Right now, today, I have a P3 desktop running CentOS6/32 as a network monitor. It's old as the hills. My phone beats it handily in performance. But it runs on about 15 watts, and does the job so reliably that, in 10 years, it's never skipped a beat. It started with the original RedHat 6.1. (before RHEL was a thing)
I know it won't actually make any records, but I'm sure it's one of the oldest 0.01%, maybe even 0.001% of computers in terms of durability. It would be a remarkable machine if it wasn't otherwise so unremarkable.
It seems absurd to have lameness filters that seem to specifically target code on a site that caters to the coder types! And the 4 minute limit is just silly. Back when it was still publicly shown, I had karma out the yin yang. I'm sure I still do, even though it's no longer displayed in any form that I can tell.
Slashdot trumps Reddit for quality of articles, Reddit bests Slashdot for UI and comment participation, though posting on Reddit has become such a land mine I don't bother unless I'm on a very small/exclusive subreddit)
The spinning disk era is coming to a close, and I welcome it! The issue is that while storage capacity has, for decades, increased almost exponentially, the actual performance of the HDD has remained virtually flat. A typical HDD spinning at about 7200 RPM can store 4 TB of data or more, but can only serve about 150 seek operations per second. Physics, she is a bitch, you know? So while you might be able to store 500 million files, it takes a month to copy them.
Everywhere I look, Enterprise or "performant" storage has moved to SSDs. We moved our DB servers 4 (5?) years ago to SSDs and saw at least a 95% reduction in query times involving disk operations. (EG: not cached) For us, spinning rust is the new tape; SSDs (or RAM) anywhere performance matters, and spinning rust for archival use. For our session cache, we use a RAM drive.
So, your program allocates some memory. Should it initialize the memory to make sure it's all a bunch of zeros? Apparently, Nvidia doesn't think so.
So, a program running on your OS requests some memory. Should the OS initialize the memory before handing it to the application? Apparently, Apple doesn't think so.
* The destruction of the Firefox web browser thanks to numerous fucking idiotic changes being forced on its users by Mozilla.
... Moves designed to protect users from MalWare. Of course that's "destruction"...
* The destruction of Linux as a viable OS, especially when used on servers, all thanks to systemd being forced by all of the major distros.
I use Linux extensively. SystemD has been a very minor speed bump. All the people screaming and crying about systemD haven't been able to down out the simple fact that SystemD works just fine and carries numerous benefits.
I personally don't care enough about the other issues to comment.
Sorry, but with silly results like this, I have to ask why such a small article so vapid of meaningful content was posted on Slashdot. Shouldn't paid shill articles be a different color or something?
No mention was given as to how this ranking was accomplished, and the list given at the bottom of the article doesn't even match the headline (where 2 and 3 are MySQL and MS SQL Server, and Microsoft Access beats Cassandra.
We *have* a Samsung 4k "Smart" TV and we don't use any of the smarts - at all. What drives it is the XB One, or the Android TV Stick. I'm not even sure how to use the "smart" part. But it said "Smart" on the box...
I have no doubt that the "smart" feature is something they added to make it more appealing in some way, but why?
Years ago, with a then-ubiquitous Moto flip phone at my hip, I "invented" what I called the "Urban Commando Phone" - the cell phone already had a clock, why not add things like flashlight, garage door remote, TV Universal Remote, etc. so that instead of having dozens of devices, you had one to "rule them all".
I had no idea, at the time, of the types of convergence that would come in the form of the smart phone, which has all of these and many more either available built in or easily available.
The term "ubiquitous computing" has been used for decades, and when I first heard the term, it was to convey the still-radical idea that every home would have a computer.
Computing seems to come in stages or "generations", where each previous generation generally powers or enables the next one. Mainframes became infrastructure for Mini computers, which (eventually) gave way to PCs, which then merged with Mini computers in the Internet revolution, which then gave birth to the smart phone era.
Following this trend, the "next big thing" will use Cell phones, PCs, and Servers to extend their capability. And this is already happening. My cell phone has a small cluster of devices that surround it that it interfaces with: Bluetooth headset, folding mobile keyboard, smart watch, etc. We are just beginning to innovate with these standards-based technologies to develop the true "IoT" that is coming; the things that link to our mobile phones to enable things we haven't begun to imagine.
Some examples that I've seen/heard of include all manner of medical devices: insulin pumps that use software on your smart phone to adjust or recommend insulin administration. Devices that provide the ability to test for common diseases "in the field" inexpensively, serving field medics and impoverished areas alike. Payment systems that use our mobile phones and networks to augment or replace credit cards.
And on and on. As always, the game is just beginning!
We switched from rsync to ZFS replication for our production environments and the difference in performance is rather extreme. (and why we made this change)
Medium sized file system, 12 TB and a few hundred million files. Doing a backup with rsync took days, and it was all just tied up in IOPs, even if the number of files changed was rather small. At this scale, it takes more than 24 hours just to get a listing of files.
Switching to ZFS with nighly snapshots and replication dropped backup times from days down to minutes. Add other features like clones, compression, hot error checking (scrub), hot swapping and RAIDZ, and it becomes pretty obvious pretty quickly that if you're serious about data you should seriously consider ZFS.
ZFS on Linux is pretty easy to install and it's been rock solid stable of our use in a 24x7 heavy use environment.
Netflix already offers "CDN" like capability in the form of caching servers that any ISP meeting a volume of viewership can negotiate to have. It's like a Squid proxy for Netflix, and it's pretty straightforward.
1) You have gobs of users on your network. 2) You negotiate for a NF content server. 3) You install a big server (or servers, depending on load) that caches the most commonly viewed shows. It automatically updates as demands change. 4) ???? 5) Profit from sharply reduced upstream or peering point bandwidth bills.
The only reason that Comcast doesn't do this is that they offer their own content and *want* a degraded network for NF users, or NF to pay $$ to Comcast, and they want that degraded performance in a way that isn't blatantly their fault.
Source: I'm a techie with friends who work at a regional ISP.
We spend big bucks at a tier 1 data center to host our data. Our application cluster is fully redundant, and we serve large amounts of high-value data. (think: business intelligence and workflow information, not video streams)
We have a DR cluster, a back-stop for when all else has failed. Although it is a redundant copy of our production cluster, it itself is, by design, non-redundant. Where our production cluster has at least two physical machines (and often more) providing any specific service, the DR cluster has just one.
It's there as insurance against the rare, once-in-a-lifetime event that has been known to shutter companies. I've experienced one such event when the data center I was hosting at went dark after they went insolvent - it took weeks to get all services resolved, because, while I had backups, they were onsite. (doh!)
While I've never since had backups that weren't automatic, daily, and off-site, I see no reason to demand the same power redundancy for the DR cluster that I do at the primary data center. Cost is more of a factor.
I think I can answer your question, as Mr. comfortable PHP guy's story isn't so different from my own.
1) PHP isn't the point. Solving a need is. Find a need and devise a solution. Bonus points if there are no other solutions on the market. 2) Make a solution that works better/cheaper than anything else that compares. May or may not involve PHP/ Python/ Go/ Lua/ Javascript/ Assembler/your dog. 3) Sell your service to people who need it. Manage it carefully, make sure it really works and keeps getting better. 4).... ????? (too many little things to mention) 5) Profit!
GP didn't "get rich quick", and neither did I. I spent years pounding the keyboard and the pavement while my solutions matured and got traction. I picked up some partners along the way, as I quickly found that I needed somebody to manage the books, negotiate contracts, run marketing/sales/support, etc.
Growth rate since the very beginning has been between 20% to 70% per year - solid, organic growth but not VC "start up" growth. We've been solid, stable, and profitable every single year and this year is no exception.
I have a server that I use as a personal media server, and backups. I bought it at a yard sale for $10. Big box, lots of cheap, added several high capacity hard drives, performance is strictly irrelevant, and the several-generations-ago AMD Athlon 64 (remember those?) gamer board supports the 4GB of ECC DDR(1) RAM that is probably overkill for the need. I have no doubt that I could get at least another 5 years out of this ancient hardware for the need and be perfectly happy with it.
Not having requirements specified in a "is this sufficient?" question is a bit like asking "is this jacket warm enough?" without specifying where you're going to go in it.
Dumb ideas that are cheap persist. That is, until there's a watershed event that puts all the stupid into sharp relief. We haven't had such an incident for IoT; give it time.
Thanks to movies and TV, people think that encryption is something you "bypass" by letting somebody who looks nerdly typing furiously in front of 3 or 4 screens in an office with lots of glass and neon lights. When it's exploited by thugs who downloaded an exploit and stole their stuff by using their security system to verify that they weren't home, the word will start to spread.
Starving people are dramatically more likely to revolt than well fed people. Somehow, mentioning this ridiculously obvious fact is universally dismissed.
If the author hasn't been played in any way, then the damage is still done: the scammers just got a great idea they'll no doubt literally capitalize on.
If you think that anybody who's written or executed ransomware hasn't already thought about ransoming medical devices, you have an astonishingly low opinion of others. Just how smart do you think you are?
Anybody who's spent the time necessary to write ransomware and attempt to profit from it has had more than enough time to consider the all reasonable possibilities, even if it took somebody as *brilliant* as you 5 minutes to come up with this idea. This isn't some global super-conspiracy; this is as brilliant as banging chips off a rock with another rock.
HOWEVER, -all- of the "download.php" scripts I've ever looked at have at least two of the same three vulnerabilities.
1) Protection from directory transversal is harder than it looks,
2) fopen_url, and
3) memory depletion from failing to disable the output buffer before reading and writing chunks of the file.
I'm a PHP dev, and the first two are relatively straightforward to prevent. EG: Check that basename($file) == realpath(Basename($file)) kind of stuff. But #3 is interesting to me; how would the following cause any problem?
Slashdot Mirror
I have little doubt that gmail SPAM filtering is Postini SPAM filtering with a few tweaks.
Right now, today, I have a P3 desktop running CentOS6/32 as a network monitor. It's old as the hills. My phone beats it handily in performance. But it runs on about 15 watts, and does the job so reliably that, in 10 years, it's never skipped a beat. It started with the original RedHat 6.1. (before RHEL was a thing)
I know it won't actually make any records, but I'm sure it's one of the oldest 0.01%, maybe even 0.001% of computers in terms of durability. It would be a remarkable machine if it wasn't otherwise so unremarkable.
Wish I had mod points! So, I'll agree, blah blah.
It seems absurd to have lameness filters that seem to specifically target code on a site that caters to the coder types! And the 4 minute limit is just silly. Back when it was still publicly shown, I had karma out the yin yang. I'm sure I still do, even though it's no longer displayed in any form that I can tell.
Slashdot trumps Reddit for quality of articles, Reddit bests Slashdot for UI and comment participation, though posting on Reddit has become such a land mine I don't bother unless I'm on a very small/exclusive subreddit)
We use ffmpeg to process video files uploaded by customers. We'll be patching our app first thing in the morning. This is a big deal for us.
The spinning disk era is coming to a close, and I welcome it! The issue is that while storage capacity has, for decades, increased almost exponentially, the actual performance of the HDD has remained virtually flat. A typical HDD spinning at about 7200 RPM can store 4 TB of data or more, but can only serve about 150 seek operations per second. Physics, she is a bitch, you know? So while you might be able to store 500 million files, it takes a month to copy them.
Everywhere I look, Enterprise or "performant" storage has moved to SSDs. We moved our DB servers 4 (5?) years ago to SSDs and saw at least a 95% reduction in query times involving disk operations. (EG: not cached) For us, spinning rust is the new tape; SSDs (or RAM) anywhere performance matters, and spinning rust for archival use. For our session cache, we use a RAM drive.
So, your program allocates some memory. Should it initialize the memory to make sure it's all a bunch of zeros? Apparently, Nvidia doesn't think so.
So, a program running on your OS requests some memory. Should the OS initialize the memory before handing it to the application? Apparently, Apple doesn't think so.
Either answer is right.
* The destruction of the Firefox web browser thanks to numerous fucking idiotic changes being forced on its users by Mozilla.
... Moves designed to protect users from MalWare. Of course that's "destruction"...
* The destruction of Linux as a viable OS, especially when used on servers, all thanks to systemd being forced by all of the major distros.
I use Linux extensively. SystemD has been a very minor speed bump. All the people screaming and crying about systemD haven't been able to down out the simple fact that SystemD works just fine and carries numerous benefits.
I personally don't care enough about the other issues to comment.
Sorry, but with silly results like this, I have to ask why such a small article so vapid of meaningful content was posted on Slashdot. Shouldn't paid shill articles be a different color or something?
No mention was given as to how this ranking was accomplished, and the list given at the bottom of the article doesn't even match the headline (where 2 and 3 are MySQL and MS SQL Server, and Microsoft Access beats Cassandra.
Any DB ranking that puts Access in as a top contender should definitely back up their claims - extraordinary claims require extraordinary evidence!
i bought a couple of TV sticks at Amazon for about $40 and they work amazingly well. Netflix, Hulu, Google Play, the works! Browses smoothly as well, works as a general purpose super-tablet, works with normal KB/Mouse without issue...
Why would anyone would spend ten times the cash?
We *have* a Samsung 4k "Smart" TV and we don't use any of the smarts - at all. What drives it is the XB One, or the Android TV Stick. I'm not even sure how to use the "smart" part. But it said "Smart" on the box...
I have no doubt that the "smart" feature is something they added to make it more appealing in some way, but why?
The fun part is having to hack the registry so that you get inundated with "free upgrade to 10!" notices all day long...
Years ago, with a then-ubiquitous Moto flip phone at my hip, I "invented" what I called the "Urban Commando Phone" - the cell phone already had a clock, why not add things like flashlight, garage door remote, TV Universal Remote, etc. so that instead of having dozens of devices, you had one to "rule them all".
I had no idea, at the time, of the types of convergence that would come in the form of the smart phone, which has all of these and many more either available built in
or easily available.
The term "ubiquitous computing" has been used for decades, and when I first heard the term, it was to convey the still-radical idea that every home would have a computer.
Computing seems to come in stages or "generations", where each previous generation generally powers or enables the next one. Mainframes became infrastructure for Mini computers, which (eventually) gave way to PCs, which then merged with Mini computers in the Internet revolution, which then gave birth to the smart phone era.
Following this trend, the "next big thing" will use Cell phones, PCs, and Servers to extend their capability. And this is already happening. My cell phone has a small cluster of devices that surround it that it interfaces with: Bluetooth headset, folding mobile keyboard, smart watch, etc. We are just beginning to innovate with these standards-based technologies to develop the true "IoT" that is coming; the things that link to our mobile phones to enable things we haven't begun to imagine.
Some examples that I've seen/heard of include all manner of medical devices: insulin pumps that use software on your smart phone to adjust or recommend insulin administration. Devices that provide the ability to test for common diseases "in the field" inexpensively, serving field medics and impoverished areas alike. Payment systems that use our mobile phones and networks to augment or replace credit cards.
And on and on. As always, the game is just beginning!
Well, sort of....
We switched from rsync to ZFS replication for our production environments and the difference in performance is rather extreme. (and why we made this change)
Medium sized file system, 12 TB and a few hundred million files. Doing a backup with rsync took days, and it was all just tied up in IOPs, even if the number of files changed was rather small. At this scale, it takes more than 24 hours just to get a listing of files.
Switching to ZFS with nighly snapshots and replication dropped backup times from days down to minutes. Add other features like clones, compression, hot error checking (scrub), hot swapping and RAIDZ, and it becomes pretty obvious pretty quickly that if you're serious about data you should seriously consider ZFS.
ZFS on Linux is pretty easy to install and it's been rock solid stable of our use in a 24x7 heavy use environment.
Yes, of course. There are three. Two are paid by Microsoft.
I've been following the development of the Transition with the hope of having one. Right as the Transition looked like it was actually going to be a thing that not only flew but might actually be sold at some point, they announced this turd bomb of the TF-X with it's magical properties.
As soon as I saw the announcement, I was like WTF? From the description I wondered if it was going to be made of unicorn farts or faerie dust...
My guess is that their flying car design was scooped by the far better looking and more ready AeroMobil and they panicked.
Netflix already offers "CDN" like capability in the form of caching servers that any ISP meeting a volume of viewership can negotiate to have. It's like a Squid proxy for Netflix, and it's pretty straightforward.
1) You have gobs of users on your network.
2) You negotiate for a NF content server.
3) You install a big server (or servers, depending on load) that caches the most commonly viewed shows. It automatically updates as demands change.
4) ????
5) Profit from sharply reduced upstream or peering point bandwidth bills.
The only reason that Comcast doesn't do this is that they offer their own content and *want* a degraded network for NF users, or NF to pay $$ to Comcast, and they want that degraded performance in a way that isn't blatantly their fault.
Source: I'm a techie with friends who work at a regional ISP.
We spend big bucks at a tier 1 data center to host our data. Our application cluster is fully redundant, and we serve large amounts of high-value data. (think: business intelligence and workflow information, not video streams)
We have a DR cluster, a back-stop for when all else has failed. Although it is a redundant copy of our production cluster, it itself is, by design, non-redundant. Where our production cluster has at least two physical machines (and often more) providing any specific service, the DR cluster has just one.
It's there as insurance against the rare, once-in-a-lifetime event that has been known to shutter companies. I've experienced one such event when the data center I was hosting at went dark after they went insolvent - it took weeks to get all services resolved, because, while I had backups, they were onsite. (doh!)
While I've never since had backups that weren't automatic, daily, and off-site, I see no reason to demand the same power redundancy for the DR cluster that I do at the primary data center. Cost is more of a factor.
I think I can answer your question, as Mr. comfortable PHP guy's story isn't so different from my own.
1) PHP isn't the point. Solving a need is. Find a need and devise a solution. Bonus points if there are no other solutions on the market. /your dog. .... ????? (too many little things to mention)
2) Make a solution that works better/cheaper than anything else that compares. May or may not involve PHP/ Python/ Go/ Lua/ Javascript/ Assembler
3) Sell your service to people who need it. Manage it carefully, make sure it really works and keeps getting better.
4)
5) Profit!
GP didn't "get rich quick", and neither did I. I spent years pounding the keyboard and the pavement while my solutions matured and got traction. I picked up some partners along the way, as I quickly found that I needed somebody to manage the books, negotiate contracts, run marketing/sales/support, etc.
Growth rate since the very beginning has been between 20% to 70% per year - solid, organic growth but not VC "start up" growth. We've been solid, stable, and profitable every single year and this year is no exception.
Boy, this is spot on!
I have a server that I use as a personal media server, and backups. I bought it at a yard sale for $10. Big box, lots of cheap, added several high capacity hard drives, performance is strictly irrelevant, and the several-generations-ago AMD Athlon 64 (remember those?) gamer board supports the 4GB of ECC DDR(1) RAM that is probably overkill for the need. I have no doubt that I could get at least another 5 years out of this ancient hardware for the need and be perfectly happy with it.
Not having requirements specified in a "is this sufficient?" question is a bit like asking "is this jacket warm enough?" without specifying where you're going to go in it.
Dumb ideas that are cheap persist. That is, until there's a watershed event that puts all the stupid into sharp relief. We haven't had such an incident for IoT; give it time.
Thanks to movies and TV, people think that encryption is something you "bypass" by letting somebody who looks nerdly typing furiously in front of 3 or 4 screens in an office with lots of glass and neon lights. When it's exploited by thugs who downloaded an exploit and stole their stuff by using their security system to verify that they weren't home, the word will start to spread.
Stop being a pedantic Aspie.
"Aspie" is not even a word. There is no such word. Seriously, what does that even mean?
I will likely be downvoted, even though what I write is absolutely true.
Revolution was predicted at least 6 years ago, a result of public land policy changes made 50 years ago and yet nobody talks about it. In fact, if anybody brings it up, they are immediately dismissed as radical, or simply silly.
Starving people are dramatically more likely to revolt than well fed people. Somehow, mentioning this ridiculously obvious fact is universally dismissed.
If the author hasn't been played in any way, then the damage is still done: the scammers just got a great idea they'll no doubt literally capitalize on.
If you think that anybody who's written or executed ransomware hasn't already thought about ransoming medical devices, you have an astonishingly low opinion of others. Just how smart do you think you are?
Anybody who's spent the time necessary to write ransomware and attempt to profit from it has had more than enough time to consider the all reasonable possibilities, even if it took somebody as *brilliant* as you 5 minutes to come up with this idea. This isn't some global super-conspiracy; this is as brilliant as banging chips off a rock with another rock.
Please call them by their proper name: Daesh. Calling them the "Islamic State" of anything grants credibility to a minute, small fraction of the population.
HOWEVER, -all- of the "download.php" scripts I've ever looked at have at least two of the same three vulnerabilities.
1) Protection from directory transversal is harder than it looks,
2) fopen_url, and
3) memory depletion from failing to disable the output buffer before reading and writing chunks of the file.
I'm a PHP dev, and the first two are relatively straightforward to prevent. EG: Check that basename($file) == realpath(Basename($file)) kind of stuff. But #3 is interesting to me; how would the following cause any problem?
$fp = fopen($hugefile, 'r');
while ($line = fgets($fp, 1024))
echo $line;
In this case, the buffered output will be spooled to Apache/end user as it fills. Or did you mean OOM errors from trying to load a 2 GB file into RAM?