We just moved to a similar system, and it has been pretty workable. We're using star names tied to a physcal server, which lasts forver, even past role changes. We then use subdomained aliases as you describe.
I am wondering how to name new VMs, though... right now all VMs in our cluster have been converted from physcial boxes under our old naming scheme. That scheme was terrible for users (80s TV characters). Do you just name VMs with function directly, or do you give them a different naming scheme?
But here's the thing: IT WAS NEVER AN UNLIMITED SERVICE, nor was it probably ever advertised as "unlimted" (at least not in the last 5 years). Your Committed Information Rate is either zero kbps or very close to it. ZERO. Read your contract, and read the updated terms and conditions they publish every so often. You can actually reject many of the changes.
For consumer internet contracts, all service is best-effort, and you pay for unlimited best-effort servcice. That means traffic shaping and capping are fair practices, as is just about anything else the ISP wants to do to make best-effort on behalf of the mass of their customers. You do not have a contract for guaranteed transit of your traffic (which costs a whole lot more).
Dude, you do not "pay for 10 Mbps". You pay for ZERO Mbps Committed Information Rate, with best best-effort service up to 10 Mbps. Read your contract.
Committed, uncapped, non-oversold internet transit bandwith is US$50-200 per Mbps in most US markets. Businesses buy these connections for guaranteed throughput. $300 for 1.5 Mbps of guaranteed bandwidth on a T1, or 45 Mbps for a couple of grand.
Consumer Internet service has always been best-effort to make it affordable, even in places like Japan and Sweden with government-subsidized fiber. There are typically no throughput guarantees at all, or very low guarantees. The backbone connectivity is massively oversubscribed at 20:1 or higher ratios.
If you want 10 Mbps guaranteed, pony up the $500+ per month that most corporate customers do.
because society only has to pay to solve (the software portion of) a particular problem once.
Wait a minute... I thought this was about choice. KDE or Gnome, MySQL or PostGresSQL, two thousand different email clients, etc. Free software seemst to "solve" the same problem several times over. THere are even dozens of alternatives to Apache.
So which is it? Either free software eliminates redundant work, or it enables choice. You can't have it both ways.
A lot of those assumptions made by the OS about disks are also incorrect for large RAID arrays or SANs as well, not just SSDs. The biggest problem I think would be read-ahead caching based on sequential block ID, which often actually becomes a performance-killing random read on a SAN or even local storage system with volume management or snapshots enabled.
It would be nice if the interface for all storage was just an array of blocks, as the GP suggests, but filesystems really do need to know about the physical characteristics of the storage to be efficient. Which is why NetApp filers, for example, often perform better in NAS mode than iSCSI or FC SAN mode (depending on the filesystem and application). NetApp's WAFL and OS know a lot about the hardware they're running on. Your FC-mounted LUN formatted with EXT3 or NTFS doesn't know jack about the characteristics of all the components inside the NetApp box.
This knowledge of the hardware is why ZFS is killer, by the way, and is not the "massive layering violation" that the Linux kernel guys seem to think it is.
yet they still expect that they are somehow entitled to money for it because Google went back on its word (not contractually..just its "word").
In the USA, a verbal agreement is a legally binding contract if there is an exchange of value. Of course, from a practical standpoint, there may needs to be some witnesses to the content of a verbal agreement in case it is challenged in a lawsuit.
Even a one-page email that says "we agree to X if you do Y" is a legally valid contract if it is accepted by both parties. Which is why you should always use phrases like "what if" and "pending management and legal reveiw" until the lawyers actually get involved.
Actually, most commercial SAN/NAS devices use a BSD-based kernel. NetApp, EqualLogic, etc. The GPL is a legal minefield that many commerical companies don't want to mess around with, especially when BSD-style licenses are available for most OS pieces. Lawyers cost too much.
Moreover, those commercial SAN/NAS that are Linux based typically only really use the kernel and some device drivers. The storage stack (file systems, volume management, clusterring, management interfaces, etc.) are typically proprietary "special sauce" technology and not open source at all.
To combat the increase of traffic why cant businesses/ISP's work more towards multicasting
Because there is no viable billing model for multicast IP traffic as of yet. How do you charge for traffic created/amplified by routers? ISP A has to trust that ISP B isn't lying about the amount of traffic actually delivered to end users, becuase ISP A has no visibility once a multicast packet leaves their network and enters that of ISP B.
That makes no sense at all. It takes more energy to grow and harvest food than is gained by eating it. Is farming "not viable"? 10000 years of human history suggests a degree of viability.
Client bandwidth limits are optional, so it requires cooperation from the BitTorrent user, who is not likely to slow their downloads voluntarily.
Filtering at the client end of the connection doesn't help at all, as the download bottleneck being discussed is occuring on one of the (slowish) last-mile links from the backbone to the end user. QoS or throttling would have to be applied upstream from the user to be of value. In general you have no absolute control over the rate at which others on the net send *you* IP packets. (A TCP window size is just a suggestion, and does nothing to slow down a dumb or malicious sender).
Motors give people enough kinetic energy to do real damage.
So do bicycles, because the people involved have a lot less protection than they do in a car. People are killed on the lakefront bike paths in Chicago regularly, almost always by cyclists (and a roller blader or three). Scale that deaths/passenger-mile rate up to mass commute traffic levels, and I don't think bicycling is a big win. Have you looked at statistics from urban China?
I myself have spent time in the hospital from a bike wreck involving nothing but myself, my bike, and a rock the size of a golf ball. Bicycles are not generally safer than cars.
I will not use a binary program, unless I have the source and can verify it is legitimate.
Surely your compiler is a binary? And your BIOS? Disk firmware? Processor microcode?
Unless you built (not assembled, mind you, but rather built from bits of copper and silicon) all the hardware yourself, at some point, there is a binary black box you have to trust.
"Fretting about overpopulation, is a perfect guilt-free - indeed, sanctimonious - way for 'progressives' to be racists."
- O'Rourke, P.J. (1994), All the Trouble in the World
Another exercise that P.J. came up with... anytime you hear an "intellectual" talk about "overpopulation", replace that word with "niggers" and you'll understand what they're really saying.
Using an example posted elsewhere this thread: "Overpopulation is a serious threat to the long-term sustainability of the world's fresh water supplies." becomes "Niggers are serious threat to the long-term sustainability of the world's fresh water supplies."
Well, until someone gave them a taser gun. Now, shoot first is the rule because they won't get sued, and don't have to worry about it.
Not true. Here in Chicago, the police get sued all the time - almost daily. Multi-million dollar judgements are commonplace, and the cops in question are invariably sacked (and sometimes financially ruined).
Back in the Windows 3.x and Windows 9x days it was fairly easy to swap out the desktop manager, but I don't think thats the case now.
Replace explorer.exe with your own window manager. Ignore the warnings that say "this is a ciritcal system file, do you really want to replace it."
Of course, to run WIndows software, you'll have to re-implement a lot of the Win32 API as well, but it would not be impossible. But what would be the point?
Yes, I did. I wasn't implying you were "wrong", but that input validation is ALWAYS necessary, no matter what. Even if you ignore security concerns, input validation very important for catching bugs and providing useful error messages.
There is still validation necessary. An attacker could insert Javascript into strings in your database, which could then be used for attack if not escaped when read out and displayed in the web page.
Since you can't rely on remebering to HTML encode all sttrings coming out of the DB into the web app (even if a framework does it automatically, somebody might end-run around the framework in the future), you really should cehck for and sanitize Javascript etc on the way into the database as well as on the way out.
And of course there are lots of other application-specific reasons to validate input; giving the user good error messages comes to mind. Parameterized SQL helps prevent SQL injection, but it does not remove the need to validate and sanitize input.
The first edition of the CGM standard wasn't published until 1987. AutoCAD was released in 1982, and previous UNIX or mini-based CAD programs were even more proprietary.
I used AutoCAD on UNIX and on the Mac in the late 1980s in school, and it was a very proprietary solution. Tablets, pucks, and plotters all had AutoCAD-specific drivers IIRC. You could not exchange files with anything; to import something meant to take a plot of it from some other system, tape it to your tablet, and trace it into AutoCAD. All other CAD systems at the time were similarly closed (our school used several, but none of my classes used the others).
Face it: poprietary file formats are the norm in the computer industry, especially when you look back in time to the mainframe, mini, and early UNIX eras. Standard formats for most application spaces simply did not exist. Microsoft is hardly the first nor the worst in this area. IBM invented "vendor lock-in" on the mainframe platform, for heaven's sake.
AutoCAD is Windows-only *now*, but the files originated on the UNIX and Mac Versions of AutoCAD from teh 1980s. The major point is that there was NO standard file format for CAD back then, so what was a vendor to do except make one up? These "standards" you talk about didn't exist for most application spaces until very recently. Even office documents!
Slashdot Mirror
We just moved to a similar system, and it has been pretty workable. We're using star names tied to a physcal server, which lasts forver, even past role changes. We then use subdomained aliases as you describe.
I am wondering how to name new VMs, though... right now all VMs in our cluster have been converted from physcial boxes under our old naming scheme. That scheme was terrible for users (80s TV characters). Do you just name VMs with function directly, or do you give them a different naming scheme?
But here's the thing: IT WAS NEVER AN UNLIMITED SERVICE, nor was it probably ever advertised as "unlimted" (at least not in the last 5 years). Your Committed Information Rate is either zero kbps or very close to it. ZERO. Read your contract, and read the updated terms and conditions they publish every so often. You can actually reject many of the changes.
For consumer internet contracts, all service is best-effort, and you pay for unlimited best-effort servcice. That means traffic shaping and capping are fair practices, as is just about anything else the ISP wants to do to make best-effort on behalf of the mass of their customers. You do not have a contract for guaranteed transit of your traffic (which costs a whole lot more).
Dude, you do not "pay for 10 Mbps". You pay for ZERO Mbps Committed Information Rate, with best best-effort service up to 10 Mbps. Read your contract.
Committed, uncapped, non-oversold internet transit bandwith is US$50-200 per Mbps in most US markets. Businesses buy these connections for guaranteed throughput. $300 for 1.5 Mbps of guaranteed bandwidth on a T1, or 45 Mbps for a couple of grand.
Consumer Internet service has always been best-effort to make it affordable, even in places like Japan and Sweden with government-subsidized fiber. There are typically no throughput guarantees at all, or very low guarantees. The backbone connectivity is massively oversubscribed at 20:1 or higher ratios.
If you want 10 Mbps guaranteed, pony up the $500+ per month that most corporate customers do.
Wait a minute... I thought this was about choice. KDE or Gnome, MySQL or PostGresSQL, two thousand different email clients, etc. Free software seemst to "solve" the same problem several times over. THere are even dozens of alternatives to Apache.
So which is it? Either free software eliminates redundant work, or it enables choice. You can't have it both ways.
A lot of those assumptions made by the OS about disks are also incorrect for large RAID arrays or SANs as well, not just SSDs. The biggest problem I think would be read-ahead caching based on sequential block ID, which often actually becomes a performance-killing random read on a SAN or even local storage system with volume management or snapshots enabled.
It would be nice if the interface for all storage was just an array of blocks, as the GP suggests, but filesystems really do need to know about the physical characteristics of the storage to be efficient. Which is why NetApp filers, for example, often perform better in NAS mode than iSCSI or FC SAN mode (depending on the filesystem and application). NetApp's WAFL and OS know a lot about the hardware they're running on. Your FC-mounted LUN formatted with EXT3 or NTFS doesn't know jack about the characteristics of all the components inside the NetApp box.
This knowledge of the hardware is why ZFS is killer, by the way, and is not the "massive layering violation" that the Linux kernel guys seem to think it is.
Well, there's always RFC 2606. But naming your Active Directory domain "mycompany.invalid" seems a bit harsh.
In the USA, a verbal agreement is a legally binding contract if there is an exchange of value. Of course, from a practical standpoint, there may needs to be some witnesses to the content of a verbal agreement in case it is challenged in a lawsuit.
Even a one-page email that says "we agree to X if you do Y" is a legally valid contract if it is accepted by both parties. Which is why you should always use phrases like "what if" and "pending management and legal reveiw" until the lawyers actually get involved.
I don't think enterprise-level customers or shareholders would appreciate that wait very much...
Actually, most commercial SAN/NAS devices use a BSD-based kernel. NetApp, EqualLogic, etc. The GPL is a legal minefield that many commerical companies don't want to mess around with, especially when BSD-style licenses are available for most OS pieces. Lawyers cost too much.
Moreover, those commercial SAN/NAS that are Linux based typically only really use the kernel and some device drivers. The storage stack (file systems, volume management, clusterring, management interfaces, etc.) are typically proprietary "special sauce" technology and not open source at all.
Because there is no viable billing model for multicast IP traffic as of yet. How do you charge for traffic created/amplified by routers? ISP A has to trust that ISP B isn't lying about the amount of traffic actually delivered to end users, becuase ISP A has no visibility once a multicast packet leaves their network and enters that of ISP B.
That makes no sense at all. It takes more energy to grow and harvest food than is gained by eating it. Is farming "not viable"? 10000 years of human history suggests a degree of viability.
Stop skipping your high school economics classes.
Because...
So do bicycles, because the people involved have a lot less protection than they do in a car. People are killed on the lakefront bike paths in Chicago regularly, almost always by cyclists (and a roller blader or three). Scale that deaths/passenger-mile rate up to mass commute traffic levels, and I don't think bicycling is a big win. Have you looked at statistics from urban China?
I myself have spent time in the hospital from a bike wreck involving nothing but myself, my bike, and a rock the size of a golf ball. Bicycles are not generally safer than cars.
I was an engineer. What is this "latin" of which you speak?
I went to a Catholic University, you insensitive clod!
Surely your compiler is a binary? And your BIOS? Disk firmware? Processor microcode?
Unless you built (not assembled, mind you, but rather built from bits of copper and silicon) all the hardware yourself, at some point, there is a binary black box you have to trust.
Except connect to a SONNET network. Or a DS3 interface. Or aggregate multiple T1s. Or suport terabit switching and routing speeds.
"Fretting about overpopulation, is a perfect guilt-free - indeed, sanctimonious - way for 'progressives' to be racists."
- O'Rourke, P.J. (1994), All the Trouble in the World
Another exercise that P.J. came up with... anytime you hear an "intellectual" talk about "overpopulation", replace that word with "niggers" and you'll understand what they're really saying.
Using an example posted elsewhere this thread: "Overpopulation is a serious threat to the long-term sustainability of the world's fresh water supplies." becomes "Niggers are serious threat to the long-term sustainability of the world's fresh water supplies."
Quite a lot of spam originates in Europe, with 4 of the top 10 spam sources being Euorpean. RIPE's system is hardly foolproof!
Not true. Here in Chicago, the police get sued all the time - almost daily. Multi-million dollar judgements are commonplace, and the cops in question are invariably sacked (and sometimes financially ruined).
Here are some examples.
Replace explorer.exe with your own window manager. Ignore the warnings that say "this is a ciritcal system file, do you really want to replace it."
Of course, to run WIndows software, you'll have to re-implement a lot of the Win32 API as well, but it would not be impossible. But what would be the point?
Yes, I did. I wasn't implying you were "wrong", but that input validation is ALWAYS necessary, no matter what. Even if you ignore security concerns, input validation very important for catching bugs and providing useful error messages.
There is still validation necessary. An attacker could insert Javascript into strings in your database, which could then be used for attack if not escaped when read out and displayed in the web page.
Since you can't rely on remebering to HTML encode all sttrings coming out of the DB into the web app (even if a framework does it automatically, somebody might end-run around the framework in the future), you really should cehck for and sanitize Javascript etc on the way into the database as well as on the way out.
And of course there are lots of other application-specific reasons to validate input; giving the user good error messages comes to mind. Parameterized SQL helps prevent SQL injection, but it does not remove the need to validate and sanitize input.
The first edition of the CGM standard wasn't published until 1987. AutoCAD was released in 1982, and previous UNIX or mini-based CAD programs were even more proprietary.
I used AutoCAD on UNIX and on the Mac in the late 1980s in school, and it was a very proprietary solution. Tablets, pucks, and plotters all had AutoCAD-specific drivers IIRC. You could not exchange files with anything; to import something meant to take a plot of it from some other system, tape it to your tablet, and trace it into AutoCAD. All other CAD systems at the time were similarly closed (our school used several, but none of my classes used the others).
Face it: poprietary file formats are the norm in the computer industry, especially when you look back in time to the mainframe, mini, and early UNIX eras. Standard formats for most application spaces simply did not exist. Microsoft is hardly the first nor the worst in this area. IBM invented "vendor lock-in" on the mainframe platform, for heaven's sake.
AutoCAD is Windows-only *now*, but the files originated on the UNIX and Mac Versions of AutoCAD from teh 1980s. The major point is that there was NO standard file format for CAD back then, so what was a vendor to do except make one up? These "standards" you talk about didn't exist for most application spaces until very recently. Even office documents!