I know that, you know that, but it's very obvious from the article that Cringely has no clue whatsoever. He still thinks that peering points are some weird rarified animal where only telcos with enormously expensive routers can play, when the reality is that anyone with $3000/month to spend can host their equipment "near the core."
(And I find the idea of Google parking a tractor trailer with their hypothetical container-box node on the street next to 111 8th ave particularly amusing. Nevermind heat dissipation issues: can they afford the parking tickets?)
No, no, no. This is exactly my point: it won't buy them anything in terms of reduced lag, because most internet traffic doesn't go over public peering exchanges any more. Putting a Google node at MAE-East would be substantially less effective than just renting a dozen racks at Equinox and ordering a fistful of 10-gigabit crossconnects to whatever carriers are available there.
Even assuming the power and heat requirements of cramming that many opterons into that small a space could be dealt with, there's another, larger problem:
It's not fucking 1997 any more.
"Peering points" -- big, open-access traffic exchange handoffs like the old MAE-East and MAE-West used to be a big deal back in the late 90s, when OC-12 circuits were still rare and hideously expensive beasts, and Gigabit Ethernet was still a gleam in some 3Com engineer's eye.
In 2005, they simply don't matter. The big players (level3, MCI/Verizon, Qwest, SBC, etc) all exchange traffic over private fiber interconnects, and everyone else buys transit from the big guys directly or ponies up for a switch port at Equinox, PAIX/Switch&Data or some other 'carrier neutral' colocation center. Dropping a datacenter-in-a-box onto MAE-east or any of its surviving ilk would buy Google precisely nothing.
(And nevermind the fact that google is documented to own thousands upon thousands of unused square feet of datacenter space already: they went on a very well-thought-out buying spree in 2000-2001 when all the dot-com datacenter companies were going out of business, and are very well provisioned for the forseeable future as a result.)
Now, a much more interesting application of the "Google node in a shipping container" idea can be summed up in one simple word: China. Why wait for the local market to develop the infrastructure you need when you can just drop a box down and then run fiber to it? I'm still dubious though...
Although I think remaking The Prisoner is an awful idea, SkyTV's involvement doesn't necessarily make it worse. Just because they're part of News Corp doesn't mean that they're the sample people who produce "Cops".
Sky is probably best known in the US for co-bankrolling Ron Moore's remake of Battlestar Galactica with SciFi Network, which I believe at least one or two people around these parts liked. (This is why the first season of Galactica aired first on Sky -- they got first-airing rights as part of the deal.)
Mmmmph. I won't make a blanket defense of "Fall Out" -- it's certainly not flawless, and doesn't necessarily hold up that well today -- but I tend to think of its flaws as artifacts of its era, rather than as outright artistic failures.
The series always wavered in between straight-up cold war spy thriller and existential parable of individuality and liberty; I can't fault McGoohan for coming down hard on the latter side for the finale, even if it did skirt the edge of "psychadelic nonsense" as a result. If he'd gone the other way, we'd probably barely remember the show at all. ("Aha, Number Two is... A Soviet agent!" Snore.)
And even at its worst, The Prisoner had neither a Space Disco nor a Cute Robot Dog...
The important difference is that the original Battlestar Galactica was very, very, very, very bad. Not good. Terrible, in fact. Unwatchable. Occasionally downright embarrassing. Anyone who has fond memories of the original Galactica probably last watched it as a hyperactive 7-year-old. Changing the original series could only improve it.
The Prisoner, to put it mildly, does not suffer from this problem.
As far as the speed of the trial: this isn't, really, moving slowly at all. Unless one side caves and settles, or unless the judge grants summary judgement, it's not at all unusual for corporate contract disputes like this to stretch on for years.
The big problem is that before the court can make a finding of fact, the court has to determine what the facts are. This means something called "discovery", wherein the plaintiff and defendent both subpeona every single document, email, and backup tape that they think could be releavent, and depose (interview) every single person who could possibly have any insight into what happened. This part alone can take years. Then you have to present all of that evidence to the judge, complete with expert witnesses, cross-examination, and motions from both sides.
Just because the facts of this case appear "obvious" to you and me does not mean that a judge can or will decide the same. A big case (like the aforementioned IBM antitrust action) can take decades. Apple Computer and Apple Records have been suing each other on and off for nearly twenty years now.
IBM has its own history of antitrust entanglement, and I believe still operates under some of the consent decrees from that era. Launching a hostile takeover bid against a company that is current suing them for breach of contract would almost certainly invite a lot of the wrong kind of regulatory and legislative scrutiny.
Also: IBM has more money than god: as of their most recent financial statements, they maintain a cash balance of $8,250,000,000 (That's billion with a B, folks.) While the expense of fighting off SCO might seem insane to you or me, and would bankrupt a smaller company, it's literally pocket change to IBM, and it is very likely that their board of directors consider making an example of SCO to be a worthwhile investment.
Credibility in business contracts is a very important thing, and SCO is making a direct claim that IBM entered into a contract with them and proceeded to rob them blind. That's not a charge that any CEO is going to take lying down, and if they have to spend a few million dollars and a few years to have a judge authoritatively dismiss SCO's claim, it's probably worth it.
I am, to this day, probably more bitter about Aureal's end than about any other failed tech company other than the one I was personally involved with. (Don't ask, you've never heard of them.) I still have my MX300 in a drawer, on the off chance that someone with a soul gets an internship at Creative and leaks the driver source so that it can be updated for XP and Linux 2.6. And I will neverever buy a Creative product in my life: it's almost five years later and they still haven't managed to come up with a positional audio codec half as good as the one languishing in their vaults...
And dear lord am I ever enjoying watching Apple stomp Creative into bloody chunks in the DAP market. Couldn't happen to a nicer pack of thieves.
I did read the date of the article -- that's really, honestly, truly the last update on the situation.
The wheels of the court system grind really slowly, especially when the issue under consideration is a highly contentious piece of international trademark and contract litigation between two exceedingly well-funded companies that are ill-inclined to settle. The discovery phase of the trial alone could take years, and in this case almost certainly has.
(You'd think that the SCO-vs-Linux thing has been dragging on since 2003 would have clued people here in about this, but apparently not.)
...if very few people buy music from digital downloads according to this suit, then what the FUCK do these guys care what price Apple sells their music at?
Jake Gittes: I just want to know what you're worth. Over ten million? Noah Cross: Oh my, yes. Jake Gittes: Why are you doing it? How much better can you eat? What can you buy that you can't already afford? Noah Cross: The future, Mr. Gittes, the future.
These deals rarely work out the way you expect them to.
Sony Electronics corporation bought its way into the music and movie market, and everyone expected that this would mean Sony would apply their technology muscle and hustle to bring content to new markets faster.
But instead, the opposite happened: the hollywood side of Sony ate the rest of the company like a cancer. It's common knowledge among ex-SEL people that Sony had a "digital walkman" product ready to ship at least a year before the iPod, but that Sony Music's execs strangled it in the crib. It's said that Nobuyuki Idei still has to take sleeping pills because the sound of Steve Jobs cackling keeps him awake at night. But hey, Tommy Mottola got to do Mariah Carey on that big boardroom table from "Wall Street", so you win some and you lose some.
More seriously here: the problem is that music labels make INSANE amounts of money, and once you've acquired one, it's really, really, really hard to argue with the business practices of a PROFIT CENTER. Plus, even if you're a complete maniac like Steve Jobs and are personally willing to swing the axe around once you've acquired a major label... you can't do it, because a music label is, in the end, comprised of nothing but very canny salesmen with very expensively cultivated rolodexes, and if you make them uncomfortable, they will leave... and take their rolodexes with them. Then all you have in hand is a bunch of worthless contracted artists that nobody expected to get a second album out of in the first place, and nothing in the pipeline to replace them. Oh, and a very large bill for Cristal, Moet and cocaine. And some very pissed-off shareholders.
Apple has to play the cards in their own hand. Which, luckily for them, are pretty damn strong: hence Bronfman and Nash whining like pussies in the national press.
With respect...did you ever use WordPerfect for Windows?
Seriously the DOS versions of WP were brilliant, but WPfW was, for the first 2-3 incarnations, a nightmare: slow, crash-prone, visually confusing and basically less useful than its DOS predecessor.
By the time they finally got WPfW right, everyone who didn't have a substantial investment in WP's legal templates had converted to Word for Windows.
Agreed. People got very distracted over the whole "browser war" circa 1998-2000 because it was an obvious, visible instance of Microsoft abusing is monopoly power, to the point where I think a lot of people forgot how MS got its monopoly power:
Visual Studio on the development side, and Office on the applications side.
In a lot of ways, MS got very very lucky: if Borland, Novell, Lotus and WordPerfect/Corel hadn't spent an entire decade shooting themselves in the foot over and over again, the competitive landscape in 2005 would be a very different place.
1) It's a bitch to install. Won't even compile on modern Linux distributions. You have to patch it to compile it and the patch isn't even hosted on qmail's site.
Look, it's annoying that Bernstein and the GLIBC authors have decided to take their mutual pissfest out on us, but "echo gcc -O2 -include/usr/include/errno.h > conf-cc; make" is not exactly going to kill you, now is it?
If it is going to kill you, there's always the net-qmail distribution.
2) It's a bitch to configure. Rather than parsing a single configuration file, qmail relies heavily on the presence of individual files in a directory.
A matter of taste, I guess. The single-config-per-file method makes it very easy to build kickstart/rpm profiles that add or remove certain features without having to carefully parse/edit a monolithic configuration file, but I can see how for a junior sysadmin it's a little more confusing than just "look in main.cf."
3) Not not not not scalable! That's a myth. Doesn't properly batch jobs together. Hell! qmail was originally designed to be run from inetd!
You really have no idea what you're talking about, do you? (Hint: qmail isn't sendmail, and qmail-smtpd isn't "qmail" any more than inetd is "unix".)
4) Heavy reliance on other daemontools.
You can use daemontools to manage qmail if you want to. It's not a requirement, and the official docs don't even suggest it.
5) Breaks well-known and understood UNIX standards.
Really? Which ones?
6) Security through lack-of-functionality.
It's an MTA. It transports mail. Securely, as it happens. This is a feature, not a bug.
7) Not really secure despite the claims.
Really? Care to enlighten us?
8) No longer maintained.
This is as close to an actual valid complaint as you've got here: it's certainly been a good long time since the last release. And yet, it still works.
9) No features. Adding them requires patching, and patching, and more patching.
Look, if you need an MTA that speaks LDAP, SQL and UUCP, has hooks into an integrated calendar, and polishes the bumpers on your car, it's probably true that qmail is not the tool you want to use. Have fun trying to manage whatever monstrosity it is that does.
It does one thing, and it does that one thing extremely well: some of us still consider that to be a virtue.
Serious sysadmins don't use qmail and for damn good reason.
I rather doubt you'd recognize a serious sysadmin if one bit you.
You're high. Building a massive production email system on Sendmail 9 is slow-motion suicide. If the security holes don't get you, the terrible configuration methods and complete lack of scaleability will, nevermind the fact that Sendmail Inc is trying desperately to replace the product.
"Most managable with [...] heavy customization?" I'd laugh if I wasn't crying. And I'm crying because I used to work for a company that deployed a massively customized sendmail infrastructure -- and I was one of the poor bastards who had to maintain it. Trust me, you don't want to do this. Ever.
Yes, milter is cool. No, it's not cool enough to justify burning CPU cycles on sendmail in 2005.
Even Sendmail Inc tacitly admits that Sendmail's design is garbage: take a look at the design document for Sendmail X, and note carefully how much it resembles Postfix and Qmail. There are very good reasons for this.
While it's true that MySQL 4 and 5 support "clustering" out of the box, it's actually a somewhat misleading statement. MySQL's clustering support is actually an entirely different database engine from MyISAM or InnoDB, much in the same way that MaxDB is. Cluster is actually something called NDB (Network DataBase), which Mysql AB acquired from a european telco (I think Vodaphone, but don't quote me on that) and is slowly integrating into the MySQL codebase.
NDB has some interesting design features, but it is not a plug-and-play HA solution for MySQL. It imposes some very serious limitations on queries and datatypes -- differences that basically make it a worst-case-scenario subset of the limitations of MyISAM and InnoDB. Additionally, running NDB requires running a number of specialized daemons that are administered in a substantially different fashion than the rest of mysqld, and which are, to put it charitably, a bit on the undocumented and flaky side.
Mysql AB appears to be working pretty hard to integrate NDB into their codebase, and I expect that in another 2 or 3 years it will be a very compelling solution. But for now, it's an interesting toy and nothing more.
Slashdot Mirror
I know that, you know that, but it's very obvious from the article that Cringely has no clue whatsoever. He still thinks that peering points are some weird rarified animal where only telcos with enormously expensive routers can play, when the reality is that anyone with $3000/month to spend can host their equipment "near the core."
(And I find the idea of Google parking a tractor trailer with their hypothetical container-box node on the street next to 111 8th ave particularly amusing. Nevermind heat dissipation issues: can they afford the parking tickets?)
No, no, no. This is exactly my point: it won't buy them anything in terms of reduced lag, because most internet traffic doesn't go over public peering exchanges any more. Putting a Google node at MAE-East would be substantially less effective than just renting a dozen racks at Equinox and ordering a fistful of 10-gigabit crossconnects to whatever carriers are available there.
Even assuming the power and heat requirements of cramming that many opterons into that small a space could be dealt with, there's another, larger problem:
It's not fucking 1997 any more.
"Peering points" -- big, open-access traffic exchange handoffs like the old MAE-East and MAE-West used to be a big deal back in the late 90s, when OC-12 circuits were still rare and hideously expensive beasts, and Gigabit Ethernet was still a gleam in some 3Com engineer's eye.
In 2005, they simply don't matter. The big players (level3, MCI/Verizon, Qwest, SBC, etc) all exchange traffic over private fiber interconnects, and everyone else buys transit from the big guys directly or ponies up for a switch port at Equinox, PAIX/Switch&Data or some other 'carrier neutral' colocation center. Dropping a datacenter-in-a-box onto MAE-east or any of its surviving ilk would buy Google precisely nothing.
(And nevermind the fact that google is documented to own thousands upon thousands of unused square feet of datacenter space already: they went on a very well-thought-out buying spree in 2000-2001 when all the dot-com datacenter companies were going out of business, and are very well provisioned for the forseeable future as a result.)
Now, a much more interesting application of the "Google node in a shipping container" idea can be summed up in one simple word: China. Why wait for the local market to develop the infrastructure you need when you can just drop a box down and then run fiber to it? I'm still dubious though...
Although I think remaking The Prisoner is an awful idea, SkyTV's involvement doesn't necessarily make it worse. Just because they're part of News Corp doesn't mean that they're the sample people who produce "Cops".
Sky is probably best known in the US for co-bankrolling Ron Moore's remake of Battlestar Galactica with SciFi Network, which I believe at least one or two people around these parts liked. (This is why the first season of Galactica aired first on Sky -- they got first-airing rights as part of the deal.)
Mmmmph. I won't make a blanket defense of "Fall Out" -- it's certainly not flawless, and doesn't necessarily hold up that well today -- but I tend to think of its flaws as artifacts of its era, rather than as outright artistic failures.
The series always wavered in between straight-up cold war spy thriller and existential parable of individuality and liberty; I can't fault McGoohan for coming down hard on the latter side for the finale, even if it did skirt the edge of "psychadelic nonsense" as a result. If he'd gone the other way, we'd probably barely remember the show at all. ("Aha, Number Two is... A Soviet agent!" Snore.)
And even at its worst, The Prisoner had neither a Space Disco nor a Cute Robot Dog...
The important difference is that the original Battlestar Galactica was very, very, very, very bad. Not good. Terrible, in fact. Unwatchable. Occasionally downright embarrassing. Anyone who has fond memories of the original Galactica probably last watched it as a hyperactive 7-year-old. Changing the original series could only improve it.
The Prisoner, to put it mildly, does not suffer from this problem.
"Larry"
Ah, okay, I misread your post.
As far as the speed of the trial: this isn't, really, moving slowly at all. Unless one side caves and settles, or unless the judge grants summary judgement, it's not at all unusual for corporate contract disputes like this to stretch on for years.
The big problem is that before the court can make a finding of fact, the court has to determine what the facts are. This means something called "discovery", wherein the plaintiff and defendent both subpeona every single document, email, and backup tape that they think could be releavent, and depose (interview) every single person who could possibly have any insight into what happened. This part alone can take years. Then you have to present all of that evidence to the judge, complete with expert witnesses, cross-examination, and motions from both sides.
Just because the facts of this case appear "obvious" to you and me does not mean that a judge can or will decide the same. A big case (like the aforementioned IBM antitrust action) can take decades. Apple Computer and Apple Records have been suing each other on and off for nearly twenty years now.
Three years is nothing.
IBM has its own history of antitrust entanglement, and I believe still operates under some of the consent decrees from that era. Launching a hostile takeover bid against a company that is current suing them for breach of contract would almost certainly invite a lot of the wrong kind of regulatory and legislative scrutiny.
Also: IBM has more money than god: as of their most recent financial statements, they maintain a cash balance of $8,250,000,000 (That's billion with a B, folks.) While the expense of fighting off SCO might seem insane to you or me, and would bankrupt a smaller company, it's literally pocket change to IBM, and it is very likely that their board of directors consider making an example of SCO to be a worthwhile investment.
Credibility in business contracts is a very important thing, and SCO is making a direct claim that IBM entered into a contract with them and proceeded to rob them blind. That's not a charge that any CEO is going to take lying down, and if they have to spend a few million dollars and a few years to have a judge authoritatively dismiss SCO's claim, it's probably worth it.
I am, to this day, probably more bitter about Aureal's end than about any other failed tech company other than the one I was personally involved with. (Don't ask, you've never heard of them.) I still have my MX300 in a drawer, on the off chance that someone with a soul gets an internship at Creative and leaks the driver source so that it can be updated for XP and Linux 2.6. And I will neverever buy a Creative product in my life: it's almost five years later and they still haven't managed to come up with a positional audio codec half as good as the one languishing in their vaults...
And dear lord am I ever enjoying watching Apple stomp Creative into bloody chunks in the DAP market. Couldn't happen to a nicer pack of thieves.
I did read the date of the article -- that's really, honestly, truly the last update on the situation.
The wheels of the court system grind really slowly, especially when the issue under consideration is a highly contentious piece of international trademark and contract litigation between two exceedingly well-funded companies that are ill-inclined to settle. The discovery phase of the trial alone could take years, and in this case almost certainly has.
(You'd think that the SCO-vs-Linux thing has been dragging on since 2003 would have clued people here in about this, but apparently not.)
...if very few people buy music from digital downloads according to this suit, then what the FUCK do these guys care what price Apple sells their music at?
Jake Gittes: I just want to know what you're worth. Over ten million?
Noah Cross: Oh my, yes.
Jake Gittes: Why are you doing it? How much better can you eat? What can you buy that you can't already afford?
Noah Cross: The future, Mr. Gittes, the future.
(Forget it, Jake. It's Chinatown...)
These deals rarely work out the way you expect them to.
Sony Electronics corporation bought its way into the music and movie market, and everyone expected that this would mean Sony would apply their technology muscle and hustle to bring content to new markets faster.
But instead, the opposite happened: the hollywood side of Sony ate the rest of the company like a cancer. It's common knowledge among ex-SEL people that Sony had a "digital walkman" product ready to ship at least a year before the iPod, but that Sony Music's execs strangled it in the crib. It's said that Nobuyuki Idei still has to take sleeping pills because the sound of Steve Jobs cackling keeps him awake at night. But hey, Tommy Mottola got to do Mariah Carey on that big boardroom table from "Wall Street", so you win some and you lose some.
More seriously here: the problem is that music labels make INSANE amounts of money, and once you've acquired one, it's really, really, really hard to argue with the business practices of a PROFIT CENTER. Plus, even if you're a complete maniac like Steve Jobs and are personally willing to swing the axe around once you've acquired a major label... you can't do it, because a music label is, in the end, comprised of nothing but very canny salesmen with very expensively cultivated rolodexes, and if you make them uncomfortable, they will leave... and take their rolodexes with them. Then all you have in hand is a bunch of worthless contracted artists that nobody expected to get a second album out of in the first place, and nothing in the pipeline to replace them. Oh, and a very large bill for Cristal, Moet and cocaine. And some very pissed-off shareholders.
Apple has to play the cards in their own hand. Which, luckily for them, are pretty damn strong: hence Bronfman and Nash whining like pussies in the national press.
The Apple vs Apple deal is currently back in court again, for reasons that have everything to do with the iPod and the iTMS.
I'm honestly surprised that Apple Computer hasn't simply made a cash tender offer for Apple Records at this point.
With respect...did you ever use WordPerfect for Windows?
:)
Seriously the DOS versions of WP were brilliant, but WPfW was, for the first 2-3 incarnations, a nightmare: slow, crash-prone, visually confusing and basically less useful than its DOS predecessor.
By the time they finally got WPfW right, everyone who didn't have a substantial investment in WP's legal templates had converted to Word for Windows.
Yes, I'm still bitter about this.
Visual C++ 1.0 shipped in 1993. True, it didn't get rebranded as "Visual Studio" until much later, but that's, well, branding.
Visual Studio on the development side, and Office on the applications side.
In a lot of ways, MS got very very lucky: if Borland, Novell, Lotus and WordPerfect/Corel hadn't spent an entire decade shooting themselves in the foot over and over again, the competitive landscape in 2005 would be a very different place.
The set currently implemented by the freebsd qmail port might be a good start...
(hi russ!)
1) It's a bitch to install. Won't even compile on modern Linux distributions. You have to patch it to compile it and the patch isn't even hosted on qmail's site.
/usr/include/errno.h > conf-cc; make" is not exactly going to kill you, now is it?
Look, it's annoying that Bernstein and the GLIBC authors have decided to take their mutual pissfest out on us, but "echo gcc -O2 -include
If it is going to kill you, there's always the net-qmail distribution.
2) It's a bitch to configure. Rather than parsing a single configuration file, qmail relies heavily on the presence of individual files in a directory.
A matter of taste, I guess. The single-config-per-file method makes it very easy to build kickstart/rpm profiles that add or remove certain features without having to carefully parse/edit a monolithic configuration file, but I can see how for a junior sysadmin it's a little more confusing than just "look in main.cf."
3) Not not not not scalable! That's a myth. Doesn't properly batch jobs together. Hell! qmail was originally designed to be run from inetd!
You really have no idea what you're talking about, do you? (Hint: qmail isn't sendmail, and qmail-smtpd isn't "qmail" any more than inetd is "unix".)
4) Heavy reliance on other daemontools.
You can use daemontools to manage qmail if you want to. It's not a requirement, and the official docs don't even suggest it.
5) Breaks well-known and understood UNIX standards.
Really? Which ones?
6) Security through lack-of-functionality.
It's an MTA. It transports mail. Securely, as it happens. This is a feature, not a bug.
7) Not really secure despite the claims.
Really? Care to enlighten us?
8) No longer maintained.
This is as close to an actual valid complaint as you've got here: it's certainly been a good long time since the last release. And yet, it still works.
9) No features. Adding them requires patching, and patching, and more patching.
Look, if you need an MTA that speaks LDAP, SQL and UUCP, has hooks into an integrated calendar, and polishes the bumpers on your car, it's probably true that qmail is not the tool you want to use. Have fun trying to manage whatever monstrosity it is that does.
It does one thing, and it does that one thing extremely well: some of us still consider that to be a virtue.
Serious sysadmins don't use qmail and for damn good reason.
I rather doubt you'd recognize a serious sysadmin if one bit you.
Um, what kind of company would spend $311 million to buy PalmOS, then STOP SELLING THE TECHNOLOGY?
Uh, I hate to burst your bubble, but that happens all the time in this industry. Just ask anyone who ever worked for Aureal.
(er, where by "9" I mean "8" -- damn fingers)
All of these systems will be running sendmail.
You're high. Building a massive production email system on Sendmail 9 is slow-motion suicide. If the security holes don't get you, the terrible configuration methods and complete lack of scaleability will, nevermind the fact that Sendmail Inc is trying desperately to replace the product.
"Most managable with [...] heavy customization?" I'd laugh if I wasn't crying. And I'm crying because I used to work for a company that deployed a massively customized sendmail infrastructure -- and I was one of the poor bastards who had to maintain it. Trust me, you don't want to do this. Ever.
Yes, milter is cool. No, it's not cool enough to justify burning CPU cycles on sendmail in 2005.
Even Sendmail Inc tacitly admits that Sendmail's design is garbage: take a look at the design document for Sendmail X, and note carefully how much it resembles Postfix and Qmail. There are very good reasons for this.
We new univers fans DEMAND equal representation in this new film deal! ...both of you?
While it's true that MySQL 4 and 5 support "clustering" out of the box, it's actually a somewhat misleading statement. MySQL's clustering support is actually an entirely different database engine from MyISAM or InnoDB, much in the same way that MaxDB is. Cluster is actually something called NDB (Network DataBase), which Mysql AB acquired from a european telco (I think Vodaphone, but don't quote me on that) and is slowly integrating into the MySQL codebase.
NDB has some interesting design features, but it is not a plug-and-play HA solution for MySQL. It imposes some very serious limitations on queries and datatypes -- differences that basically make it a worst-case-scenario subset of the limitations of MyISAM and InnoDB. Additionally, running NDB requires running a number of specialized daemons that are administered in a substantially different fashion than the rest of mysqld, and which are, to put it charitably, a bit on the undocumented and flaky side.
Mysql AB appears to be working pretty hard to integrate NDB into their codebase, and I expect that in another 2 or 3 years it will be a very compelling solution. But for now, it's an interesting toy and nothing more.
(Which is to say, I shouldn't have used the word "won" in my original reply up there. My bad.)