Gigabit Ethernet maxes out at between 70 and 100 Megabytes per second, depending on your file sharing protocol. When Gb-E was first introduced this was faster than local disk, so it meant that workstations could get data to and from the server faster than they could from local storage. This was a good thing.
Now that even a cheap 1TB consumer hard drive (not to mention a SSD) can push more than double this data transfer rate, working off a server is (relative to local storage) getting slower and slower.
There was a good progression from 10 Base-T to 100 Base-T to 1000 Base-T and the adoption curve was pretty constant. 10 Gb-E has been around for a long time now (10 Gb-E was ratified in 2002 over fibre and 2006 for twisted-pair copper cabling) but it's adoption has been slow. Even Apple who have historically been an early adopter of faster network technologies haven't put 10 Gb-E in anything, even the Mac Pro.
2.5 Gb Base-T is a nice stop-gap measure, it more than doubles the speed whilst retaining a lot of (and being backwards compatible with) the existing infrastructure and 5 Gb Base-T will be a nice improvement over that for workstations and other data intensive tasks, without the expense of going all the way to 10 Gb.
A single 2.5 Gb port will have more throughput (and be easier to configure) than a pair of bonded 1 Gb ports. This is going to be great in the majority of the environments in which I work, where workstations have been feeling the squeeze of "slow" 1 Gb links for some time now, leading people to do things like work off their desktop instead of off the server (meaning no backups and no-one else can share their work)
Seeing as you can already push 10Gb-E down a Cat6 or 6e cable, with regular RJ45 plugs on it, I'd say that they're keeping 2.5Gb-E and 5Gb-E backwards compatible and using Cat5e (for short runs) Cat6 (recommended) or 6e (more better) with RJ45 termination https://en.wikipedia.org/wiki/...
Google Code Project Hosting offered a free collaborative development environment for open source projects. In 2016 the service was shut down, see this post for more info. Projects hosted on Google Code remain available in the Google Code Archive.
We’re trying to make great products for people, and we have at least the courage of our convictions to say we don’t think this is part of what makes a great product, we’re going to leave it out. Some people are going to not like that, they’re going to call us names [...] but we’re going to take the heat [and] instead focus our energy on these technologies which we think are in their ascendancy and we think are going to be the right technologies for customers. And you know what? They’re paying us to make those choices. [...] If we succeed, they’ll buy them, and if we don’t, they won’t, and it’ll all work itself out.
Now, guess what they're talking about? Removing the headphone jack from the latest iPhone? No, this is Steve Jobs from 2010 talking about having the courage to not support Flash on iOS. Many, many people were up in arms at the time, however history has shown us all that whilst this was a tough decision to make, we're all now better off because of it.
And, forcing everyone to use bluetooth, so they can sell more Beats bluetooth headphones, is exactly why there's no way to plug regular headphones into the iPhone 7. This is why they don't even include a free adapter in the box with each and every single iPhone.
According to Wired - http://www.wired.com/2016/08/a... Apple are buying iPhones, made in China, by an Irish subsidiary and then selling them from their Irish subsidiary to US distributors, so the profits on the sale of the iPhones is booked through Ireland - although Wired don't explicitly say that it's US distributors, it could be their EU disties instead...
Yes, but if by some magic process of having my brother, who is a foreign resident, send me an invoice for "services rendered" or "brand licensing", and I paid this invoice and claimed back all of the the tax I'd otherwise need to pay on the expense, when all my brother is doing is holding the cash on my behalf in a low-tax jurisdiction, then we'd have a situation akin to what's going on with Apple, Google et. al.
Oh, you mean the unique Advertising Identifier that's in every iOS device? Yes, that one that you can reset at any time? https://support.apple.com/en-a...
I'm guessing that the throughput is slow enough that sending a command to the spacecraft takes 20 seconds of tx time - e.g. it might take 10 bytes to send a command and are getting a data rate of 4 bits per second.
Latency isn't as much of an issue in this case, as once they send the wake up command, they can have the other commands in flight on their way to the satellite, but it's going to stop listening and do something they don't want it to do ~2 minutes after it gets the wakeup command - likely due to a fault with a sleep timer or similar.
Yes, I know, that's why I tried to ping him. Throughput isn't too bad once the connection is established (sending a 256 GB MicroSD card taped to the bird's leg) but the latency is a killer!
I was talking in broad hypothetical strokes there (in other words, making stuff up to illustrate a point), and you're right, an old ISA card will be trouble, you might be able to find a specialised board with regular PCI or PCI-X with a bit of looking.
Yep, I remember having to write a TSR myself that would count from 1 to 100,000 or something like that at every tick of an interrupt just so Prince of Persia would play at an acceptable speed on my 386 DX2 66 when it was optimised for a 286.
Have an old legacy proprietary DOS program that you need for your business? Then throw it into/dev/null, and hire some talented programmers to write a modern free open-source replacement for GNU/Linux, that will get published on github.
Why, that just sounds like an absolutely wonderful idea. Why would anyone insist keeping on using some old software that has been paid for many moons ago, and we all by now know exactly where and when it does and doesn't work because it's been doing the same task for 20 years? Why not instead pay thousands and thousands of dollars for someone to attempt to write a replacement for it, possibly reverse-engineering a proprietary and undocumented hardware interface (costing thousands and thousands of dollars more more in time) only to give it all away to the handful of other people on the planet who also use the same version of WHATEVER.EXE that I'm using?
No, not many things use actual EEPROMs these days - they're expensive and not (easily) field reprogrammable - most devices use flash to store their initial OS "ROM" and subsequent updates simply reflashes new ROM image to the flash.
Modern drives will silently remap sectors without telling you (unless you look at the SMART status). Once they exhaust their pool of spare sectors, then they start telling things higher up the chain that there are bad sectors. By the time a disk is reporting bad sectors to the OS (as a bad sector, instead of incrementing a SMART counter and silently carrying on) it has remapped so many bad sectors that it can no longer automatically remap them and is now telling you there is a problem.
In my experience, every single drive that I've seen reporting even a single bad sector will soon go pear-shaped and shouldn't be used.
Wow, that's a pretty intense workload! Yes, it's very difficult to get SLC SSDs these days, everyone has gone for MLC, even the Intel DC grade units...
Yeah, but then you can't accurately know the cost of de-duplication, additionally you're doing work against already committed files which is a big no-no if you want stable storage. If I commit a file, I don't want a background process to read/write it and a software bug to screw it up years down the road.
Additionally, you're taking away resources from a system that will already be taxed. My file server has a load of 1.2-2.5 on an average day (because I'm running against the IOPS limits on my 5-year old SSD's), doing ANYTHING (even streaming a backup) has to be meticulously planned so as not to affect the system.
Wow, what are you doing on your server that you're thrashing your SSDs with 500+ IOPS 24/7?
Slashdot Mirror
Because shutting down a site that has controversial and illegal (and borderline-illegal) material has always been a good way to stop the bad behavior?
No, it's not a good way to stop bad behaviour, but if said bad behaviour causes the site to implode, then why should we help them stay afloat?
Gigabit Ethernet maxes out at between 70 and 100 Megabytes per second, depending on your file sharing protocol. When Gb-E was first introduced this was faster than local disk, so it meant that workstations could get data to and from the server faster than they could from local storage. This was a good thing.
Now that even a cheap 1TB consumer hard drive (not to mention a SSD) can push more than double this data transfer rate, working off a server is (relative to local storage) getting slower and slower.
There was a good progression from 10 Base-T to 100 Base-T to 1000 Base-T and the adoption curve was pretty constant. 10 Gb-E has been around for a long time now (10 Gb-E was ratified in 2002 over fibre and 2006 for twisted-pair copper cabling) but it's adoption has been slow. Even Apple who have historically been an early adopter of faster network technologies haven't put 10 Gb-E in anything, even the Mac Pro.
2.5 Gb Base-T is a nice stop-gap measure, it more than doubles the speed whilst retaining a lot of (and being backwards compatible with) the existing infrastructure and 5 Gb Base-T will be a nice improvement over that for workstations and other data intensive tasks, without the expense of going all the way to 10 Gb.
A single 2.5 Gb port will have more throughput (and be easier to configure) than a pair of bonded 1 Gb ports. This is going to be great in the majority of the environments in which I work, where workstations have been feeling the squeeze of "slow" 1 Gb links for some time now, leading people to do things like work off their desktop instead of off the server (meaning no backups and no-one else can share their work)
Yes, RJ-45 jacks and plugs can already handle 10 Gb-E transmissions:
https://en.wikipedia.org/wiki/...
Yes, an RJ45 plug and jack can already do 10Gb-E
https://en.wikipedia.org/wiki/...
Seeing as you can already push 10Gb-E down a Cat6 or 6e cable, with regular RJ45 plugs on it, I'd say that they're keeping 2.5Gb-E and 5Gb-E backwards compatible and using Cat5e (for short runs) Cat6 (recommended) or 6e (more better) with RJ45 termination
https://en.wikipedia.org/wiki/...
Here's some inspiration for you: http://27bslash6.com/halogen.h...
On their own hosting? You mean Google Code?
Here's a direct quote from Apple:
Now, guess what they're talking about? Removing the headphone jack from the latest iPhone? No, this is Steve Jobs from 2010 talking about having the courage to not support Flash on iOS. Many, many people were up in arms at the time, however history has shown us all that whilst this was a tough decision to make, we're all now better off because of it.
And, forcing everyone to use bluetooth, so they can sell more Beats bluetooth headphones, is exactly why there's no way to plug regular headphones into the iPhone 7. This is why they don't even include a free adapter in the box with each and every single iPhone.
Oh, hang on a minute...
According to Wired - http://www.wired.com/2016/08/a...
Apple are buying iPhones, made in China, by an Irish subsidiary and then selling them from their Irish subsidiary to US distributors, so the profits on the sale of the iPhones is booked through Ireland - although Wired don't explicitly say that it's US distributors, it could be their EU disties instead...
Yes, but if by some magic process of having my brother, who is a foreign resident, send me an invoice for "services rendered" or "brand licensing", and I paid this invoice and claimed back all of the the tax I'd otherwise need to pay on the expense, when all my brother is doing is holding the cash on my behalf in a low-tax jurisdiction, then we'd have a situation akin to what's going on with Apple, Google et. al.
Oh, you mean the unique Advertising Identifier that's in every iOS device? Yes, that one that you can reset at any time?
https://support.apple.com/en-a...
No, the article specifically says that after it boots they have approximately 2 minutes before it goes into a fault condition
I'm guessing that the throughput is slow enough that sending a command to the spacecraft takes 20 seconds of tx time - e.g. it might take 10 bytes to send a command and are getting a data rate of 4 bits per second.
Latency isn't as much of an issue in this case, as once they send the wake up command, they can have the other commands in flight on their way to the satellite, but it's going to stop listening and do something they don't want it to do ~2 minutes after it gets the wakeup command - likely due to a fault with a sleep timer or similar.
Yes, I know, that's why I tried to ping him. Throughput isn't too bad once the connection is established (sending a 256 GB MicroSD card taped to the bird's leg) but the latency is a killer!
[user@localhost ~]$ ping yvan256.amish.org
PING yvan256.amish.org (144.131.380.158): 56 data bytes
64 bytes from 144.131.380.158: icmp_seq=0 ttl=59 time=14.368 hrs
64 bytes from 144.131.380.158: icmp_seq=1 ttl=59 time=11.156 hrs
64 bytes from 144.131.380.158: icmp_seq=2 ttl=59 time=12.062 hrs
64 bytes from 144.131.380.158: icmp_seq=3 ttl=59 time=11.772 hrs
64 bytes from 144.131.380.158: icmp_seq=4 ttl=59 time=11.867 hrs
^C
I was talking in broad hypothetical strokes there (in other words, making stuff up to illustrate a point), and you're right, an old ISA card will be trouble, you might be able to find a specialised board with regular PCI or PCI-X with a bit of looking.
Or, you know, use FreeDOS to run it on modern hardware - which is kind-of the point of the article.
Yep, I remember having to write a TSR myself that would count from 1 to 100,000 or something like that at every tick of an interrupt just so Prince of Persia would play at an acceptable speed on my 386 DX2 66 when it was optimised for a 286.
Have an old legacy proprietary DOS program that you need for your business? Then throw it into /dev/null, and hire some talented programmers to write a modern free open-source replacement for GNU/Linux, that will get published on github.
Why, that just sounds like an absolutely wonderful idea. Why would anyone insist keeping on using some old software that has been paid for many moons ago, and we all by now know exactly where and when it does and doesn't work because it's been doing the same task for 20 years? Why not instead pay thousands and thousands of dollars for someone to attempt to write a replacement for it, possibly reverse-engineering a proprietary and undocumented hardware interface (costing thousands and thousands of dollars more more in time) only to give it all away to the handful of other people on the planet who also use the same version of WHATEVER.EXE that I'm using?
Yep, what a smashing idea!
No, not many things use actual EEPROMs these days - they're expensive and not (easily) field reprogrammable - most devices use flash to store their initial OS "ROM" and subsequent updates simply reflashes new ROM image to the flash.
Does Windows RT have The Windows Subsystem for Linux (WSL)?
If so (and I assume not, but haven't looked) then you can run native Debian binaries right from CMD.EXE
Modern drives will silently remap sectors without telling you (unless you look at the SMART status).
Once they exhaust their pool of spare sectors, then they start telling things higher up the chain that there are bad sectors.
By the time a disk is reporting bad sectors to the OS (as a bad sector, instead of incrementing a SMART counter and silently carrying on) it has remapped so many bad sectors that it can no longer automatically remap them and is now telling you there is a problem.
In my experience, every single drive that I've seen reporting even a single bad sector will soon go pear-shaped and shouldn't be used.
Wow, that's a pretty intense workload! Yes, it's very difficult to get SLC SSDs these days, everyone has gone for MLC, even the Intel DC grade units...
Yeah, but then you can't accurately know the cost of de-duplication, additionally you're doing work against already committed files which is a big no-no if you want stable storage. If I commit a file, I don't want a background process to read/write it and a software bug to screw it up years down the road.
Additionally, you're taking away resources from a system that will already be taxed. My file server has a load of 1.2-2.5 on an average day (because I'm running against the IOPS limits on my 5-year old SSD's), doing ANYTHING (even streaming a backup) has to be meticulously planned so as not to affect the system.
Wow, what are you doing on your server that you're thrashing your SSDs with 500+ IOPS 24/7?