomes a consumer item, the reality is that the economics of meeting consumer needs is just so incredibly lucrative as long as you can hit the scale that it's just as likely to be a consumer commodity before it's helping doctors diagnose brain diseases.
We have an office in a small, very remote town. Until recently, we got 100 Mb Internet though a WISP by installing a microwave tower. Recently, the local power company installed fiber optic Internet, so now we have 100 Mb Internet at the same price, without packet loss!
Meanwhile, in our home town, Comcast recently announced support for 2 Gbit Internet for $350/month, and the entire area is blanketed with 4G LTE through ATT/Verizon/TMo.
I climbed a remote mountain. I facebooked the pix I took at the peak before hiking down.
What is this "darkness beyond the 'burbs" you speak of?
Slashdot Mirror
... and the replication systems are typically not worth much more than a dime, sadly.
We have a pretty beefy set up; 4x 16 Core Xeon DB servers with 128 GB of RAM each and Enterprise SSDs, serving hundreds of instances of like-schema databases, one per (organizational) customer, serving an aggregate peak of about 1,000 queries/second in a mixed read/write load.
And we've never been able to get replication to work reliably, ever. In every case we've ever tried, we've seen a net reduction in reliability. Every single time. Not that we've stopped trying, it has just never reached "just works" territory.
Replication is PG's Achilles's heel.
To be fair, I copy/paste all the time on my Android 4.4.x phone all the time. It is a little awkward, but it works fine.
Venus has a mass comparable to Earth's. Mars' mass is 1/10th Earth's. While it's nice to put hard numbers to it, the reason why Mars' atmosphere got blown away isn't hard to understand.
I have a 2.5 year old phone that I otherwise love and while it's EOL, I still use it extensively.
The idea that a phone can be not even 3 years old and not have any hope of getting updates is something I balk STRONGLY at.
I've been dealing with metabolic syndrome for years, and so far, my blood sugar remains in normal range, weight, cholesterol, etc. is normal, though I do still take some pills to reduce hypertension. I started with The Diabetes Diet by Dr. Bernstein which laid out the relationship between sugar, blood sugar, and diabetes decades ago. Bernstein is literally the guy who changed the treatment of diabetes in the 1970s and at least doubled the life expectancy of diabetics.
If I keep my diet to simple meats and vegetables, I feel far better, sustaining much higher energy and work performance levels, even as my blood sugars stay down (A1C of 6.0) and "all the numbers get better".
Starch, simple sugars and saturated fats are just death. Just stay away. Granted, that means that you can't eat at least half of what the grocery store sells, but are those deep fried starch crackers really all that great?
Even the Star Trek world is forced to admit that scarcities exist. For example, if they truly were "post scarcity", why would the Enterprise have to negotiate for vaccine on Ligon II?
So let's do away with this "post scarcity" nonsense.
For all intents and purposes, we already live in a "post scarcity" world. Even homeless bums and mentally ill in the first world do not starve to death. I spend less than 30 minutes per day earning the food I eat - everything else pays for other stuff. And yet, scarcities exist, and money-based economy is still chugging away. I want the latest indie song. My clients need me to write a program to help solve a regulatory issue. Etc.
Depending on the corporate structure, you doom your career with the company if you ask for such orders in writing.
Note at all:
From: PropellerHead@vwtech.com
To: SupremeManager@vwtech.com
Subject: Smog modification clarification
Mr Smegbert,
I just wanted to clarify your verbal request to disable the backfeed loop on the emissions detector; did you want that to happen all the time or only automatically when the engine was not in drive? The possibility exists also to make this a button that the driver could push.
Thanks.
Smarty McSmartpants
Sr Propeller Head Engineer guy
Well, the idea is solid, but there's so much revisionist history here....
Linux isn't a fork, it's a rewrite. FreeBSD was in no way derived from Win 85. MS-DOS wasn't a fork of CP/M, it was a hackish clone.
And those are the parts I'm somewhat familiar with...
5 years ago, it seemed that BTRFS was rapidly getting there, and its inclusion into the kernel made it feel like a rather sure bet!
(crickets)
5 years later, BTRFS is still "rapidly" getting there. I've tried it numerous times and had horrible data loss events literally every single time, and this as recently as a month ago.
Meanwhile, we're using ZFS on Linux in a complex production environment in a worst-case mixed read/write use case and it's been absolutely rock solid bullet proof, demonstrably more stable than EXT4. Yes. More stable than EXT4. And this while bring so many incredible features to the administration table! Until you've lived with snapshots, replication, clones, pools, zvols, extendable pools, and dynamic resource allocation, it's like trying to explain Monet to a blind person.
I sincerely hope that ZFS finally becomes a first class citizen in the Linux community.
Regardless of whether QNX is superior or not technically, it no longer matters. They've lost because people want to use what is popular (and has apps), and Android and iOS are it.
I'm just glad that, a few years ago, when Windows/OSX ruled the roost, that the hairy hippies didn't say this about Linux. We can crow now, that Linux is installed on more devices than any other kernel or O/S, but Linux wasn't always such a sure bet.
Diversity is good. I welcome it. I'm hoping they digest the Android ecosystem and learn to use it to strengthen QNX.
I have a precision M3800 and love it! Lightweight, decent battery life, gorgeous 4K screen, wickedly fast i7 processor, dual HD ports, (one mSATA) HDMI support...
All of which makes it a beautiful laptop, but add to that native Linux support... I'm a Fedora fan so I bought with windows and dual boot. It "just works" with a Fedora install.
Your confidence is proof of your inexperience. Data... dies. Sorry, that's just the truth.
If you've ever tried to do a data recovery on years-old data, whether it's audio tape, film, HDD, flash, CD/DVD rips, whatever. They all have an error rate that increases over time.
The only way to preserve data long term is to actively manage it. Keep redundant copies. Use error correcting code to identify data errors and correct them. Media must be periodically re-read and written to ensure "freshness". Non-digital data must be redundantly copied in line with its utility, analog data should be digitized to minimize generation loss.
We maintain a large ZFS file store. We scrub everything weekly, a process that does all the above to proactively identify small data errors and fix them before they become big, unrecoverable ones. We store all our data in redundant storage pools, and replicate constantly.
This, or a process like this, is what's required to keep data squeaky clean, secure, and accurate.
What you're describing for "unlimited" is what would be termed in a data center "unmetered". If I buy a 100 Mbit unmetered pipe, I can do exactly as you say, max out the 100 Mbit pipe 24x7 as I please.
What customers really want, most likely, is something like a "burstable" connection with reasonable limits. Let's say I buy a 100 Mbit "burstable" connection with a 10 Mbit commit. That means I can use up to 100 Mbits at any moment, but if the average is over 10 Mbit I pay more. (It's actually not average, it's 95th percentile, but we'll call it "average" for this conversation)
So there are limits! Fine. I'd happily go for an agreement that
1) states an average data rate,
2) Allows me to burst up to 4x or 5x that rate,
3) Throttles later in the month to maintain the average data rate or less.
4) As technology advances so that bits are cheaper/faster to send my average data rate climbs, or monthly price drops
I think the problem isn't with 1, 2, or 3, but with #4 It's much cheaper to send a GB of data now than it was 3-5 years ago. Why hasn't my usage cap gone up, or my monthly price dropped? Until that question is answered, all we're dealing with are lies and spin.
I'm advising everyone to install Linux from now on, this crap is not worth it, not even for free.
If you're this late in the game and *finally* saying this, well, welcome to the club!
I switched almost 20 years ago to Linux, when my Windows 98 computer emailed a word file of customer names and (private) contact info with a virus. Realizing the risk of staying with an insecure platform, I jumped to using Linux for my workstation full time.
I've never looked back.
RedHat Linux became Fedora/RHEL/CentOS but picking the "main" commercial distro at the time has paid enormous dividends over the years! In the intervening years, I went from newbie to experienced software developer, with pay scale to match. Security has been excellent; the constant plague of malware and virus updates are a long distant memory.
This while serving thousands of users at hundreds of clients 24x7.
Yes, I still Windows - for games. And that is dwindling.
If they had local wifi and a small server with last year's hit movies on demand for $1, they'd have a plane full of happy customers. To watch a movie, you wouldn't get access to the Internet - that would still cost $5 just like now, since the $5 customers are corporate.
This is grownup LEGO.
No, it isn't. It's an attempt at a shunk down, big-box PC. You know, the boring beige boxes that nobody buys any more? I see no way that this saves money over time. The branding is in software, which this doesn't fix. See: Cyanogenmod which works with many already existing phones. It's highly impractical, expensive, and architecturally prone to failure, as you have a mobile, device commonly subjected to strong impacts, which is exactly when you don't want removable, (flimsy) interlocking pieces.
> I'm not going to buy a phone until I can get something like this, and I don't really care if it's made by google or someone else.
You're gonna be waiting a long time. Sorry.
There's no hard, fast answer, although it would probably be popular around here to assume that the right place is with the Tech dept. This is certainly supportable; I've seen plenty of clueless administrators blinded by blinking lights and flashy fluff make architecturally very poor choices!
At work, we are a vertical stack cloud-based software vendor. We work with hundreds of clients and deliver a very excellent product that saves our clients $$$. Several times now, I've seen IT departments that have ballooned into inefficient "candy stores" for developers who are mostly intent on increasing their take of the organization's $$. It mostly happens because the managers at our client organizations aren't techies in any sense of the word, so they take whatever techno mumbo jumbo blurted out by the techies as gospel.
When the powers that be at the organization bring us in, and ask the tech department, they are almost universally ice cold to the idea of working with us, as their job is potentially on the line. Change = BAD! And so we see a fight while the corrupt IT department and the management duke it out. We've lost a few, we've won most. In any case, we often come in as little as 1/5 the cost of the bloated, internal IT department's offerings, while offering better service, better security, and strongly worded privacy and availability clauses.
So there isn't a right answer, you know? Some CxOs are clueless or corrupt. Some IT departments are similarly incompetent or corrupt. It all really comes down to "people are people".
There is only a difference of semantics between the following two statements:
1) Conserving resources during an emergency by strongly discouraging the waste of a suddenly valuable commodity.
2) Taking advantage of an emergency by gouging customers in need of a suddenly valuable commodity.
There is literally no difference in practice between the two, the difference is intent of the seller, the actions could, quite literally be exactly the same. If we can use greed to make a bad situation better, shouldn't we?
This is the foundation of economic theory, and it rarely works out well to ignore economics altogether.
I note that you are talking about the sound of the plane(s) at approach speeds, not the hypersonic speeds for which the Concorde is unique.
The issue wasn't that the Concorde was loud during take offs and landings, the issue was that the Concorde was ridiculously noisy at altitude, flying at 3,000 MPH or better.
You never experienced the Concorde at full speed. And that's why the Concorde wasn't economically successful.
Disclaimer: I ZFS.
We had a problem that ext* just couldn't handle. We have a medium sized filesystem with about 250 million data files that we needed to back up. Every day. Rsync completely failed at the job, taking between 1 and 2 days to do the job.
Desperate to find a solution, we tried ZFS and snapshot replication. Our time to replicate to DR, dropped from days to a few hours, backup storage requirements dropped through the floor, and server load dropped at the same time! This is on a reasonably priced set of systems, Xeon-based intel systems with just 32 GB of RAM and 6x 4 TB drives.
ZFS is pretty decent, and has proven to be more reliable for our use than ext*. However, its licensing presents a developmental pit fall. On Linux, it won't ever be a "first rate citizen" even though the ZoL project has done a great job making it very available. ZFS also has a number of pretty terrible problems:
1) You can't remove a vdev from a ZFS pool without destroying the pool.
2) You can't upgrade a vdev's redundancy level once you've added it to a pool.
This means that, if you're careful, ZFS is wonderful. But it's easy to make a mistake that you can't easily back out of. See the section hating your data to see what I mean.
BTRFS has been "only a few years away now" for quite a few years now. I'm not convinced it will ever reach production ready status. Apparently it has some architectural problems that have been criticized pretty soundly. I'm no longer convinced about the future inevitability of BTRFS.
I sincerely hope that BCacheFS really delivers on these promises, I'd love it!
+1 If I had mod points you'd get one!
There have been a *lot* of smart people in the history of Computer Science over a very *long* period of time, and the best of the best of their innovations we now call "classic solutions". Solutions like SQL, POSIX, etc.
It has become popular to decide that such solutions are "antiquated" in the face of some new "great thing".
Remember NoSQL? Well, yeah. There actually *is* a very small set of problems for relating data not best served by SQL. But even those cases often collapse into something best left to the tried and true "stodgy" technology of SQL and a competent admin.
Object storage is another way of saying "REST API roolz doodz!" and while REST is a fantastic technology for integrating disparate product stacks, it's hardly a replacement for a proper, local, filesystem.
Running the fiber optic cables could get rather expensive though...
At my work, one of the products we offer is to archive documents. We have over 250 million documents that were never printed.
Give it some time! We need to let the AI mature like a fine wine, and filter down into consumer devices.
The thing is, that consumer devices don't, themselves, *need* to have AI in them at all.
Try using Google Maps on your phone without an Internet connection. It's dead, Jim! Try using Siri without an Internet connection. Nope. Try using voice-to-text on your phone without a network connection. Bzzzzt!
AI doesn't need to be on your phone to be useful. As AI is developed, it'll be hosted in massive server farms (a la Watson) and time sliced for consumers. And even though we think AI will turn up in "high end" uses before it becomes a consumer item, the reality is that the economics of meeting consumer needs is just so incredibly lucrative as long as you can hit the scale that it's just as likely to be a consumer commodity before it's helping doctors diagnose brain diseases.
We have an office in a small, very remote town. Until recently, we got 100 Mb Internet though a WISP by installing a microwave tower. Recently, the local power company installed fiber optic Internet, so now we have 100 Mb Internet at the same price, without packet loss!
Meanwhile, in our home town, Comcast recently announced support for 2 Gbit Internet for $350/month, and the entire area is blanketed with 4G LTE through ATT/Verizon/TMo.
I climbed a remote mountain. I facebooked the pix I took at the peak before hiking down.
What is this "darkness beyond the 'burbs" you speak of?