As a grizzled vet of many a foo'd environment, I'd really like to say forget the hallmark holiday. I know what I'm paid to do, and I'm damned good at it. If I'm appreciated, that'll show that in day-to-day interactions, and if not - well, I still have a job to do. I'm not paid for sitting around and whining about how hard my life is, ooh poor me.
Funny you should mention that. We recently upgraded our colocation circuit to gigabit ethernet, and one of my staff members from the cage downloaded a 250MB file from - I don't know, YouTube, maybe? - in 10 seconds. Obviously that's not coming close to saturating gig e, but I think I could live with ISOs being transferred in 30 seconds. (it helped that YouTube was 2ms away, as well)
Of course, as you say, yes - depends on the remote provider.
Is Innovation one of those words like Hero that is starting to use meaning through dilution? Sun might be one of the few companies that could pull off the transition from high-margin to low-margin computing (which is ultimately what staved in the heads of SGI, Digital, Data General, Sequent, HP and tons of others). Sun's just going to take longer to die than its brethren, thanks to generally conservative single-vendor fanboys who track Sun like gospel. They *might* be able to pull it off, but how one can lose half a billion dollars over the last three quarters amidst and still survive long-term - that's a good question....
Basically on topic and to the point, except I think your characterization of Mac users is a little overbroad. Lots of us Mac users have strong technical background, and muck with all manner of computational foo-ery during the day. The last thing I'm looking forward to in the evening is monkeying around with more of the same. So yes, you're right about the appeal of Apple, just that it applies in more cases than just those who want ignorance. "Just play me my music, will ya? Do the right thing, damn you..."
Same has happened here - but we've got some pretty decent management. Most of our market is B2B. We've gotten the occasional "fine customer" who through helpful heaps of abuse, have actually brought members of our support team to tears. Unbelievable abuse, laced with filth and personal attacks on gender, intelligence, ethnicity, and sexual orientation (including their mothers, sisters, fathers, etc.).
However, we also have VPs who call other VPs at the offending companies, and explain to them exactly the content of the conversation that occurred (a good audio replay usually gets people's attention). That is usually about all it takes to clean up abusers. Too bad that's not usually something that can happen in the enduser support arena....
(aside: honestly, must we report everything a vendor says about their own product? When would that NOT be just spin and hype?)
Yesterday, we read a blog from a Microsoftie indicating the problems Vista developers are facing: a) aversion to truth b) too many cooks (2000 developers, I believe was mentioned) c) impossible for one developer to comprehend Vista as a single entity d) 50 layers of dependencies (including circular dependencies). e) code approaching 50 million LOC
Now, I suppose with the slow pace and unit testing going on in MS development, MAYBE you've got 2000 developers who are writing secure code. Seems highly unlikely, though, given MS' past performance and daunting challenges.
Really, I fail to see how every Linux user needs to read complex sysadmin books and learn everything about the command line.
Good point.
As a syadmin, I live and die in the command line, and there's enough cool shit out there on the command line that every now and then I'll just sit back and sigh happily....
Been doing this so long, it's hard for me to think of Linux as user apps (though I use them).
Not sure a book is the best way to go... most linux boxen have enough on there that you could bootstrap yourself into linux mastery without too much difficulty. Start with these commands: cd ls mv rm chmod chgrp man
Learn about the command line switches via man. When you think you've got a handle on that, add these: bash find grep sort uniq vi|emacs|nano (not touching this one editorial-wise - use whatever editor rocks your boat, I've got my favorites, you'll have yours)
Then top it off with one or more of the following: awk perl python
and for network, try: netstat (many many options here) traceroute ping route ifconfig
That's pretty much the shit you'll use most days if you're larking around linux. Then start browing around with man.
Minus these, I can see many systems that could fail with a little effort. One of the problems I see with our current infrastructure is the notion of machine-to-machine communication - when really, what we want to know is in effect, remove anonymity from the equation (there will be discussion this point, I realize). Machines talk to each other as machines. We ultimately want to know WHO did X, or Y, so we can find them and hurt them in some fashion (bullets to the temple, fines, whatever...). (okay, substitute we for I if it makes you feel better).
This is a really nasty point. Privacy versus safety. Or, in this case, utility. The internet does no one any good if denial of services render it unusable - and of course, a good DDoS exploits the behavior of its regular users, so that effective rebuttal becomes increasingly difficult.
I find myself disillusioned by the human race. There are no sacred cows so holy that someone won't shit all over it.
Poor old SGI. They built amazingly excellent hardware, bleeding edge software, paid their workers well, treated employees like kings and customers like emporers, and donated heavily to the open source movement.
SGI did a lot of things like you mentioned well, but treating their customers well - was a mixed bag in my opinion. Support was top-notch, just at about NetApp levels.
Let's talk about sales.
SGI Salespeople seemed to be constructed primarily out of ex-Porn producers replete with coke spoons around their necks. You were SO LUCKY to be ALLOWED to even THINK about buying their incredible hardware! And yes yes, incredible hardware, but also wildly priced out-of-proportion: drives were ridiculously expensive, and excuse me? $5000 for a 100BaseT copper card in 1997 (this was for an Onyx)? NO WAY!
All that might have been excused. And then came Farenheit, engineer exodus to nVidia, Jim whatever-the-hell-his-name-was Microsoft Lacky. All these participated in the final ass rape of SGI, but more than that was, as has been mentioned, the simple commoditization of hardware. Customer "wins" while they "lose".
Despite it all, still worth lighting a candle for the splendor that was SGI....
As has been mentioned, BIND can do this without much issue - EXCEPT it doesn't QUITE follow all expectations. Apparently, BIND uses a common name cache - so although a recursive query with an uncached answer can arrive at a DNS server and be rejected, if that BIND server ALREADY has a cached answer (eg, from an internal query) - it'll serve that up to the requestor. This was true in implementations of BIND about 6 months ago - I'm not sure things have changed since...
If by "the way to go" you mean, hard-disk-based backups, then I'd agree with you. In this particular case, this acts as a poor-man's replacement for RAID-1 (mirroring), with the same problems inherent in that system that make it unsuitable for general backups. Consider a simple command - "rm -rf s *". Ooops! With a point-in-time backup, you're not necessarily SOL, though of course, you weigh that against the data lost between your backups.
I've seen it time and again, most recently at my current employer when I proposed a NAS based on Linux that would cost less than half of what we ended up buying (the difference, mind you, was more than I get paid annually). The manager in charge of purchasing it didn't 'trust' that 'this Linux thing' would stay free or that he'd be able to keep it running if I left for another job.
I've built and bought NAS servers for years - Solaris, EMC Celerra, HP, SGI, Digital, Linux, NetApp - and frankly, I got tired of having to dick around with Unix-based hosts. I've got a lot to do, and worrying about upgrading my redundant Linux NAS is not one of the things that I'd rather do. Dedicated NAS is not without its drawbacks (flexibility of architecture, and mandated used of NFS or CIFS spring to mind), but there's a lot to be said for a system that takes me 5 minutes to setup, does NFS better than just about any Unix-based host I've seen, gives me new and super-easy methods of recovering files (yay, snapshots), does failover without fuss, and ties me in directly to top tier vendor support.
That said, I'm not going to buy a NetApp for internal systems where it's not completely critical to have this functionality - there's better use for those dollars - but for production uses, I don't think I'll be buying much else for a few years (until I change my mind again). And I won't be buying EMC Celerra again, ever.
(the main point of your post I feel remains valid - if you have management who are unwilling to add appropriate resources to a problem (eg, additional trained and competent staff) and expects their vendor to fix everything for them, you've got larger problems).
Yes yes, but this demonstrates why the notion of machine=>your identity is not valid. The current protocols treat network traffic as machine to machine. The notion of personal identity doesn't map well onto this concept.
If the protocols (eg, alternative to TCP/IP) could be reworked so that concepts like person-to-person, person-to-service, and service-to-service connections were possible (and unspoofable), that'd go a long way towards allowing us to build enormous, decentralized mesh networks where Internet connectivity is ubiquitous.
Implementation is left as an exercise to the reader.;)
I think your scenario is possible and maybe even likely (wrt hardware sales for 2005 chomping hard). However, the folks at Apple aren't idiots - from their press release:
The press release contains forward-looking statements about the Company's plans to deliver Macintosh products using Intel microprocessors by this time next year, the Company's current plans to transition all Macs to Intel microprocessors by the end of 2007, and the expectation that Microsoft and Adobe will create future versions of Microsoft Office and Creative Suite, respectively, for the Mac that support both Power PC and Intel processors. These statements involve risks and uncertainties and actual results may differ. Potential risks and uncertainties include the Company's ability to make timely delivery of the new products with Intel microprocessors and related hardware and software technological changes and innovations to support Intel microprocessors; the development and availability of components and services essential to enable the Company to deliver such products in a timely manner; the effect that the Company's dependency on manufacturing and logistics services provided by third parties may have on the timing, quality, quantity or cost of products manufactured or services rendered; the Company's ability to evolve its operating system and attract sufficient Macintosh developers; the Company's dependency on third-party software developers such as Microsoft and Adobe to timely develop future applications that run on Macs that support Intel microprocessors and Power PC microprocessors; the effect competitive and economic factors and the Company's reaction to them may have on consumer and business buying decisions with respect to the Company's products; and the potential negative impact the transition of all Macs to Intel microprocessors by the end of 2007, or the announcement of such transition, might have on sales of current or future Mac products with Power PC processors.
I think it's doubtful about MacOS X Server being able to run on Dells, though from our point of view (company I work for), MacOS X Server could then be considered for some portion of our workload (and would be welcomed).
Interesting times.
Mail.app was part of NeXTStep in the late 80s. Whether it had photo lookup out of address book, I don't know. It certainly had support for Faces, which should in itself be considered prior art that invalidates this patent.
A major portion of IP law that needs to change is the ability to transfer it. This often results in IP brokerage houses that do nothing useful except collecting licensing fees for their held patents.
Patent protections need to start reflecting their original intention - to grant to that PERSON the right to solely benefit from their invention.
Moderators, wake up and mark parent down (or at least funny, or troll)!
Several severe reality problems with this "advice" (it's surely a troll, people - come on, "DNS cyber buffer?"): While that's a sure fire way of killing cache poisoning for your own records, setting DNS TTL to 0 for all records *will* cause severe Internet Armageddon as the root DNS servers explode (client DNS servers would be screwed in short order as well).
Since DNS is a distributed system, run by admins clueful and otherwise, setting DNS TTL to 0 everywhere is not possible (short of owning every single DNS server out there).
Further, setting DNS TTL to 0 does nothing to prevent caching of records on your own DNS server (and serving it to your clients).
Slashdot Mirror
As a grizzled vet of many a foo'd environment, I'd really like to say forget the hallmark holiday. I know what I'm paid to do, and I'm damned good at it. If I'm appreciated, that'll show that in day-to-day interactions, and if not - well, I still have a job to do. I'm not paid for sitting around and whining about how hard my life is, ooh poor me.
Funny you should mention that. We recently upgraded our colocation circuit to gigabit ethernet, and one of my staff members from the cage downloaded a 250MB file from - I don't know, YouTube, maybe? - in 10 seconds. Obviously that's not coming close to saturating gig e, but I think I could live with ISOs being transferred in 30 seconds. (it helped that YouTube was 2ms away, as well)
Of course, as you say, yes - depends on the remote provider.
Is Innovation one of those words like Hero that is starting to use meaning through dilution? Sun might be one of the few companies that could pull off the transition from high-margin to low-margin computing (which is ultimately what staved in the heads of SGI, Digital, Data General, Sequent, HP and tons of others). Sun's just going to take longer to die than its brethren, thanks to generally conservative single-vendor fanboys who track Sun like gospel. They *might* be able to pull it off, but how one can lose half a billion dollars over the last three quarters amidst and still survive long-term - that's a good question....
Basically on topic and to the point, except I think your characterization of Mac users is a little overbroad. Lots of us Mac users have strong technical background, and muck with all manner of computational foo-ery during the day. The last thing I'm looking forward to in the evening is monkeying around with more of the same. So yes, you're right about the appeal of Apple, just that it applies in more cases than just those who want ignorance. "Just play me my music, will ya? Do the right thing, damn you..."
Same has happened here - but we've got some pretty decent management. Most of our market is B2B. We've gotten the occasional "fine customer" who through helpful heaps of abuse, have actually brought members of our support team to tears. Unbelievable abuse, laced with filth and personal attacks on gender, intelligence, ethnicity, and sexual orientation (including their mothers, sisters, fathers, etc.).
However, we also have VPs who call other VPs at the offending companies, and explain to them exactly the content of the conversation that occurred (a good audio replay usually gets people's attention). That is usually about all it takes to clean up abusers. Too bad that's not usually something that can happen in the enduser support arena....
sloth_jr
(aside: honestly, must we report everything a vendor says about their own product? When would that NOT be just spin and hype?)
Yesterday, we read a blog from a Microsoftie indicating the problems Vista developers are facing:
a) aversion to truth
b) too many cooks (2000 developers, I believe was mentioned)
c) impossible for one developer to comprehend Vista as a single entity
d) 50 layers of dependencies (including circular dependencies).
e) code approaching 50 million LOC
Now, I suppose with the slow pace and unit testing going on in MS development, MAYBE you've got 2000 developers who are writing secure code. Seems highly unlikely, though, given MS' past performance and daunting challenges.
sloth jr
Good point.
As a syadmin, I live and die in the command line, and there's enough cool shit out there on the command line that every now and then I'll just sit back and sigh happily....
Been doing this so long, it's hard for me to think of Linux as user apps (though I use them).
sloth jr
Not sure a book is the best way to go... most linux boxen have enough on there that you could bootstrap yourself into linux mastery without too much difficulty. Start with these commands:
cd
ls
mv
rm
chmod
chgrp
man
Learn about the command line switches via man. When you think you've got a handle on that, add these:
bash
find
grep
sort
uniq
vi|emacs|nano (not touching this one editorial-wise - use whatever editor rocks your boat, I've got my favorites, you'll have yours)
Then top it off with one or more of the following:
awk
perl
python
and for network, try:
netstat (many many options here)
traceroute
ping
route
ifconfig
That's pretty much the shit you'll use most days if you're larking around linux. Then start browing around with man.
sloth jr
Minus these, I can see many systems that could fail with a little effort. One of the problems I see with our current infrastructure is the notion of machine-to-machine communication - when really, what we want to know is in effect, remove anonymity from the equation (there will be discussion this point, I realize). Machines talk to each other as machines. We ultimately want to know WHO did X, or Y, so we can find them and hurt them in some fashion (bullets to the temple, fines, whatever...). (okay, substitute we for I if it makes you feel better).
This is a really nasty point. Privacy versus safety. Or, in this case, utility. The internet does no one any good if denial of services render it unusable - and of course, a good DDoS exploits the behavior of its regular users, so that effective rebuttal becomes increasingly difficult.
I find myself disillusioned by the human race. There are no sacred cows so holy that someone won't shit all over it.
Sloth Jr
SGI did a lot of things like you mentioned well, but treating their customers well - was a mixed bag in my opinion. Support was top-notch, just at about NetApp levels.
Let's talk about sales.
SGI Salespeople seemed to be constructed primarily out of ex-Porn producers replete with coke spoons around their necks. You were SO LUCKY to be ALLOWED to even THINK about buying their incredible hardware! And yes yes, incredible hardware, but also wildly priced out-of-proportion: drives were ridiculously expensive, and excuse me? $5000 for a 100BaseT copper card in 1997 (this was for an Onyx)? NO WAY!
All that might have been excused. And then came Farenheit, engineer exodus to nVidia, Jim whatever-the-hell-his-name-was Microsoft Lacky. All these participated in the final ass rape of SGI, but more than that was, as has been mentioned, the simple commoditization of hardware. Customer "wins" while they "lose".
Despite it all, still worth lighting a candle for the splendor that was SGI....
Sloth Jr
As has been mentioned, BIND can do this without much issue - EXCEPT it doesn't QUITE follow all expectations. Apparently, BIND uses a common name cache - so although a recursive query with an uncached answer can arrive at a DNS server and be rejected, if that BIND server ALREADY has a cached answer (eg, from an internal query) - it'll serve that up to the requestor. This was true in implementations of BIND about 6 months ago - I'm not sure things have changed since...
If by "the way to go" you mean, hard-disk-based backups, then I'd agree with you. In this particular case, this acts as a poor-man's replacement for RAID-1 (mirroring), with the same problems inherent in that system that make it unsuitable for general backups. Consider a simple command - "rm -rf s *". Ooops! With a point-in-time backup, you're not necessarily SOL, though of course, you weigh that against the data lost between your backups.
Living together
Of course! Who gives two shits about a million dollars? It's what could be done with that money that's exciting. Money makes the world go round.
This sure sounds like it came directly out of Seagate's marketing department.
I've built and bought NAS servers for years - Solaris, EMC Celerra, HP, SGI, Digital, Linux, NetApp - and frankly, I got tired of having to dick around with Unix-based hosts. I've got a lot to do, and worrying about upgrading my redundant Linux NAS is not one of the things that I'd rather do. Dedicated NAS is not without its drawbacks (flexibility of architecture, and mandated used of NFS or CIFS spring to mind), but there's a lot to be said for a system that takes me 5 minutes to setup, does NFS better than just about any Unix-based host I've seen, gives me new and super-easy methods of recovering files (yay, snapshots), does failover without fuss, and ties me in directly to top tier vendor support.
That said, I'm not going to buy a NetApp for internal systems where it's not completely critical to have this functionality - there's better use for those dollars - but for production uses, I don't think I'll be buying much else for a few years (until I change my mind again). And I won't be buying EMC Celerra again, ever.
(the main point of your post I feel remains valid - if you have management who are unwilling to add appropriate resources to a problem (eg, additional trained and competent staff) and expects their vendor to fix everything for them, you've got larger problems).
Yes yes, but this demonstrates why the notion of machine=>your identity is not valid. The current protocols treat network traffic as machine to machine. The notion of personal identity doesn't map well onto this concept.
;)
If the protocols (eg, alternative to TCP/IP) could be reworked so that concepts like person-to-person, person-to-service, and service-to-service connections were possible (and unspoofable), that'd go a long way towards allowing us to build enormous, decentralized mesh networks where Internet connectivity is ubiquitous.
Implementation is left as an exercise to the reader.
sloth jr
Uh, talk about unearned power, priviledge, and position - give a portion of the estate to the state!
Please read the first linked article. Answered in second paragraph.
He was forced to resign. Hard to see how that saves him financially in any way.
Mail.app was part of NeXTStep in the late 80s. Whether it had photo lookup out of address book, I don't know. It certainly had support for Faces, which should in itself be considered prior art that invalidates this patent.
"because it's source-based, it's notorious for its speed" - so what, no other distribution uses source to compile its distro?
A major portion of IP law that needs to change is the ability to transfer it. This often results in IP brokerage houses that do nothing useful except collecting licensing fees for their held patents.
Patent protections need to start reflecting their original intention - to grant to that PERSON the right to solely benefit from their invention.
Moderators, wake up and mark parent down (or at least funny, or troll)!
Several severe reality problems with this "advice" (it's surely a troll, people - come on, "DNS cyber buffer?"):
While that's a sure fire way of killing cache poisoning for your own records, setting DNS TTL to 0 for all records *will* cause severe Internet Armageddon as the root DNS servers explode (client DNS servers would be screwed in short order as well).
Since DNS is a distributed system, run by admins clueful and otherwise, setting DNS TTL to 0 everywhere is not possible (short of owning every single DNS server out there).
Further, setting DNS TTL to 0 does nothing to prevent caching of records on your own DNS server (and serving it to your clients).