Sun has announced a new set of products and services targeting developers, startups, and internet companies seeking to build and deploy their web infrastructure on Solaris 10. The three offerings are Solaris Express, Developer Edition, Solaris + AMP (Apache/MySQL/PERL or PHP), and an expanded Sun Startup Essentials program.
Sun is pushing up hardly Solaris, OpenSolaris and Java… Jonathan Schwartz’ work is impressive because he is leading the company to its new brighter age.
Sun is Rising! I like how they have balanced the needs of the individual enthusiast programmer with the corporate developers. Basically, Sun isn’t leaving anyone out.
At this point, I would say Sun is the #1 open source comapny in the world. Can a company prosper by providing excellent products and service _WITHOUT_ hording IP?
Let’s hope so!!
Surely that title must go to IBM, right? Whatever the result, with Red Hat, IBM, Sun, Novell, and Canonical out there, it’s a new age for the Open Source business model.
Surely that title must go to IBM, right?
I would agree that IBM is one of the largest proponents of OSS, but I’m not sure they are the biggest creator of OSS.
SUN: Solaris, Java, Netbeans
IBM: Eclipse
I’m sure there are more for each, but Solaris and Java just seem so huge for me. That would be like IBM opening up DB2 and AIX.
Having said that, IBM certainly has done wonders in their promotion of OSS, especially of Linux. Personally, I love their commercials!
I would agree that IBM is one of the largest proponents of OSS, but I’m not sure they are the biggest creator of OSS.
Sun has established stronger branding around their OSS stuff, but IBM is a leader in many key areas of the OSS development community. For example, they are the leading corporate contributor to the Linux kernel project and the Apache project. I would imagine they sell more Linux clients and servers than anyone else by a considerable margin. They employ more Linux kernel hackers than any other hardware vendor.
That would be like IBM opening up DB2 and AIX.
I’m an AIX developer. So take it from me–you don’t want IBM to open AIX. Besides the licensed code and patents it contains, it’s just too complex. We can only build the damn thing through a Rube-Goldberg device of a build system that I wouldn’t wish on my worst enemy. Everything in the kernel has to be painfully aware of how it accesses memory because the in-kernel virtual memory implementation is very aggressive: the kernel address space is pageable (kernel memory can be swapped to disk); pages can change size and can be relocated to a different segment at any time; and interrupt context is interruptible. With every release, we add more sophistication that makes things more challenging, and the coming AIX 5.4 release is certainly no exception.
As a platform for high-end databases, AIX is unmatched. But it’s an incredibly challenging development environment that’s completely unsuitable for distributed development. I’m not as familiar with DB2, but I imagine that similar design trade-offs led to similar kinds of development hurdles.
Furthermore, AIX only runs on System P hardware, and IBM provides AIX (including fixes, updates, and new releases) to its System P customers at no extra charge. It’s not like anyone has a machine that could run AIX if IBM opened it up.
Furthermore, AIX only runs on System P hardware, and IBM provides AIX (including fixes, updates, and new releases) to its System P customers at no extra charge. It’s not like anyone has a machine that could run AIX if IBM opened it up.
They shouldn’t have killed AIX/PS2! Ability to Hot key between OS2 and AIX was cool.
There’s still some time to rethink AIX on x86
In the platform market, there is tremendous value in targeting commodity hardware. Solaris is an excellent example. Maybe Sun didn’t need Solaris on x86 in order to continue funding their SPARC offerings, but it has certainly helped a great deal. Solaris seems like less of a niche platform now that it has an x86 port, and OpenSolaris has also increased the awareness of Solaris as an alternative to Linux. I’m not privy to the internal discussions (and if I was I wouldn’t comment), but I would imagine that the recent moves to broaden the target market and brand strength of Solaris are quite troubling for IBM.
AIX on System P, on the other hand, is a niche platform in both perception and practice. With every product generation, the target market moves uphill to a smaller and smaller niche of customers that absolutely need the best performance for their massive database workloads. There was an Itanium port for AIX (the infamous Monterrey Project that was implicated in the SCO suit), but that turned out to be the gold standard for niche platforms (i.e. nobody wanted it).
At this point there’s really no turning back. Most of what makes AIX excel in its market is tied to the capabilities of System P. Scale it back to the x86 feature-set, and you have what amounts to a somewhat more sophisticated off-shoot of Solaris with horrible hardware and software compatibility and vanishingly small developer mindshare.
My theory is that there’s a race going on between us System P guys making UNIX systems big and mainframe-like and the System Z guys making mainframes accessible to more of the enterprise market via Linux. Then you have the BladeCenter folks that want to make discrete AIX, Linux, and Windows systems fit in the same box with centralized management.
The market for a big, virtualized UNIX machine is getting squeezed on the low end by the blades and on the high end by the resurgence of the mainframe. It’s nice to have an option between discrete UNIX blades and Linux on the mainframe, but this market is threatened by the late arrival of viable x86 virtualization (if they can make it manageable, that is).
With this competitive landscape in mind, I envision blades and the mainframe as the winning horses. The former solves the pressing problem of consolidating the management of discrete cross-platform workloads, and the latter solves just about every problem in IT… as long as you can afford it and your workloads are Linux-compatible. P will continue to live in the form of P blades, but AIX seems to be going the way of the dodo bird, destined to be replaced by Linux. Solaris will be a viable Linux alternative for blade environments, while Linux rises from the 386 to conquer the biggest enterprise mainframes.
I won’t be out of a job, though, since I’m known as the “Linux guy” of my department. They’ll probably just move me to Z, and as long as they let me stay in Austin, I’m fine with that.
Edited 2007-03-10 06:31
I think you underestimate the value of some of AIX’s features. Although my AIX skills are rusty, the ability to create LPARs and manage resources is nothing to sneeze at considering Linux does not have those capabilities, or if they are present, they are not “production quality”.
We are in the process of building a SunFire 4800 (production) and a V480 (development) using Zones and Containers to limit access to CPU’s in order to support a lower cost Oracle licensing model. Enterprise features don’t come cheap, and for demanding customers Solaris and AIX will be there to meet their needs.
There is no doubt that AIX is specalized, but it was the easiest flavor of Unix I ever administered (thanks to smitty and pimpworks.org). If it came down to me using a Unix variant other than Solaris I would use AIX before HP-UX.
Hmm, I always get a bad feeling everytime I see HP-UX; not that its a bad operating system or anything, but the way it is treated by HP.
Give HP’s relationship with Microsoft and its willingness to bend over to meet Microsofts demands on any occasion, I question the long term road map of HP-UX when compared to other offerings out there.
Its like Compaq and when it bought two stella companies, Tandem and Digital, then ran them into the ground; same thing is occuring with HP, HP-UX and their new management; focus on quarter to quarter sales than long term, and how a relliance on Microsoft for future growth will leave them in a tricky situation.
One of the problems with HP is their investment in Itanium and the mistaken belief that their Enterprise customers would drop PA-RISC and go with them without question. One of the places I interviewed with when I was unemployed had a signifcant investment in HP software and hardware (including a PA-RISC SuperDome machine). It would have cost them a fortune to transition to Itanium.
Expecting your customers to change platforms and eat the cost of that change is arrogant to say the least. On the x86 and x86-64 side of the house, things are pretty good. We use HP exclusively for our x86 needs and the machines are pretty much bulletproof. That is why you hear nothing but Windows and Linux out of HP, is because nobody wants to hear about HP-UX and their on again, off again strategy for HP-UX.
One of the problems with HP is their investment in Itanium and the mistaken belief that their Enterprise customers would drop PA-RISC and go with them without question. One of the places I interviewed with when I was unemployed had a signifcant investment in HP software and hardware (including a PA-RISC SuperDome machine). It would have cost them a fortune to transition to Itanium.<?i>
When all this cafuffle went down, even Sun Microsystems was actually surprised at the abruptness of the change over – IIRC, a manager/engineer at Sun said that PA-RISC was an interesting architecture and he was confused how a company could move from a well known, well respected achicture and beat the whole company on something that is untested and and majorly problem prone – both in hardware and software availability. Given the current situation with how things are going with Itanium, apart from a few ‘paid for conversions’ by Intel/HP; you’ll probably find that the focus will move back to x86, or as Scott NcNealy said, “it ain’t pretty, but it gets the job done”.
[i]Expecting your customers to change platforms and eat the cost of that change is arrogant to say the least. On the x86 and x86-64 side of the house, things are pretty good. We use HP exclusively for our x86 needs and the machines are pretty much bulletproof. That is why you hear nothing but Windows and Linux out of HP, is because nobody wants to hear about HP-UX and their on again, off again strategy for HP-UX.
I think the one thing that would have really caused some damage to competition would have been to firstly port OpenVMS to x86-64, then opensource OpenVMS – couple that with improvements in x86-64 being imported from PA-RISC.
As for HP-UX, given what we know about Solaris (being opensourced), pull off the parts which made HP-UX unique, such as the management tools and so forth, port them to Solaris, rebrand OpenSolaris as their own UNIX which would address the current relationship between HP and UNIX.
Why keep maintaining AIX when it would be better to kill off AIX, port features to Solaris, and port Solaris to their POWER based platform.
Infact, actually, Sun opened the door to that possibility around 2 years ago; IBM turned down the offer; Solaris could have been the platform that spanned the great divide between the last remaining RISC UNIX vendors, and x86 world.
This is like Microsoft posting on the LKML saying that they’ll give all the Linux hackers a free copy of Vista if they agree to stop developing Linux. Or Toyota offering their cars to Honda at a steep discount. This is not a deal to be taken seriously!
1) IBM doesn’t outsource technology unless they absolutely have to. They don’t want to be beholden to any third-party vendor, especially a direct competitor like Sun.
2) Porting Solaris to POWER would be relatively easy, but porting Solaris to System P would be much harder. There are many, many features that are unique to this platform that aren’t addressed by any other operating system. They aren’t even attempted with IBM’s own Linux on P implementation.
3) If IBM was to drop AIX and replace it with another UNIX OS, it would be Linux. This is the same point I made the other day when you suggested that Dell sell Solaris servers. It just won’t happen. You don’t replace your own work with that of your direct competitor. That’s stupid.
Solaris could have been the platform that spanned the great divide between the last remaining RISC UNIX vendors, and x86 world.
But it didn’t happen, and Linux runs on all of these architectures anyway.
Edited 2007-03-10 10:08
1) IBM doesn’t outsource technology unless they absolutely have to. They don’t want to be beholden to any third-party vendor, especially a direct competitor like Sun.
But they’re quite happy with outsourcing their operating system to a third party like Red Hat or Novell; and worse still, any contributions made not only benefit IBM but since Oracle now has its own distribution, they’ll directly benefit too.
Atleast with Sun, they don’t have anything comparable to DB2 on their books – lets remember, IBM does make alot of its money via services, which actually involves servicing Sun hardware and software as well – they’re hardly a distant company from Sun.
2) Porting Solaris to POWER would be relatively easy, but porting Solaris to System P would be much harder. There are many, many features that are unique to this platform that aren’t addressed by any other operating system. They aren’t even attempted with IBM’s own Linux on P implementation.
If you kept up with the play, IBM is moving to one standard architecture for their whole high end systems, POWER processors from mainframes to services, from blades to workstations, and everything in between.
How is contributing to Solaris going to benefit Solaris given that Sun doesn’t sell POWER hardware? and given that there is a move to x86 hardware any way, and IBM already provides services and products for Solaris, they’re not going to lose out.
3) If IBM was to drop AIX and replace it with another UNIX OS, it would be Linux. This is the same point I made the other day when you suggested that Dell sell Solaris servers. It just won’t happen. You don’t replace your own work with that of your direct competitor. That’s stupid.
But Dell is quite happy with selling Red Hat loaded machines even though a good portion of the high value stuff is actually being won by IBM.
You do realise that not all decisions are made on price; there are people who choose Sun hardware, irresptive of what others offer; same goes for Dell; a Dell customer who goes to run Solaris, went with Dell because they offered something Sun didn’t.
Dell might have offered a complete suite of desktops, workstations and servers loaded with Solaris at a cheaper price than if that client were to purchase a server off Sun and desktops/workstations off another vendor.
But it didn’t happen, and Linux runs on all of these architectures anyway.
Have you run Linux on those other architectures? its absolutely terrible; Linux’s only rising star is on the x86, everything else its a broken dodgy experience that I wouldn’t want to repeat for all the tea in China.
Edited 2007-03-10 11:53
There is Godson2 MIPS64 which is similar to PPC and is going to be very cheap. This is a target for AIX. But NetBSD would be there before AIX.
Also, in a sense, Sun has a chance to be the “Open Apple”. They can provide the Hardware, Operating System and Development Tools. Since their engineers are the creators of all three, I would expect a very high level of support.
I’m dreaming of some yummy Sun laptops. I wonder how they are doing as far as ACPI and wireless support? I noticed their developer edition of Solaris Express specifically mentions laptops. I’m definitely interested.
[EDIT: Spelling errors]
Edited 2007-03-09 17:00
There are Solaris ready Sparc laptops today, but they are expensive.
And I don’t think they do wireless. Also, I tried to find them [Sun Ultra 3 Mobile Workstation] on Sun’s website and they said discontinued.
Yes, vry expensive:
http://www.tadpole.com/
That doesn’t really change the fact they’re still excellent laptops, especially if you’re looking for a real, mobile UNIX workstation. When I purchased my Tadpole, I wasn’t given Microsoft Office or Windows XP, but rather, Solaris 8 media and a separate CD containing StarOffice. See the difference? 🙂
Also, the Tadpoles that I’ve worked with do not use ACPI, but rather Tadpole’s own power management interface, which Tadpole will be contributing to ON (OpenSolaris) at some point. Tadpole has drivers for the Atheros-based wireless cards, as well as a wireless configuration application.
(I’m writing this from a Tadpole SPARCbook 6500 with a 650MHz UltraSPARC-IIe, 4GB of memory, and 2x60GB drives, running Solaris Nevada b57.)
Edited 2007-03-09 19:31
Not to be nosy (but I am), but I couldn’t find any prices on the site. What does you range laptop go for? Feel free to ignore this if you want!
From my understanding, Tadpole cannot publicize their prices, because they’re a government contractor, as are most of the other SPARC portable vendors.
I will tell you that a SPARCLE (the model below a SPARCbook) runs around $3,000-$4,000 with a 500MHz UltraSPARC-IIi, 256MB of memory, and a 20GB disk. I didn’t pay the full price for my SPARCbook (purchased it through a liquidation), but I’d suspect it was somewhere between $20,000-$30,000 new in 2003.
In a previous job we had a SPARC Book in the back room that had been unused for years. They are definitely nice laptops that can take some abuse, but they are clearly for those with deep pockets.
There is Naturetech as well. The deliver Sparc laptops also.
I know two years ago we got prices from them, but i cannot find that email.
Its been my experience that all of the SPARC portable vendors have similarly configured low-end and mid-range systems within the same price-range. In some cases Tadpole is cheaper, but prices per configuration and model tend to change once you get into the high-end, specialized portables. (ex. Tadpole Bullfrog).
Runs on x64/x86 too.
Sun doesn’t really make consumer level products, such as most laptops. However, OpenSolaris is getting to the point that you can buy almost any major vendor hardware and have it work. I recently bought an Acer Aspire laptop off-the-shelf, and got everything major on it working: graphics, sounds, wireless. It doesn’t report battery life, which is a little problem when my laptop just suddenly shuts off, or suspend yet, but that’s rarely a big issue for my use. Wireless and sound were the biggies for me.
“I’m dreaming of some yummy Sun laptops. I wonder how they are doing as far as ACPI and wireless support? I noticed their developer edition of Solaris Express specifically mentions laptops. I’m definitely interested. “
Then the Sun SparcStation Voyager would be a nightmare for you, such as it was a yummy dream for me. Fantastic device… 🙂
I’m dreaming of some yummy Sun laptops. I wonder how they are doing as far as ACPI and wireless support? I noticed their developer edition of Solaris Express specifically mentions laptops. I’m definitely interested.
Hmm, hardware support has always been the achilles heal of Solaris x86 – its sad though, given how great the operating system is over all; what is needed is for Sun to hire 1000 programmers, and all they should be doing is eating, breathing and living driver writing and nothing else.
Pull the code from all sources, be it FreeBSD, NetBSD, OpenBSD – pay companies for access to documentation if required; if failure of adoption of Solaris is due to the fact that people can’t run it on their hardware, its possibly the worse way one can lose in the marketplace.
You might want to read this document to give you an idea of the contributions Sun makes to OSS:
http://ec.europa.eu/enterprise/ict/policy/doc/2006-11-20-flossimpac…
“””
Surely that title must go to IBM, right?
“””
I’d still give the title to RedHat. IBM’s commitment is solid enough, for now. But a change of CEO, poor performance of their stock valuation (for *any* reason), etc. could conceivably change that.
Large corporations change focus all the time, issuing grand visionary statements about the wondrous benefits of their new direction, in impeccable market-speak.
I’m not saying that I am expecting that with IBM. But if they did they’d lose a few employees.
The same could happen with RedHat… but they’d lose almost all of their technical employees, much of their non-tech staff, and most of their credibility with customers.
RedHat’s heart and soul is Open Source. With IBM it’s just a market strategy.
Sun, now that they are getting some steam up, probably trump IBM on that count. But they have a way to go before they match RedHat in the heart and soul department.
Sorry if that seems a bit gushy. I do have a RedHat background. However, I’ve pretty much moved to Ubuntu for my desktop and am considering such a move for some of my customers.
In the end, though, it’s hard to find a publicly traded company more fundamentally devoted to OSS than is Redhat.
Edited 2007-03-09 19:49
Well said! Redhat is quite an amazing company and community both.
One thing I’m curious about is understanding Sun’s business strategy here.
I mean, it’s great that they’re opening Solaris, but where is the gain for their business model?
I can’t see it generating the same sort of collaborative model that Linux does due to the licensing and copyright assignment reasons. Don’t get me wrong, I’m not trying to start a whole CDDL/GPL type thing going on here. But I think the one point that is often overlooked in the success of the linux collaborative devleopment model is the lack of copyright assignment; I don’t think the corporate backers would be as willing to contribute if they had to assign rights to another party.
It’s great that Sun is contributing something this significant to the community, but what’s the business justification? How does this increase either sales or profits? If you look at the linux service support models for companies like IBM or Red Hat, their profit comes largely from the diminished development costs since that cost is spread across thousands of developers. Sun, on the other hand, owns Solaris and virtually funds the development internally. If Sun is unable to truly leverage the community to aid in development and improvement, it reduces their move to basically making openSolaris freeware instead of FOSS. Sure there will be community distros built around it, but Sun’s corporate flagship will have to set it’s own course and fund it’s own development. That’s why I’m wondering about the business motive.
Is this just to seed Solaris and hope that it leads to greater corporate sales? Is it to try and diminish linux’s growth by splitting the market for FLOSS *nix-based platforms? Is it a marketing ploy to generate community buzz knowing full well that little will change in the FLOSSosphere? Do they intend to open their development model to better encourage community contributions that they are willing to adopt without requiring proprietary rights?
Any one of those is a valid business strategy, I’m just not entirely sure where they’re heading. I keep seeing stories about how great this is going to be, and I don’t want to deride Sun’s contribution since I’ve always maintained that any OSS contribution serves to strengthen, but I haven’t seen anything equating this move into a solid business strategy for growth. And shareholder-owned multi-billion dollar corporations don’t do things without having a solid business plan behind it, lest the move been seen as impulsive or non-commital.
Java I can understand being GPL’d, I believe there were valid business reasons there over and above “embracing the community”, I just can’t see the angle on openSolaris yet.
Then again, I just saw Troy again recently, and maybe I’m simply being cynical and suspicious of strangers bearing gifts.
“I just can’t see the angle on openSolaris yet.”
No matter what sun got from resources and power, they cannot improve their solaris without inputs from contributing developers over the world who test their OS, hunt code bugs and possible contribute to solve problems(debug, code, design).
Remember how log was solaris using CDE, even as we are in year 2007 sun still use it; now what if sun was forced to open new markets for itself, say desktop market, how would they enter such a place with CDE.
The same is true for SPARC based machines; we all know that SPARCs are powerful now, but what if in the future they are not (like what happened with SGI RISC processors), then if at that moment they do not have x86 or x64 compatible versions of solaris they would suffer alot.
So maybe they are afraid from the future and thus they want to buy themselves the escape route ticket in advance!
It’s just my guess.
“ey cannot improve their solaris without inputs from contributing developers over the world who test their OS”
Well, OpenSolaris is on its way to creating a user community to help bring this project up to speed. Check out http://www.gnusolaris.org otherwise known as Nexenta.
I mean, it’s great that they’re opening Solaris, but where is the gain for their business model?
Easy, it is a logical expansion of their existing Solaris model; before, only a few people had access to Solaris via an NDA; Sun Microsystems never made money off selling Solaris, but instead maintenance contracts for big companies.
What Sun hope is that for the same sort of people who are enthusiastic about contributing to open source projects, will find that OpenSolaris would be a good project to to contribute to.
With that being said, they do need to do more in regards to man power contributing to OpenSolaris by boosting the number of programmers they have working on OpenSolaris; the number of bits of hardware on Solaris is appalling, and waiting for volunteers to do it for free is nothing more than exploiting any good will that might exist.
snv_59 is very unstable compared to the most stable build snv_55. 59 crash during gdm installation process no matter what gdm used (gnome, cde).
Their OS seems to be very sensitive to NIC changes; which would trigger registration process again.
Other than that it is amazing; sun was able to do 3x more faster development than Linux, even though linux is many years ahead of sun solaris in the arena of hardware support. Software wise, “pkg-get” simplify things alot together with “pkgadd” but the problem remains that SPARC based solaris is better treated than x86.
Nvidia drivers installation and configuration is un-necessary at all, because it is already installed and configured and this is never seen in any distro including those that contain drivers already (like xandros) because these distros were unable to configure the correct resolution for my monitor with 1920x1200x32@60Hz unless you touch the “xorg.conf” file.
Well Done Sun! and keep the good work
Bit off topic, though. I registered for the free Solaris 10. I am still waiting for the DVD. Can anyone point me where to post a bug? This is a bug, isn’t it?
Never saw my DVD either 🙁
Still waiting on mine as well.
Just to elliminate all doubt that Sun isn’t sending out free media, let me say that I’ve received both sets of media that I’ve requested: (1) Solaris 10 6/06 media for SPARC and x86 and (2) Solaris Express: Developer’s edition. Unfortunately, at the time I requested the SX:DE, no SPARC media was available, so I received a single DVD for x86, which I promptly handed off to a friend of mine.
I just got mine in the mail last weekend.
I had completely forgotten about it. Made my weekend 🙂
My understanding is that the demand for Solaris 10 DVDs were overwhelming and so an order for even more was placed. Those orders should be filled in the near future from what I heard.
I do have the feeling that Sun takes OSS more seriously and if they really bring opensolaris (and maybe java too) under the GPL v3, I think they would be the OSS king.
Especially if Linus stays with his argument that the kernel stays at GPL v2. I think that then a nice combo would be the opensolaris GPL3 kernel and the rest of the gnu userland.
UBEATABLE imho.
CDDL *is* an open-source license. It is compatible with many other open source licenses. GPLv2 is incompatible with many, whiich is a shame. It remains to be seen whether GPL3 is incompatible with other licenses. Remember that GPL does not coexist with other licenses – you have to relicense.
Why do you think that GPL is a ncesity to be ‘open source’? Is Apache not open source? Nor Boost?
Why do you think that GPL is a ncesity to be ‘open source’? Is Apache not open source? Nor Boost?
Many times Stallman has actually said not to get confused between the very liberal term of opensource and the GPL – to me, its up to the individual as to whether something is ‘truely opensource’.
For me, if I can download the source, add to it, and distribute it via a fork, then it is opensource source – as for ‘contributions back’ thats up to the individual; personally, if I was a company and embraced the *BSD kernel, the features I added might be hardware specific or something, so I might decide to donate 10% of the profits derived from the products sold to give back to the said *BSD project.
Although I do see a noble intent behind GPL, it however beings alot of things into question as to whether one can call it ‘free’ as in ‘libertarian’ or ‘free’ as in ‘freedom from poverty’ as many socialists might use – then again, thats up to the individual to make that judgement call.
It needs time and energy to build a community, and I am not sure if SUN has really understood that part of the game. It’s not enough to release software under some kind of open source license – that’s really only the beginning of a long and stony path. And I am not convinced SUN is really heading in that direction. They seem to have the need to urgently earn a lot of money and press really hard. But that’s not the way it usually works in the open source world …
It needs time and energy to build a community, and I am not sure if SUN has really understood that part of the game. It’s not enough to release software under some kind of open source license – that’s really only the beginning of a long and stony path. And I am not convinced SUN is really heading in that direction. They seem to have the need to urgently earn a lot of money and press really hard. But that’s not the way it usually works in the open source world …
Sun returned to quarterly growth a while back, and they posted a profitable year last year, with growth looking again to be a positive thing in this year.
If growth and profits are a negative thing that accompany the so-called ‘bad direction’ then I say keep going in that ‘bad direction’ because obviously Johno plan is obviously working – and where is IBM grandstanding now? well, Java is now opensourced; the ball is now in their court, what do they have to say?
Are we going to see them opensource DB2? AIX? Notes? Domino? I mean, if we’re going to talk about opensource contributions, when are we going to see IBM open up the crown jewels for people to pick and poke at?
This highlights something I’ve been thinking in the back of my mind as I’ve watched this discussion unfold. Sun’s contributions to OSS are far greater, IMO, than what I’ve seen from IBM, and I’ve been a customer of IBM for some time (TSM, AIX, pSeries, etc.). Sun has opened up several products that no one else in the industry has bothered with — namely, the Availability Suite (basically, Veritas Volume Replicator).
IBM may be a huge contributor to open source in terms of volume, but not in quality. Its not difficult for a corporation to hire several employees and developers to contribute to an already open, existing project, like Linux kernel development. What is difficult is exactly what Sun has done — opening up a once proprietary project, which took years of going through the code and looking for possible legal issues that could occur if that code were released in its current state. If you take a look at OpenSolaris.org, you’ll notice a lot of work has gone into educating new contributors (those outside of Sun) about internal procedures.
I’d challenge IBM to do one of two things: (1) encourage AIX development. Its difficult to ‘encourage’ development when the entire AIX product stack is non-free. For those of us that have POWER systems (I’m currently sitting behind a 7026-B80 with 2x375MHz POWER3-IIs and 2GB of memory), our only way of getting media is from eBay or finding a decent sales rep that’s familiar with the ‘one-time media purchase’. XLC is also very expensive and non-free. Note, that Sun has not yet opened Sun Studio, but they’ve still made it possible for anyone with interest to download it and use it. That’s what encourages development or (2) go beyond what Sun has done and open up AIX and the compiler suite, along with some of your Tivoli products, and educate new contributors on internal procedures and policies, as Sun has done with OpenSolaris.
Edited 2007-03-11 17:04
This highlights something I’ve been thinking in the back of my mind as I’ve watched this discussion unfold. Sun’s contributions to OSS are far greater, IMO, than what I’ve seen from IBM, and I’ve been a customer of IBM for some time (TSM, AIX, pSeries, etc.). Sun has opened up several products that no one else in the industry has bothered with — namely, the Availability Suite (basically, Veritas Volume Replicator).
Hence my curiosity as to why, at minimum, Lotus Smart Suite hasn’t been opened up – give that it doesn’t play a major role in their middleware – why not opensource it?
IBM may be a huge contributor to open source in terms of volume, but not in quality. Its not difficult for a corporation to hire several employees and developers to contribute to an already open, existing project, like Linux kernel development. What is difficult is exactly what Sun has done — opening up a once proprietary project, which took years of going through the code and looking for possible legal issues that could occur if that code were released in its current state. If you take a look at OpenSolaris.org, you’ll notice a lot of work has gone into educating new contributors (those outside of Sun) about internal procedures.
You’re right about that; what they have contributed has either been of little or no benefit to Linux (JFS anyone?) or simply a matter of stamping their name onto something that was going to occur irrespective of whether IBM got involved in the frist place.
It’ll keep getting better when Sun releases Solaris Enterprise System as opensource; a completely top to bottom stack of software for the enterprise, backed by a world class software and hardware vendor.
I’d challenge IBM to do one of two things: (1) encourage AIX development. Its difficult to ‘encourage’ development when the entire AIX product stack is non-free. For those of us that have POWER systems (I’m currently sitting behind a 7026-B80 with 2x375MHz POWER3-IIs and 2GB of memory), our only way of getting media is from eBay or finding a decent sales rep that’s familiar with the ‘one-time media purchase’. XLC is also very expensive and non-free. Note, that Sun has not yet opened Sun Studio, but they’ve still made it possible for anyone with interest to download it and use it. That’s what encourages development or (2) go beyond what Sun has done and open up AIX and the compiler suite, along with some of your Tivoli products, and educate new contributors on internal procedures and policies, as Sun has done with OpenSolaris.
I doubt it’ll ever happen; for Sun, opensource and openstandards aren’t new; its always been a part of the basic DNA make up of Sun, right from the very beginning when it was first founded.
For IBM, its just a glorified PR stunt, pull in the customers, sell massive overpriced services under the guise of ‘integration’, when infact its more like ‘money extraction’ – Linux happens to be the perfect vehicle to use for that; low cost, high margin, and when it comes to ‘integrating legacy with new’ IBM will promise the earth, cost the earth, but never tell you that you’re better off trashing the whole damn thing and replacing it with a whole new system that’ll yield a lower costing over the life time of the infrastructure.
As for Sun; OpenSolaris can be compiled with GCC, and at the end of the day, I think Sun realises one thing; the big movers within Solaris, are going to be in the x86 market, where GCC is a great performer, and given that in only a few areas where Sun’s own compilers really shine – on UltraSPARC hardware and a few niche areas like scientific computing, and the given improvement in the quality of code spat out by GCC, there is very little incentive to opensource it.
I’m still enjoying running Apple for Home apps
and NOW running Solaris 10 in fusion.
So, as to a Solaris laptop, you can look at Apple.
– Fusion is currently in beta, so it’s still running debug libraries.
( Hotter and a bit slower till they release the full product. )
The Java Desktop ought to have a Create User dialog under administration. I like and prefer to su to root, not run as root 100% of the time. But, I’m currently reading the man page for “adduser” so it’s probably not a big deal.
The Terminal program, GNome Terminal, isn’t as nice as Apples. But, that’s why Apple is Apple.
Looking forward to checking out DTrace.
Thanks Sun.
“The Java Desktop ought to have a Create User dialog under administration. I like and prefer to su to root, not run as root 100% of the time. But, I’m currently reading the man page for “adduser” so it’s probably not a big deal.”
When you first log into the desktop there pops up a window that execatly asks about a new user … (at least it was in snv_55 and it is still there in snv_59).