Slashdot Mirror
Sun Chief Calls Out IBM, Demands Compatibility
downbad writes "Sun's President, Jonathan Schwartz, yesterday published an Open Letter to the CEO of IBM, Sam Palmisano, in which he alluded to "behavior reminiscent of an IBM history many CIOs would like to forget" - a reference to Sun's frustration that IBM isn't supporting Solaris 10 with WebSphere, DB2, Tivoli, Rational and MQSeries products. In his "Dear Sam" letter - circulated via his blog - Schwartz refers first to the "long history of partnering" between Sun and IBM, and claims Sun customers have made repeated calls to IBM about having the choice to run IBM products on Solaris 10." *cough* Kettle, meet Pot.
So... if you make Solaris compatible with Linux won't this solve the problem somewhat?
circulated via his blog
gotta love that
Don't be a looter...and yes, I know that it's spelled with an "A" instead of an "E".
So...he's not a black pot, then?
...well that makes everything better, then, doesn't it?
It is quite hypocritical of Sun to be saying this when so little of their software runs on anything but Solaris.
A battle between two for-profit companies over market share adn interoperability. Why is this a headline on Slashdot? I mean, we're all commie pinko Open Source advocates that work soley for charitable organizations.
Was that necessary?
Sun doesn't make all that many software products that aren't OS-type products. Off the top of my head, I can think of one big product they've made -- Java -- and they seemed to try to make it available on all platforms, though based on their rules (which hey, is true for any GPL-based software also. It's all about letting the people who created the software determine how it's released).
It is, however, a little offensive to publicly decry a company not releasing their product on your platform, especially when that platform hasn't yet actually shipped its first non-beta version. Seems a little petulant.
...has made a number of bad business decisions in the last couple decades, most notably selling their souls to Windows/DOS, something they've paid for (literally) in the long run.
This is where I stopped reading.
If I want Solaris professionally, I'll buy a SPARC to run it on. If I want to play around with Solaris, I'll download it for x86.
Allen Zadr is the Director of IT for a small software company
Kinetic stupidity has a new brand leader: Allen Zadr.
Maybe their delaying as the shake-out over the SCO fiasco takes place. Heck it maybe shown that SUN actively supported the attack (if it hasn't been already).
Virals Suck
Kettle, meet Pot. Except in NEBRASKA!!
It's the most secure OS the world has ever seen
Huh? Since when? I think someone's tooting his own horn. But anyway, this blog is mostly just an indignant "pretty please help us", offering silly remarks whilst asking what's pretty simply a favour. I don't see why this should even make slashdot.
No, Jonathan, IBM won't switch from Linux to Solaris just because you got the OSI to look at your license. Perhaps make it GPL compatable and try the FSF and see if it'll pass there - then I at least might listen.
I hope IBM's response is "make Solaris GPL compatable and we'll support it too, once we merge it with Linux"
If people would use PostgreSQL instead of DB2 and Jetty/JBoss (or other free alternatives) instead of Websphere they could run their apps on just about any OS. Or if they used a free OS, particularly one supported by IBM, they could run their proprietary IBM software. Or run free software on a free OS and be ready for anything.
Let's see - will Sun be making the API for their new file system's extra-special features available so that other *nix OSs can support it with their own native file systems?
No?
Well, will Sun make their new file system available for other *nix OSs?
No?
Well, will Sun have ANY compatibility between Solaris with their new, all-signing-all-dancing file system and any other OS?
No?
Then to Sun I say - "SHUT THE FBOMB UP ABOUT OTHER PEOPLE'S COMPATIBILITY UNTIL YOU ARE COMPATABLE YOURSELF!"
www.eFax.com are spammers
Help us make money and give your customers an alternative to your products.
Thanks,
Sun
.. like griping that M$ does not produce a versions of it's Games, Office suite, Visio toos etc. for Linux? With IBM backing Linux why should they support Solaris? Corporate Wolf bites Corporate Coyote...
Only to idiots, are orders laws.
-- Henning von Tresckow
duh.
Kinetic stupidity has a new brand leader: Allen Zadr.
I wonder what the OpenBSD folks have to say about this...
Also, I haven't had the need to run any IBM software on my BSD (or anywhere till now), so maybe he's just making things up about about the "repeated" calls and so on. Or is he?
At this point, it makes more sense for IBM to port their applications to OS X. Now that they supply the CPU for Apple's server hardware, there's a strong case to be made for this.
If WSAD were ever ported to OS X, my boss would be placing a nice order for xServes and powermacs on the Apple website.
I'm pretty sure Schwartz is the COO not CEO, unless McNeely's finally been overthrown.
Of course, IBM still has strong roots as a "hardware" company. What's IBM's incentive to rewrite their software (little profit) on Sun's hardware (no profit)? Not a whole lot of incentive there.
[
...how many Sun OS products I can run on my z-series mainframe...?
The Kai's Semi-Updated Website Thingy
Apparently you have not read your own literature. I refer you to the web page at
http://www.sun.com/2004-0803/feature/
In which you state:
"3.Aug.04--Customers who want the stability and security of the Solaris Operating System and the flexibility to also use Linux applications won't have to wait much longer. The forthcoming Solaris 10 Operating System (OS) will include a remarkable new feature that allows customers to run Linux applications unchanged on the Solaris OS. By enabling this functionality, code-named Project Janus, administrators can create an environment for running a range of Linux applications at near-native speeds. Sun is offering Project Janus as an optional kernel service of the Solaris OS, enabling administrators to run Linux applications in a new and unique way on x86 platforms. In keeping with Sun's long-standing support of industry standards, Project Janus is designed for compliance with the Linux Standard Base specification.
Ergo, if your version of *Nix was as compatible as you claim, there is no issue at all.
Thanks for taking the time to write, and while I have your attention, how are efforts to open Java for improvements by the open source community coming?
Signed,
IBM
Posted ... January 24, "yesterday published an Open Letter" and at OSnews http://osnews.com/comment.php?news_id=9478/ ... on 2005-01-22
Posted
"Steve Jobs invented the world" -- Bill W. GATES
Just as soon as you Open Source Java.....
.... forever.
*bump*.
This guy Schwartz has a mouth big enough to put an entire army of PR handlers into fat city
Go tree-huggers.
Interesting how Schwartz titles his missive as an "open letter". Hah. Schwartz' letter is the only thing "open" about Sun. How about opening Java?
(I work for IBM, though my opinion does not reflect my employer)
Should run on Solaris but i'm not sure if this is just on Solaris 10 x86 or whether it applies to sparc too.
However even if you can run IBMs linux binaries on Solaris that doesn't mean they are "supported" it only means that they work.
I mean Sun Created Java and they don't have a credible Java app server (I know iPlanet with it's whopping 2% market share) the big boys are Wepshere, BEA, Oracle and Jboss......
If Sun wants all this then they need open up java, and try to make Solaris more compaible with 3rd party products (JBoss anyone). It's more than hypocritical, it shows there is some desperation on Suns part. The Ultra-Sparc line is Ultra Slow and Ultra priced. If IBM were to start turning out PowerPC based Risc Boxes running Linux, would Sun even be relevant? I know about all of Solaris's great OS features, but how long will it take Linux to catch up? Especially with the other big boys pushing linux.
Now add to that these new Cell CPU's IBM & Sony are making. A Linux Server with a big cluster of Cell processors, Sun Who??
So Long and Thanks for all the Fish.
Is there going to be some sort of dance rumble for street cred?
The surprise isn't how often we make bad choices; the surprise is how seldom they defeat us.
can the very first post be redundant?
emt 377 emt 4
just say please :)
or else junst mention IBM's application alternatives like apache, mysql and so forth...
maybe even java? heh
Ok... it its binary compat as the letter states, why is a simple recompile required?
/\/\icro/\/\uncher
Okay, I deserverd that nearly instantaneous Troll mod. It's Monday morning, and I apologize for being a bit cranky.
I was once able to toot my own horn. Ah, how I yearn for my more able years as a flexible young lad...
Schwartz has to throw his weight around while he can, as he drives his company into the Sunset. Maybe if letterware like this produces enough derision in the industry, he'll get the axe faster than you can say "Gilbert Amelio". And Sun might have a chance to churn Solaris tech into the kind of superstable Linux that IBM produces by hybridization with AIX.
--
make install -not war
I think they are getting hurt again from the time Sun tried to sell x86 Solaris machines and then abandoned ship without so much as a good bye we're sorry.
This has seriously pissed off Intel which has since been making trying to beat Sun into hamburger. Maybe some companies have a long memory. Strange as that may seem.
sri
You stole my customers ! Prepare to die !
Sure, call me a troll, but Schwartz hasn't exactly been known as a straight shooter, or industry visionary. They (Sun PR) have tried desperately to make him out to look like a visionary (the blog is the most blatant attempt). But, from what I hear from friends inside Sun, the guy is pretty much thought of as a tool.
Don't know why he's attacking IBM. IBM is his friend. No, really.
If IBM were to start turning out PowerPC based Risc Boxes running Linux,
Where have you been?! IBM has been touting LINUX along with AIX5L (thats what the "L" stands for!) for over two years.
P.S.- we played around with Sun's app server... it blows. Chunks.
In the future, I would want to not be isolated from my friends in the Space Station.
Dear Jonathan,
What about Java for ppc?
"My OS is becoming irrelevant! Lots more stuff runs on Linux! Save me, IBM!!!"
Seriously, IBM will port their software if they can make more money selling the Solaris versions than it cost to port and support. That's it.
IBM may show largesse toward open source, but that's because they view it as strategic. Solaris isn't strategic for them, no matter how much Schwartz may wish it so.
Garg
Garg
Alumnus, Xavier's School for Gifted Youngsters
Jonathan is complaining about IBM not _supporting_ those products in Solaris 10, not whether the run on Solaris 10 or not, I'm pretty sure they do. If I were IBM, I would not publicly support a competitor's beta product either.
Maybe, but I doubt it. They gave the same sort of song and dance about 2.8 upgrades as well and it didn't really work that way. They promise even more with Java and also fail to deliver 100% complete portability.
"iBM iz such loosers. i not let them on my Freinds list. ha ha! i tell them to port DB2 and i might think about it. Loosers. Sam will not tell me my poetry sux N E more now. i write what i feel. i want to ask the Q T girl out but i am shy. Maybe i will send her some of my poetry. iBM better port WebSphere 2 or i still keep them off my Freinds list. ha ha! Take that you loosers!"
IBM should simply say, "We don't feel it in our best interest to begin supporting a dying product.", and leave it at that.
Table-ized A.I.
I had issues with Apple not supporting other vendors formats and media formats as well with iTunes/iPod. In fact, specifically targeting other companies that were compatible and actively and openly shutting them off with unrelated firmware upgrades. I read all the replies on /. about how Apple choosing to force people into their products was good for the consumer and anyone who wanted compatibility was stupid. I'm confused on the change of thought here.
As expected, this will be marked as troll because it is anti-Apple. Instead of moderating, how about a discussion about this instead of trying to hide it from the others?
Yes, that was quite necessary. Contrary to your statement, Sun makes lots of software products that aren't OS-type products. In fact, Sun has lots of huge buildings filled with people trying to make and sell these products.
The problem is, they almost all suck, and so no one has heard of these products.
For starters, there's Sun's glorious development tools (C compiler and make, etc). These are marginally useful, but they are hobbled by Sun's incredibly stupid Flexlm license manager. For EVERY single cc call in a Makefile, you have to go out and check the license. Performance goes into the tank.
This is one reason why gcc is popular within Sun. And I note that even the final Solaris build process has figured out a way to avoid this hinderance, which isn't used by Suns' customers.
Also, one has Teamware, which is where BitKeeper came from (both developed by the same guy). Teamware would be a contender in the marketplace, if it wasn't hobbled by the stupid Flexlm licensing as well. IMO, it is far superiour to Perforce and that other one (starts with a C IIRC - can't remember it offhand).
And if any fanatics dispute that, all I have to say is look at it in practice. HP, last I heard, had 2 dozen people supporting their OS builds, just because of their stupid build environment. Sun has designed their builds right, and needs about 1/10th the resources.
the fact that we funded 50% of sco's attack on you, please disregard that. We decided to buy a $25 million dollar license for the first time last year. The fact that we never bought a license prior to their confidential inquiries and confidential plans to sue you that they showed us as a basis for us to take the license...disregard that also.
We had nothing to do with the lawsuit (other than funding them just enough to last several years in court), it was a license for their code. Really.
Thanks again,
Sun.
...IBM has never issued letters of support for a non-IBM platform which does not yet exist. Why should that be different for Solaris 10?
I guess today is a passable day to die.
...the Product Managers of the named products would probably very much like to support Solaris, but can't get approval from their superiors to purchase any Sun h/w or s/w licenses to compile and test on.
Of course an unreleased OS is the most secure ever - it's never been exploited! Even OpenBSD can't say that...
Slashdot - Mutual Assured Discussion
Wow. So Sun is releasing a port of OpenBSD :)
This guy is way out there
Then we have Sun CEO Scott McNealy complaining before congress in 2000 that, "We already half way through the fiscal year, capped out on the number of really bright Israelis and Indians.". He gets more and more H-1b visas allocated.
Then we have Sun's stock going from above $60/share to below $3/share.
And now Sun is complaining about something else and we're supposed to consider this "news that matters"?
Seastead this.
I've dealt with IBM Software in the past. They typically lag behind in their "official" supported platforms because they need to go through a lot of tests to validate their software works as designed. When I've run into issues like this they simply say "it may work, but we haven't tested it enough yet".
That is why they pick a flavor (or two) of Linux as supported instead of saying "we support Linux". Other distros will probably work, but they only have so much time to validate & test. For a long time WebSphere (at least on z/Series hardware) was only supported on a 2.2 kernel. It ran fine on 2.4, but it wasn't officially supported.
That being said, if you do have a problem and you have a support contract IBM will work with you to solve the issue, but they don't like to make gurantees about unsupported hardware / software interacting with their stuff.
Their Java support for linux has been stellar. Likewise, their backing of Linux on Sparc has also been wonderful. I can call them up, and get full backing on their systems.
What do you mean no?
I prefer the "u" in honour as it seems to be missing these days.
Schwartz refers first to the "long history of partnering" between Sun and IBM, and claims Sun customers have made repeated calls to IBM about having the choice to run IBM products on Solaris 10.
There are several possible logic interpretations of that foggy statement:
possibility #1: Some Sun customers found Solaris lacking features and wanted to run those same superior IBM solutions as do their competitors.
possibility #2: Some IBM customers wishing to defect to Solaris on cheapo hardware, but are not sufficiently brave to say that to IBM eyes to eyes. So they used Mr. Schwartz as herold.
possibility #3: Mr. Schwartz tries to save dying Solaris desperately by generating some media PR/FUD at all costs. Slashdot comes handy for such task, as usually.
possibility #4: Sun is negotiating secretly with IBM about monetary compensation for future opening of Java. Mr. Schwartz suggests they would accept some customers instead of cash. On really big markets, customers are commodity, you know...
possibility #5: Mr. Schwartz used to write on his blog after a very long long drinking party.
possibility #6: All of the above.
There you are, staring at me again.
I stand corrected:x run/
http://www.sun.com/software/linux/compatibility/l
Umm... Solaris 10 is compatible with Linux. You run binaries on x86 just fine. (You can also run binaries just fine on FreeBSD, NetBSD, or OpenBSD as well.)
IBM to SUN: We want you to Open Source Java
...
Sun to IBM: Get Bent we are keeping it to ourselves
Few months goes by
SUN to IBM: Please oh Please make your stuff run
on slowarus.
IBM to SUN: Get Bent we are keeping it to ourselves.
See how simple things are!
Got Code?
PS. And please put a frikking Windows key on your ThinkPads.
Solaris is now an obscure niche platform. IBM doesn't support VAX VMS either, but no one complains. If it made financial sense for IBM to support Solaris they would. This rant is only a futher sign of the waning of Sun in the marketplace.
an ill wind that blows no good
Sun is spreading itself too thin. They support Linux, Solaris, Java, Windows and one or two others. No one wants to admit that Unix is dead. Given the choice, people run Linux, and then maybe a VM session of a Unix (inside of Linux).
...desperately want to run as little Sun software as possible! Help us, IBM!
Seriously, does anyone take Sun seriously anymore? Bad prices, complete sellout, ancient repackaged ugly technology, and second-rate hardware? Sign me up!
adam b.
Java
OpenOffice?
Fuck off and die!
Sun is getting pretty desperate. Are they gonna pull a SCO next?
I hate sigs.
Schwartz' letter smacks of someone who is desparate and who thinks that Sun's hardware business will shrivel and die. In Schwartz' mind, in order to survive, Sun must grow its software business by using the Microsoft model. Namely, Sun locks people into Solaris, then with a captive audience, Sun charges monopoly profits for Solaris upgrades and exclusive Sun applications like StarOffice.
The big clue is the disappointing financial report from last quarter for Sun. Revenue actually fell. Sun hardware is simply not competitive with IBM hardware or, even, HP hardware. Since Sun still largely depends on hardware revenue, Sun must quickly ramp up its sotware business and service business. Otherwise, Sun as a company will die along with its hardware division.
Note that Sun is expected to lose money in the remaining 2 quarters of its fiscal year.
For whom does the bell toll? It tolls for Sun.
Why man? It wasn't even a good troll. Wear your crakiness with a little lust for life.
I'm still waiting for the "+1 Troll" moderation option.
Something so breathtakingly forbidden people will be dumbstruck upon reading.
"No Indy, covor yar heaart."
Careful with the kettle/pot reference ... before you know it, Omarosa's going to accuse you of saying the N-word!
He gotta be kidding... the last trick (well actually not the last...) sun did was to support SCO in their attack on IBM, I say, mod him funny or is he just trolling?? I'm sure someone here already wrote this, but I am to lazy to read through the posts.
" How many of IBM's platforms does Java run on? None."
All of them. Name a platform of IBM's that you can't run Java on.... MVS, TPF, AIX, OS/2, AS400....hell.... they all run Java.
Maybe you meant something else? Or you had a typo?
http://shit.slashdot.org/article.pl?sid=05/01/24/1 430235
Although sun has many products that run on Linux, overall I view them as an enemy to the Linux community. Thats just my opinion. They have this 'Now we love Linux, now we don't' attitude that bothers me and they are also butt buddies with Microsoft and SCO now so, to me, they can't be trusted.
I appeciate Java and OO but those apps running on Linux help Sun just as much as they help Linux,if not more. It just seems to me that everything they do has some alterior motive.
IBM isn't a service company, they're a sales company. They'll sell anything if there's a buck in it.As for their software business, it's incredibly profitable. Websphere goes for millions of dollars and suckers line up to buy it.
Don't buy IBM's propaganda, they're a big evil closed-source software company just like the others.
Cheers Koz
"Games, Office suite, Visio toos"
That implies that there is demand for those products
Yeah, come on IBM why don't you support Solaris like Java supports Free BSD?
The end.
Regards, Picky
Richard Moore of IBM's Linux Technology Center has already provided a dynamic probe and trace capability that was developed by IBM. Several papers and presentations are here
when did Schwartz become CEO of Sun? did McNealy get ousted without me knowing?
... that if Sun's executives stopped writting such long and such a prolific amount of mindless and pointless blog entries and more time spent RUNNING THE DAMN COMPANY that Sun might be doing better than it is?
Mac OS X and Windows XP working side by side to fight back the night.
Whoever modded this interesting is a complete moron.
Leaks in third-party APIs? Holy shit. This has nothing to do with Java.
There are numerous benchmarks showing Java faster than C and C++ with run-time optimization.
You sound like a bad developer blaming the platform. Perhaps you need a career change?
sun, they are so irrelivant they don't even know it yet.
I have no idea who Sam Paisano is, on the other hand.
The original URL is gone, but its contents might be archived someplace.
The that were not available to you in the old physical world geographic dislocation was just what I think Australia is not necessarily a big disadvantage. This is you ought to be net importers of brilliant people. I am fighting with our government to allow H1B visas cap to be raised. I was in at the White House talking to the chief of staff to get the H1B visa cap raised. We already half way through the fiscal year, *capped out on the number of really bright Israelis* and Indians.
I look at Sun started by an Indian and German immigrant. Lord knows we do not want more Sun Microsystems in the United States. Java invented by a Canadian. Oliver Spark? technology was invented by Chinese immigrants and other Asian immigrants. Lord knows more of those folks coming in there and creating all this value. We are bringing them into the bay area where fixer upper park bench is $US4½ million or a five block commute is your lunch hour. You should all be absolutely doing a brain drain on the rest of us if half managing knew about Sydney which is a warmer San Francisco with no traffic and that 50c in the dollar, I would lose my whole team. I am telling them Sydney is an awful place.
My own sense, McNealy is just another corporate welfare queen buying legislation trying to keep his sick company alive. Personally, I really don't want another Sun Microsystems in the United States if it means giving away immigration rights worth _billions_ of dollars free of charge. If McNealy was such a value generator he could pay fair market value for those immigration rights it might be one thing-but he can't.
Dear IBM:
Please buy us so I can have job in 12 months.
The text mistakenly says Schwartz is CEO of Sun, he's "only" COO.
Database engine for analyzed or annotated text
Sun says that customers are demanding it, but I wonder how real that demand is...
I know for a fact that Sun have sent letters to some customers asking them to ask ISVs about support for solaris 10. From what I can tell, this is to try and create the illusion of a grassroots movement where none exists.
So in light of this, how much of this 'demand' is actually real, and how much of it is Sun customers doing what Sun tell them in the hopes of a more favourable relationship when their support contract comes up for renewal ?
They ruined Linux (SCO proved it, remember? ;) and now they've ruined Solaris. WHAT NEXT?! WINDOWS?!?!
Reality is nothing but a collective hunch.
...Some corps are like a credit rating, they never forget. I mean try and sell an xSeires server from IBM to someone and you inevitably hear 'Microchannel', 'IBM Proprietary' and an 'OS/2' for good measure.
As a rock-in-roll Physicist once said, No matter where you go, there you are.
Village idiot in some extremely smart villages.
First, whatever happen to black font, Mr Schwartz?
I read his blog fairly often, and he's got insights, sometimes, but way too much marketese and "We're the greatest" lines in there.
I quote: "As you're no doubt aware, Sun is set to ship the newest release of our Solaris operating system, Solaris 10. It's the most secure OS the world has ever seen - bringing mainframe features, like logical partitioning, to every platform on which it runs."
* WTF? set to ship? it's not shipping yet?
* The most secure OS the world has ever seen? Not looking very hard are we? Besides, the customer will decide.
* Logical partitioning. didn't linux already bring that a while back?
More quotes: "They love that we can run linux apps unmodified"
* Yeah, but why not run them on linux?
I quote again:
"A strategy to trap them into IBM's proprietary Power5 platform only."
* IBM is a FOR PROFIT company. They are better at the Profit part that Sun is too. Don't blame them for knowing how to make money. Sun should learn from IBM.
* Trap them? As ESR so eloquently put it, GPL Java.
Another point on "They love that we can run linux apps unmodified."
Ah, so what about all those Solaris Apps everyone is dying to run... Oh, that's right, they're Java apps, and they should run everywhere, because that was the promise, except not on my debian boxes, because it's proprietary binary with a cloudy future.
You want java on ALL linux install out there? GPL it. The ball's in your court, Sun, Jonathan, and as long as you want to hold on to Java, you'll not make inroads except through high power meeting with PHBs in 3,000USD suits.
We don't care about Solaris and Java, because Sun is trying to seduce us with "ease of use and transition" while hiding the patents and copyrights that will doom us into an ever-downward spiral of proprietariness, Hereinafter refered to as "The Trap".
I think IBM, by supporting linux, and opening patents for Free Software use, is a much better long-term dance partner for the linux crowd.
So, Jonathan, GPL Solaris and GPL java and J2EE, and then we'll talk. Until then, go change your stylesheet: p {color: #000000;}
"Piter, too, is dead."
Comment removed based on user account deletion
Please remember that once IBM lawyers walked into Sun's offices with their patent portfolio and pretty much bullied Sun out of $20 million.
The story is told here:
http://www.forbes.com/asap/2002/0624/044.html
If Sun is getting worried again, maybe we should all take it seriously.
Hey, if this could somehow turn women into hottie rabbit/cat-babes, all the anime fiends would be forever grateful.
Besides, ask any feminazi, men have already been pigs for centuries.
Chas - The one, the only.
THANK GOD!!!
If you don't like how the iPod and iTunes work, DON'T USE THEM.
That's good advice. I don't use them, I think they're shit. A shiny silver case and a scroll wheel on your player do not makeup for DRM and a pricing structure that mirrors RIAA price gouging without even so much as giving you a CD with liner notes.
iPods/iTunes are cool because the music industry are getting their ton of flesh, and are prepared to say that they're cool. If that's good enough for you, go to town.
When I see ogg vorbis on hardware with legacy mp3 support, free as in beer music online with integrated support for buying tickets and t-shirts then maybe you'll get my money. Until then, I'll stick with mp3/DC++/Winamp/CDR.
-1 Uncomfortable Truth
Jonathan Schwartz is Sun's president and chief operating officer. Scott McNealy is chairman and CEO.
There are a few reasons to use proprietary RDBMS's but I don't think they are as you say. My company offers consulting on open source database solutions and I consider myself to be reasonably knowledgable on the topic.
1. db2 just set the world record for transaction speed - at about 50,000 transactions a second. Last I heard mysql was trumpeting 800 transactions a second with innodb. Not sure about postgresql.
Take a look at the hardware and the cost per transaction per second. I am not saying this could not be done on PostgreSQL (it could) but it would expensive as well. As for MySQL, their performance in situations with large numbers of concurrent updates has never been good, and has actually been decreasing in recent years as they add features. PostgreSQL's performance has been improving, OTOH.
In general, your performance bottlenecks with PostgreSQL are either CPU (add more if you want), Disk IO (add a better RAID 1/0 array with battery-backed caching), or issues with PostgreSQL not being tuned reasonably for the system it is running on (not enough shared buffers, etc).
Finally, these records are not necessarily representative of real-world loads, and in many cases, you can get equal performance by using reasonably decent solutions (PGPool with Slony-I, for example).
2. with partitioning, parallelism, and clustering, you can get subsecond response time from db2 *adhoc* queries against tables containing over a terrabyte of data. Postgresql, Mysql, and Firebird aren't even in the ballpark here. Note: mysql "speed" will end up requiring you to index every single column, which will kill your insert speed, double the size of the data, and their optimizer won't use the indexes anyway whenever you want to access more than 5% of the data.
No complaint about your analysis of MySQL. However, unless you have all the information in memory or unless the the data is spread across a large number of disk arrays, I don;t see how you can even get the information to process it in less than a second. With PostgreSQL 8.0, you can partition the table across a large number of block devices and then read quickly. So yes, it is possible, but in all cases it is not easy. I don't think that Firebird has these capabilities, but I am not sure. MySQL certainly doesn't.
Most of your other reasons are similarly out of date. However, here are a few which I would add:
1) No open source database management product has any sort of mature, *multimaster* replication solution which will work well over slow or unreliable links.
2) PostgreSQL and MySQL (along with Firebird, IIRC) lack supported Two-Phase-Commit structures required for distributed transactions. PostgreSQL is expected to have something on this area in the next version.
3) If you are already using a commercial database management system, it is almost always less expensive to continue using what you are than going to a new system where you have costs of getting up to speed, porting applications as necessary, etc. Open source, OTOH, is less expensive in green-field situations where you don't have existing applications to work with.
I would suspect that within the next couple of years, only this last reason will be the primary reason for sticking with a proprietart RDBMS.
LedgerSMB: Open source Accounting/ERP
Care to back up such inferences, Mr. Editor, or are you just another troll making claims without any support? What precisely is there in Sun's portfolio of app software that only runs on Solaris? Not Java. Not iPlanet/SunOne/N1/whatever you want to call it this month. Care to back it up with facts instead of random allegations?
7 November 2006: The day Americans realized corruption and incompetence weren't addressing 11 September 2001
Nothing DEC made survived the purchase by Compaq.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
I'm sure you read a lot further than Sam did.
Check out the OpenPower stuff they recently announced. It's all Linux all the time. You an even choose to configure it without an operating system!
I suppose IBM could reply with:
"We'll port all of our software to your OS when you port your OS to POWER hardware, which, by the way, leaves your SPARC boxes completely in the dust. Don't your customers want their OS to run fast?
SirWired
That would be like asking Toni Soprano for a loan. That would result in a regular St Valenti's massacre.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
Why doesn't SUN offer to pay IBM to port those applications to solaris!
Pay just like a customer for IBM
If Solaris 10 is as compatible with Linux as you claim, why would it need to be ported over? Don't you claim "we can run linux apps unmodified" in the very same "open letter"? So which of your statements is B.S. - that Solaris 10 runs all Linux apps unmodified, or that all of I.B.M.'s Linux apps have to be ported over to run? Inquiring minds want to know.....
And not if Sun makes their API 100% compatible, including performance quirks.
.NET doesn't work the same on different platforms and you have to change your development to suit.
While testing and support make up a chunk of the development cost it also radically affects the development itself. Do you minimize the design to maximize portability? Do you code to the standard API or write your own custom driver?
Even
" How many of IBM's platforms does Java run on? None."
All of them. Name a platform of IBM's that you can't run Java on.... MVS, TPF, AIX, OS/2, AS400....hell.... they all run Java.
Maybe you meant something else? Or you had a typo?
I think the original poster mean that Sun's Java would not run on those platforms. I think it was based on some kind of strange belief that Java isn't really ported to an OS unless Sun provides the VM.
IBM is bringing out new Power systems targeted at Linux.
Might additional industry accolades be gained if this platform could also offer Solaris for the Power architecture?
This sort of tit-for-tat negotiation is better conducted behind closed doors. I imagine that Solaris on Power, plus more participation and control in Java, would make IBM very open to supporting Solaris wherever Sun wanted.
Screaming in a blog isn't going to do much good.
Don't know which Anonymous coward is standing corrected :-) (parent or gp), but it's quite clear the lxrun is a x86 tool. lxrun works only on x86. Snippit:
Lxrun is a freely downloadable utility that executes Linux applications on the Solaris Operating Environment, Intel platform edition. Lxrun is a software layer that sits between Solaris and the Linux Intel binary executable and remaps system calls "on the fly" allowing them to run unmodified on Solaris.
Try NetBSD... safe,straightforward,useful.
...Uhh...code away.
Here's the correct link.
Seastead this.
As virtually no effort is needed for AMD64/x86-Solaris certification of IBM's AMD64/x86/Linux apps, it is obvious that IBM does not want customers to consider Solaris on AMD64 (or x86).
Virtually no effort?! Even Java products require significant testing effort when shipping them for a "new" platform and simply assuming that everything that works on Linux x86_64 will work seemlessly on Solaris x86_64 is a recipe for unhappy customers. As I work day-to-day on the DB2 UDB internals and I am just down the corridor from the DB2 linux support team, I can say with some authority that it would take a lot more than just adding a tick to the Solaris supported column on the marketing sheets to see DB2 on x86_64/Solaris.
I do not know of any plans to release DB2 on Solaris/x86_64 but I also am not aware of any plans NOT to release DB2 on Solaris/x86_64. Ultimately, it will come down the normal equation of testing and support effort required to get the release out and running against the customer demand. IBM is in the IT market to make money. If customers want DB2 on Solaris/x86_64 in sufficient numbers to make it financially viable, you can bet that there will be a release. There is no business proposition in sitting on your hands to spite Sun if there is financial reward to be had. This isn't a charity, nor is a playground spat.
Cheers,
Toby Haynes
Anything I post is strictly my own thoughts and doesn't necessarily have anything to do with the opinions of IBM.
ROTFLMAO!!!
Why doesn't he get Bill Gates on the line and ask him why Microsoft's being a big meanie and not porting SQL*Server to Solaris? :)
> Take a look at the hardware and the cost per transaction per second. I am not saying this could
> not be done on PostgreSQL (it could) but it would expensive as well.
Good point - a database performing 50,000 transactions a second isn't running on your grandmother's pc. But it wouldn't be correct to assume that all databases scale equally. This is where the greater tuning opportunities of db2 or oracle come into play:
- instead of just a single 8k page size, db2 has 4k,8k,16k, and 32k page sizes.
- db2 has much more flexible statistics gathering than either mysql or postgresql
- db2 has a far more robust & intelligent optimizer than any open source database
- db2/oracle/etc have query parallelism
- db2 in particular is very good at performing most database operations asynchronously: data writes are sent to a buffer pool by one agent, to a log buffer by another, from the log buffer to the log file by another, from the log buffer to actual storage by another, etc, etc.
- both mysql & postgresql have very primitive memory tuning capabilities: mysql can't share io buffers between innodb & mysql. Postgresql just doesn't have many options to manage sort memory, separation of memory buffers by tablespace, etc, etc.
But keep in mind - many of these weaknesses are also strengths: since the missing configurability is translated to simpler management for smaller & less-demanding databases.
> However, unless you have all the information in memory or unless the the data is spread across a
> large number of disk arrays, I don;t see how you can even get the information to process it in less
> than a second.
Ok, here's an exeample architecture using db2. Would work identically on informix, and probably Terradata. Oracle has similar capabilities, though a little different since they're shared-disk rather than shared-nothing for db2/informix/terradata.
Step 1: spread data across 10 2-way 64-bit AMD blades. Each blade has 8 gbytes of memory and a fibre connection to storage. You can expand this up to *hundreds* of blades (nodes) - and get almost linear performance improvement the entire time. You've now got a 20 CPU system for about 10% the price you'd pay for a Sun E10000 or HP Superdome. Since the data is spread across all nodes by hashkey - every query will get 20 CPUs processing it together - each on 1/10 (100 gbytes) of the original data.
Step 2: partition the data on each node by range or value via MDC. Say, by 'day' & 'department' - so that you've got (for example) 1000 days & 1000 departments. Now, each query will only tablescan the data within the blocks it needs: whether it needs to scan 5% or 75% it gets a linear performance improvement corresponding to the reduction of data.
Step 3: implement parallelism on each node. Now, this sample config only has 2 CPUs - so the opportunities aren't as great as they are on an 8-way. But a two-way will still usually cut your query time in half.
So, now a few examples:
1. query for 10 days of data for 1 department:
a. clustering cuts 100 gbytes on each node down to about 1 mbyte to scan on each node.
b. query parallelism cuts down scan on each node to about 500 kbytes
2. query for 100 days of data for 100 departments:
a. clustering cuts 100 gbytes on each node down to about 1 gbyte to scan on each node.
b. query parallelism cuts down scan on each node to about 500 mbytes
3. query for 365 days of data for all departments:
a. clustering cuts 100 gbytes on each node down to about 40 gbytes to scan on each node.
b. query parallelism cuts down scan on each node to about 20 gbytes
All three queries - whether very selective or not, all use the same facility for a linear performance improvement (unlike indexes). The last query is unlikely to complete in 1 second - but I'll bet it'll be fast nevertheless - especially since it'll probably end up with over 2
db2 doesn't have anything on Access!
POSIX standards? POSIX standards? Young man, Solaris IS POSIX standards, and nearly every other UNIX standard going. Where did you think the standards came from in the first place?
Linux is the Johnny-come-lately as far as POSIX is concerned.
Stick Men
Kettle, meet Pot.
For a while now, Sun has been pushing Java platform that would allow all applications written by Sun to run on IBM hardware and under WebSphere if the customer so chooses. It's quite reasonable for them to ask IBM to compete on advantages of each product rather than trying to pull a Microsoft and lock-in users into unwanted components.
Besides, there must be lots of existing Rational customers running, say, clearcase servers on Solaris. It's not really fair for IBM to force them to buy new hardware.
something like Craig's list, if you had to pick between MySQL and Postgres only, not based on claims by the distributors, but based on the capabilities you elaborate in your post and what you know personally. Starting from scratch.
If possible, in layman's terms (or as close as you can get), which one, and why? Assume that you are using multiple low-cost, off the shelf hardware (throwing single AMD-64 boxes at it, or even single opterons), and either linux software raid or 3ware sata raid cards as your options, sata drives. Add speed by adding more single AMD-64 or equivalent computers, spreading out tables by state or city.
2nd choice, as business income grows, dual opterons instead of single AMD-64 or single opterons.
Would there be a big difference in speed/capability between MySQL and Postgres, based solely on the feature availability of the two databases? Or not, due to limitations of hardware above?
Assume Linux w 2.6.x kernel on all.
tia for the insight. Not looking for a benchmark opinion (I can get into that argument elsewhere), but an insight into features for running something like Craig's list.
Who says any company has to be the enemy to get bashed? Heck we bash Redhat at times like it was situated in Redmond.
We bash companies till they're all about opensourced and GPLed software. Companies with high-stakes software like Sun with Java being the extremely high-stake, will be bashed excessively till they give in. Sure we love Sun for not being Microsoft, and Sun will be praised in Microsoft-related articles, but the larger IT companies, Sun, IBM, nVidia and ATI, Adobe, Dell etc all know they will be armtwisted into opensourcing software and selling nothing but Debian Linux.
Sun is not the enemy but its not a 100% GPL software house, and who cares if they need to make paychecks.
"Give orange me give eat orange me eat orange give me eat orange give me you." -Nim Chimpsky
I'm sure no one will read this, but I very recently took some class files that run under Windows XP, loaded them to a node in an AIX cluster, and they worked exactly the same, just at a different speed since the machines are so different. I took them home, ran the exact same files on my Sun Blade . . . and it behaved exactly the same way. It wasn't anything intensive, but there were absolutely no problems, either.
Does anyone else think it's really, really, really cool? Or am I just easy? About logos. :)
Sig not available, please try again later. If the problem persists, then the submitter is an idiot.
One of the issues here is that we are talking about different types of "Enterprise" products.
PostgreSQL cannot compete with Terradata in this space. Backplane might eventually, but it is certainly not even worth considering as a mature solution at the moment.
PostgreSQL works very well on databases ranging from 1 GB to a bit over 1TB in size. I don't know anyone running larger databases, but it could be happening, and give it a few years..... I certainly consider many of the larger PostgreSQL deployments to be "enterprise-grade."
But you are talking about databases which are much, much larger than that. PostgreSQL will never likely provide the parallelism you speak of simply because it is process-based rather than thread-based so usually one query can only hit one processor. It might be possible to create a similar solution using existing tools on PostgreSQL but ti would not be elegant.
Also one question-- suppose you have your storage go bad on a blade-- what happens to your relational integrity if the records are spread across various servers?
LedgerSMB: Open source Accounting/ERP
> PostgreSQL cannot compete with Terradata in this space.
right, but this is a huge space: and given the rate of data explosion, 3TB databases have become extremely common. And that's a good thing for all database products & customers: it means that db2, oracle, etc can keep adding more exotic features to support multi-TB databases while postgresql, etc can quietly pick up more of the smaller ones.
> Also one question-- suppose you have your storage go bad on a blade-- what happens to your
> relational integrity if the records are spread across various servers?
Depends on which dbms you're talking about: on oracle it's no big deal, since they use shared-disk, and one node can quickly take over for another on 10g (was OPS). Of course, that's a complex, high-priced product, and doesn't scale as well as the shared-nothing options like db2 & terradata. I can't answer for terradata, but if you're on db2 then you've got a few options:
1. implement a HA solution between pairs of nodes, so that one node can take over for another. This can be via HACMP or Veritas (both complex) to take over the storage from the crashed node, or even HADR (client reroute - not complex) to support a second duplicate node.
2. implement a failover database. This keeps the architecture very simple, and is a more complete solution (read: disaster recovery) - since it can protect you against datacenter floods, etc: by locating the failover in an entirely different data center.
Since we're generally talking read-only data here, you've got more options than on a large transactional database. I'd typically prefer #2, followed by #1/hadr. HACMP and Veritas are a pain in the butt, IMHO.
I'm sure no one will read this, but I very recently took some class files that run under Windows XP, loaded them to a node in an AIX cluster, and they worked exactly the same, just at a different speed since the machines are so different. I took them home, ran the exact same files on my Sun Blade . . . and it behaved exactly the same way. It wasn't anything intensive, but there were absolutely no problems, either.
I read it, and it matches my experience too. However, anti-Sun and anti-Java rants are pretty popular here. I have given up trying to understand why.
Isn't an "open letter" directed to an individual a contradiction?
Mathematics is made of 50 percent formulas, 50 percent proofs, and 50 percent imagination.
you answered the questions exactly the way I was hoping for. I've asked similar questions in the past, looking for something like you laid out, only to be ignored.
/. uses it. But that's because they started with it. Would they do it all over again the same way if they were starting now? Or would they be using php/mysql instead?
The information you supplied, along with other info I've received from reading other posts on local mailing lists (with very knowledgeable sys admins and database admins, I won't name which list but you'd know it if I listed it), and feedback to my posts on that mailing list, along with my reading of the docs to both mysql and postgres, along with reading several books about both databases leads me to lean toward postgres also. I've avoided attempting to work with either one up until now, and instead have been concentrating on administering and learning the guts of apache first, along with linux and the related other server apps (ntp, others) that I'll need to keep everything humming. Low resources and funds are always an issue, and I didn't want to add a database to computers with low memory resources.
One of the things that was repeatedly pulling me in the direction of MySQL was the lack of books and documentation on the combo of Postgres and modperl for web sites or ecommerce sites. The net is littered with books on php and mysql. From a newbie's standpoint, on support alone, and then throwing in documentation, mysql and php have postgres and php/modperl beat hands down. Sure, the postgres site itself has documentation. And I've even downloaded most of the docs and printed them out and read them from the earlier releases of postgres. I even downloaded, re-formatted manually, re-assembled and then printed out the online book on postgres to better understand how it works. But most of the docs assume you already know something before you start. This is where postgres falls down, at least for me.
Is modperl out of favor? Sure
The $500 per server is more news to me. My understanding was, if I distribute an application with mysql embedded in it, or dependent on mysql, then I have to pay for it, otherwise not. If it is the case that I would have to pay $500 per server (even if running on Linux?) starting from server #1, then that really settles the question for me, its postgres from the beginning, and no looking back.
I've been curious for quite some time on market share between MySQL and Postgres. The recent market share stats that one of the analyst companies released were quite a surprise to me, and at the same time, not a surprise. MySQL has a far greater market share than Postgres, and Firebird, even though it doesn't have as much buzz as MySQL and I think even Postgres, has a greater market share than Postgres. Postgres' market share is going to climb a lot now that it has a port to Windows, but the market share stats confirmed why MySQL has more buzz, more talk, more people following it than Postgres. It has a far higher market share. Windows has a lot to do with it, but I couldn't figure out why MySQL was/is so much more popular than Postgres if you compare the features and license. And I had a tough time understanding where Postgres was, would I be adopting a database where I would have trouble getting help with? If Postgres is closer to Oracle for example, would I be running into users similar to the responses one gets on the OpenBSD list, for example? Or like the guy on the debian-user list who wanted to cut my throat and drop me on the mepis list for suggesting that if you apt-get remove mepis apps from a mepis installation, and apt-get update regularly from only sarge sources, the distro becomes debian?
I learn best by example. Reading a book on apache didn't do it for me. Editing the config file and trial and error did it for me. Following an example of how to set up a web site that uses a database (the classical book store example) is what is going to teach me the basics of getting a dat
We're very impressed with the Opteron workstation we're testing. The specifications on their 1U/2P server (V220?) is far ahead of any Dell out there (never mind that they're Intel only). 16GB of RAM on a 1U/2P is pretty damn impressive, and you get the Opteron 850 in that. Tasty.
I would run Solarix X86 on Opteron, if I needed a Unix platform for a Solaris-based application in a heartbeat. I am even hoping to entertain using one for VMWare, if VMWare gets around to supporting it (they claim to be very interested in Opteron, which will no doubt make their ESX even better - and it's pretty damn good on Intel).
There exists no way of exchanging information without making judgments. --Bene Gesserit Axiom
First the open letter is about Solaris on x86. IBM will no doubt support Solaris on Sparc because most of their Websphere revenue comes on that platform.
Second, Sun trying to say that Solaris is the most secure OS on the planet to the company that sells z/OS and i5/OS hardware is counterproductive.
At least in terms of an ODM to registry comparison. Whatever IBM doofus conjured putting a registry into a UNIX system deserves a good kick in the pants. Sorry, but the ODM is a registry, plain and simple. Non-editable, binary files used for basic system configuration have no place on a UNIX server. Neither do tools (smitty) which directly edit and modify the structure of critical system init scripts. Idiots.
Yes if you don't even change one line of code you have to regression test to ensure compatibility.
But the exhorbitance of the regression and performance testing comes because of imcompatibility issues between different vendors platforms... ESPECIALLY in one VERSION of the same OS.
If the cost is that much of an issue it behooves the enterprise customer to demand that those costs be minimized by forcing the OS vendor to ensure compatibility in version updates or the software won't be updated.
And besides, if the coding is free compared to test and support, you're better off just rewriting the app from scratch for each new version to better adapt to your company's processes (which are always dynamic).
AFAIK, those effects have been gone for years; I stopped encountering them sometime in the mid- to late- 90s.
In the late 90's, your AIX system was still hosed if you ran out of disk space during a SMIT operation: not only was the system critically dependent on ODM, SMIT and the ODM library were also so poorly written that they couldn't cope with this common system state. So, no, whatever dependencies there were on ODM weren't fixed at that time.
some of the things IBM originally put into AIX to "industrialize" it were things that folks complained Aren't The Unix Way, but they've ended up in other Unices as the years go on. LVM, enhanced security, dynamic kernel, a systems management interface, etc. Yet, 15+ years later, I still hear complaints about how different & not-nomral-Unix AIX is. Whatever.
That, too, is just BS. LVM is rarely used even on Linux--convenient as it may seem to users, it is just one of the technically most stupid ways of managing disk space imaginable. And IBM can hardly take credit for GUI-based management (which they got woefully wrong with SMIT, both technically and in terms of user interface) or dynamically loadable kernel modules.
AIX has always been incompatible in ways that range from merely annoying to seriously bad. Many of their decisions were indeed driven by the desires of their mainframe customers, but that still doesn't make those features good ideas.
And although I thankfully don't have to deal with AIX anymore, I doubt it's much different today. Trying to portray the p.o.s. that AIX was/is as some kind of progenitor of modern UNIX or Linux systems is ridiculous.
Prove it.
1. db2 just set the world record for transaction speed - at about 50,000 transactions a second. Last I heard mysql was trumpeting 800 transactions a second with innodb. Not sure about postgresql.
I seriously doubt that those benchmarks were on the same hardware, so the comparison makes no sense.
2. with partitioning, parallelism, and clustering, you can get subsecond response time from db2 *adhoc* queries against tables containing over a terrabyte of data. Postgresql, Mysql, and Firebird aren't even in the ballpark here.
First of all, at least learn the correct metric prefixes before you start spouting benchmark data.
Now, yes, with clustering and keeping stuff in memory, you can get good response times. That's not some deep, fundamental property of the database, it's a simple add-on that sits on top of the database. I believe there are already solutions like that for PostgreSQL and MySQL, although they are not quite as flexible as those for DB2 yet. But very few people need that sort of thing.
So, on a big project where the database is critical - you will actually *save* money going with a commercial database.
Or maybe not. With a commercial database, you may have much more downtime because the commercial vendor is taking months to fix a critical software problem that you yourself could fix instantly if you only had the source. And with a commercial database you are in a real bind if the vendor makes incompatible changes or discontinues the product altogether.
You see, the cost savings from open source software aren't a result of the software not costing anything, they are a result of reduced risk. Let me stat that again: open source software has substantially lower risk than commercial software and that is where your cost savings come from with open source software.
6. Consistency: since most organizations will require a commercial database for their most demanding applications - and they can benefit from a complexity reduction by using the same database on all applications.
You're kidding yourself if you think that any large organization can consistently deploy a single platform for everything. And if you choose a commercial database like DB2 consistently for every database deployment in your organization, licensing costs are going to be a major issue.
Not to say that the open source solutions aren't great: they are, and can pick up much of the database work these days. But there's still a huge case to be made for commercial products
DB2 is actually a decent database with some advanced functionality. People may want to use it for specific applications. But your arguments for why and when people should use it are pretty naive. People who need DB2 will know it; for almost everybody else, one of the FOSS databases is fine. And if you stick to SQL standards, it's actually pretty easy to move between them (imagine that).
It would suck if Purify wasn't available for Solaris. For some reason it never quite worked right on linux x86.
"The big clue is the disappointing financial report from last quarter for Sun. Revenue actually fell. "
Revenue fell, but they are now making money again. Not much, but cash flow is positive. $18M/quarter is still much more than many smaller OS firms. Having 7B$ in the bank also helps.
The real question is whether smaller Open Source software development firms will be able to compete for the bigger contracts with IBM and Sun sucking an increasing share of these as newer technologies/architectures begin to emerge.
Besides, Sun won't die because the Pentagon can't afford to let it. Too many of the tactical battlefield command systems run on a modified version of Sparc/Solaris, not to mention their new DARPA-based project for tight junctioning of multiple core CPU's needed to highly parallelize encryption/deencryption so that Echelon can process and interpret ALL internet traffic in real time.
Last I checked IBM's own pSeries boxes shipped with Linux couldn't run WebSphere MQ. DB2 Server was released for it just within the last year.
Can I use your money to build my system?
LedgerSMB: Open source Accounting/ERP
No, it just shows that few within the large Linux community are conversant with issues other than those around x86-based computing.
I anticipate uninformed blathering will fall off as Solaris becomes open source and as more large development shops begin exploring integration among newly emerging architectures, such as AMD 64.
Sun remains far better positioned to exploit this area because of its depth of engineering experience, long history of multiple-platform apps, and connections to open source.
Admittedly, their PR operation will remain poorly funded relative to IBM or M$ so this won't be apparent to the average computer geek for perhaps several more years.
Sun could help itself if it could train its executives to stick to the knitting associated with focusing on increasing the benefits and timeliness of its own products and innovations and less time stirring the pot among a customer base that must necessarily use a variety of different platforms/hardware because of "personnel/expertise" lock-in. They spend far too much time issuing goofy messages and giving goofy talks laden with too many off the cuff remarks that step all over a simple focused effort to market their products. They need fewer cheer-leaders and more astute business persons.
If Johnathan Schwartz was doing his job, he wouldn't need to be worrying about what IBM is doing. That would take care of itself.
IBM to SUN: Please oh Please make Java Open Source so we can hijack it.
...
Sun to IBM: Get Bent. We already have a community contributions from the jcp (jcp.org). Why should we Open Source it when all YOUR major software is PROPRIETARY AND CLOSED SOURCE and you only intend to hijack the Java platform.
Few months goes by
SUN to IBM: You seem to be back to your old proprietary lockin tricks.
IBM to SUN: Get Bent we are making a profit from our proprietary lockin. Screw the customers - we are in denial over Solaris 10.
See how simple things are!
Solaris x86 needs a real, good, strong selling point. What is it?
Solaris 10
-Fastest OS on the planet for x86, AMD64 and SPARC
-DTrace
-Containers (allow over 8000 isolated partitions on one system).
-ZFS
-Native Linux Support
-Military grade security
-Free as in beer AND speech
-Massive scalability (hundreds of CPUs handled efficiently).
-etc.
Not the sort of features you will find in Windows.
Now that Solaris and Linux are both Open Source Sun is really not the enemy. We now have more alternatives to MS.
Fact: Sun has donated more lines of source code to Open Source than any other company.
******* 5un has join #chat ***********
5un: Yo!
1BM: Er.... what do you want?
5un: We're mates right?
** 1BM looks at 5un
5un: Sure we are! We've been mates for ages!
5un: Hey, how's the Solaris 10 ports going?
** 1BM looks at 5un
5un: Geez, can't wait to see them!
** 5un kicked by 1BM (l053r)
If this is any indication of a company's direction immediately after releasing a new product, I'd be selling. Now. After the deal they just made with Microsoft, and the current crop of lame "why don't they think is cool?" comments, you're looking at the most defensive OS rollout I've ever seen. It is the difference between "what can we make" and "what can we save". I haven't seen anything this undignified since the slow collapse of DEC.
insecurity asks the wrong question irritation gives the wrong answer
I take it you're not so sure now. ;>
Clueless Sun complains IBM isn't porting apps to Solaris 10
Quick summary: Sun should pay IBM to port the apps or adopt Linux wholeheartedly rather than further splinter the Unix marketplace.
I think you could, but here are the limitations you will run into:
1) Disk IO
2) Memory limitations
3) CPU limitations
Now, you didn't tell me what sort of concurrency we are talking about or the database size. If I have enough memory to cache the whole database in memory and then some, and if I have one CPU per concurrent process, and if I can stripe this across a number of disk arrays to maximize disk I/O, I think it could be done. It is primarily a question of how much one wants to spend on it.
And, of course, the WAL would have to be striped across a pretty good RAID 1/0 array to keep up.
I suspect that DB2 would be less expensive than PostgreSQL in this configuration.
Also, you could run into an issue if this is heavy wrt updates. You would probably need to concurrently run vacuum on another CPU repeatedly in order to keep the dead tuples down.
PostgreSQL doesn't cluster very well at the moment, so we are looking at some sort of huge, expensive machine.
LedgerSMB: Open source Accounting/ERP