Slashdot Mirror
What's Coming in Solaris 10
raptor21 writes "Ace's hardware has an article with feature list of technologies in Solaris 10 or whatever it is called today. Interesting stuff like DTrace, FireEngine, military grade security and a new filesystem called ZFS, Zetabyte File System."
Buzzwords like "DTrace" and "Fire Engine TCP/IP" don't sound very useful, they sound like bloat. And who needs a zetabyte filesystem? We haven't even reached petyabytes, for $DEITY's sake!
The only bright spot is that Sun is releasing it for x86-64 platforms as well as x86, so all of this "oh-so-wonderful" stuff can be used on Opteron boxen.
SCREW THE ADS! http://adblock.mozdev.org/ Proud user of teh Fox of Fire - Registered Linux User #289618
Wait until you see what is coming down the pipe from Sun. The new Athlon 64 workstations and servers will breathe new life into Solaris. For the first time you will be able to run a fully 64-bit kernel with all of the stability and reliability of Solaris, along with all of the advanced features of Solaris. Features like this won't even make their way into Linux for another 5 years or so.
Solaris 10 will be the first release of Solaris that supports native 64-bit mode on the new AMD Opteron and Athlon 64 processors.
Not to mention the ability to address terabytes of memory without using PAE hacks.
The only question in my mind is: Will you be able to run the IA-64 port of Solaris 10 on a home-built Athlon 64 box, or will it require Sun hardware to run?
"When the president does it, that means it's not illegal." - Richard M. Nixon
That's what I want: a free fire engine with every mainframe I buy. A nice Tonka fire engine would sway my decision towards Sun products, especially if they threw in a fire man's hat.
Woo hoo! Off to play with my toy trains!
Thank you for your time.
Sincerely yours,
The CTO of your company
That's "Mr. Soulless Automaton" to you, Bub.
People don't pay through the nose for the software. They pay through the nose for the hardware which is second to none IMHO. Where I work the AC went out in the datacenter over a weekend. Unfortunately, the first box to go down was the intel box that monitors everything else. On Monday the only boxes still running were the Sun boxes. It was over 120 degrees in the datacenter. It was approaching 200 degrees inside some of those Sun boxes.
According to the related article, it includes a complete rewrite of the TCP/IP stack. Conventional wisdom has it that all TCP/IP stacks out there borrowed heavily from the BSD code.
Will Fire Engine then be the first non-BSD TCP/IP stack?
This is a case of people actually providing a product that gives the customers something new and exciting.
Bravo SUN. And they recognize Linux as having a place.
To be honest, I'd rather have a SUN monopoly than a MS monopoly. At least the software would be a bit more stable.
The US military doesn't go in much for booby-traps.
In the event that a secure installation seems about to be overrun the sensitive equipment is stacked up and destroyed with WP grenades.
You'll be sad to learn that WP grenades don't explode.
Sorry to disappoint.
-Peter
I was a computer operator on a carrier long long ago. The computer room was two decks under the flight deck, right under the arresting gear ... pretty high up. My abandon ship station was to take a fire ax and whack the computer. I guess that old supply computer had too much vital technology. It wasn't the data in the computer, because it had no permanent storage, it was a tape operating system, and my job did not include whacking the tapes. So they were more concerned with the enemy capturing our carrier and recovering the computer technology than recovering the records of how much toilet paper we used. Must have been the water tight seals around the tape drive doors -- they claimed it could operate under water, tho how deep I never heard. And being so high up, 40 or so feet above the waterline ... if that had ever gotten under water, I wasn't planning on being the duty operator.
Infuriate left and right
All the talk lately has been about how Sun is on the fast-track to irrelevance, getting creamed from below by Windows and Linux, and from above by IBM. They splashed onto the scene years ago by being "faster, cheaper, and better" than DEC etc., and now that they are no longer any of those three, it's unclear what they offer - other than Java, which, while interesting, isn't part of their core "mid-to-high-range" server business. A lot of idiotic hot air about thin clients and network computing hasn't helped matters.
So are they going to go the way of DEC? Perhaps soon Gateway or Dell will be interested in an acquisition? Who knows. There are some things that look like signs of life, though.
One is dealing with AMD. This is very smart for both companies. Sun will package their excellent new chip in well-reputed new lavender-and-white boxes, and customers will run "Linux or Solaris." If they can do this cheaply enough, this could finally be a competitive angle for new business, on the low end...
But the "Linux or Solaris" is the interesting point. Right now Sun seems to make its money selling hardware and software (read: Solaris). Traditionally Linux - which is surely eating their breakfast, lunch, and dinner - has been Sun Enemy #2 (right behind Mr. Bill). And yet Linux may be exactly what they need - an open platform upon which to build a Sun-branded infrastructure that is "better, faster, and cheaper" than the competition.
And you get the feeling that some inside Sun, at least, know this. But clearly others do not. And the fight goes on, and on... and you get things like this:
The net result seems likely to be that Sun will promote Solaris ahead of Linux ("cheaper and better") except to customers who already have a large Linux installed base or they specifically want Linux solutions. For Sun customers with SPARC servers, using the same operating system across all servers would make life simpler as well - less complexity means lower TCO. By the end of the year, Sun also plans to publish benchmarks showing Solaris x86 matching, and in some cases beating, Linux on identical hardware.
This is an article detailing a strong feature-offensive against Linux - trying to make Solaris into a superiour solution that people will pay a premium for.
But I am wondering if this is such a good idea. Why not adopt Linux, and contribute this work to it, rather than try to compete with it?
The scenario with the sparc chip and the Opteron is perhaps an object lesson. It takes billions to keep your hand in the chip game. Sun is perhaps not in a very good position to continue to spend that money. The AMD move looks like the first step in a viable exit strategy. Partner with a chipmaker - a "cheaper, better, and faster" one like AMD - since the economics of trying to beat them all don't look as good.
Take Sun's software shop. While well-honed for its common uses, any given little piece of Sun's proprietary Unix fork is generally sub-par - from their SSH implementation (just found a bug in that the other day) to their shell. Clearly it is not Sun's highest priority to deal with all these little Unix details. So most people pile on open-source solutions to make it livable. But then where is the support? Yes, the world still tolerates this state of affairs while tending to their Apache and Oracle servers... but for how much longer?
Could Sun become a "services company" in the mold of IBM and Redhat, "building solutions" with AMD and Linux for the server market that are cheaper and better than competitors? Isn't there money to be made "just making these things work" for all segments of the server market - building and supporting? Wouldn't they be well positioned to do it?
In a sense this seems like a hedge. They will try to do both - package Linux and support their Solaris business. They may be waiting to see which one will thrive; they may think it's better to be in both games. But does that make sense, even now? They are hemorrhaging ca
Want to Know How to Cheat the GPL? Read On!
I was told that SUN was primarily a hardware company, and that the more exposure they got, even from software, would create more hardware sales
:-)
Because if customers want to use Sun systems for their shop, but there isn't an office product that runs on it they still have to have PC's lying around. If Sun supplies them with the hardware, OS, and decent office tools, it's yet another reason the customer can use Sun.
suddenly SUN was a wee bit worried. They tried Solaris 9 for x86, then pulled it back later on
Uhm. I've got a copy of Solaris 2.6 x86 downstairs in my software library. If you think that Solaris 9 was the first x86 release of Solaris, you're not very educated on Sun products/offerings. The reason Sun "pulled back" from x86 is because they were ready to relinquish the x86 market to Linux. Customers SCREAMED at Sun NOT to do this. They WANTED Solaris reliability and functionality on x86 CPU's and didn't trust Linux completely. Sun happily obliged.
Does anyone else think that they're competing with themselves?
Huh?
They're not a software-as-a-service business model. They're not really even an OS Software "manufacturer" business. They're a hardware company who has tried their hand at everything from a programming language (Java), an office suite (staroffice), and OS/desktop (Solaris, Java Desktop).
Yes. Your one-stop-shopping place for all of your workplace needs. You need the hardware? Got that. You need an OS that offers seamless integration with the hardware? Here 'ya go. Want a built-in filesystem with the features of VXFS without having to pay a license fee to Veritas? ZFS comes in 10. Want to write your code in one language and run it on all of your other systems? Use Java. OS Desktop? That's just icing designed to take more $ from Bill G's pocket.
When Linux pulls through
Linux is a good OS and I am no stranger to it whatsoever, but it has a long way to go to catch up to Solaris. This announcement about Solaris 10 is demonstrating just that.
Oh, and by the way. Some of us in my office are playing with the internal-only betas of Solaris 10. Very sexy IMNSHO. For the heck of it, I started calling it SunOS X as a parody of MacOS X. The rest of the engineers on my team have followed suit, though as of yet none of us know what the "official" release name will be.
***
Disclaimer
***
I DO work for Sun but this is my PERSONAL opinion. It is NOT intended in any way, shape, or form to be construed as an official Sun position.
- Kate
"DNA is life. The rest is just translation."
Low end is last year's high end.
I'm looking at a 12CPU E4500 loaded with sub GHz Ultra2 chips. They paid a LOT of money for that 2 years ago. It's ok for databases, and has terrific IO (if you use non-Sun disk systems), but how long until a 4Way Opteron comes out and smokes it? For Looking at rows of V880s (a waste of money, IMHO - fiber channel disk? pure price inflation). About $1 million dollars worth. In two years people will roll their eyes when they have to use those.
Sun's got themselves in a bad situation.
Linux works on the desktop and that means that there are THOUSANDS of folks < 30 years old who know Linux (and BSD) pretty well. Moving up in seniority and skills in the decision trees of companies.
Moving to Solaris means giving up /usr/ports/ or pkgsrc (BSD), means giving up a userland with tools developed since 1992 ("df -h" is FAR more handy on a machine with 2 TB of disk attached from the SAN than Suns "df -k" and counting).
If someone strong in Open Source OSs can make a case that using a 4x3GHz Intel box for all but the MOST high end (and costly) services, then Sun loses MORE ground.
Shall we talk about SGI's far far superior hardware and OS features and how strong THEY are in the market place? Want an 8 CPU SGI? Take the 4CPU one you have, get another, join them. All the way up to 512 CPUs. Want to start with a 4CPU Sun and move up to a 64CPU Sun? Sorry, you buy a mostly empty chassis for several $100 thousand and hope that you can still get the CPUs later on.
(my 4800 with 800MHz CPUs won't work if I add the new 1200MHz CPUs - I either replace them all or scrounge for old ones. Nice. Almost a year old too).
Sun won, in large part, the Workstation wars - the high end desktops where PCs just couldn't compete. Built in Ethernet, graphics and SCSI, for around $15k. Hows the PC world going to compete with that?
"Sun will never sell a $2000 computer - that's not the space they want to be in. - they make servers and workstations."
They've lost the desktop.
They have to innovate to keep the server space.
Hardware partitioning is KEY - a 48 way machine is mainly useless to me, but being able to chunk it - dynamically - into (say) 30 or even 50 virtual machines in very attractive.
We have a big thing going on next month where we're building up a pair of boxes to handle a HUGE load that will be for just the month.
If I had a monster machine and could say: "Oh, give those servers 24 processors during the day, but only 8 at night so the other systems can run their stuff" then we'd save money. Now.
I'd free up 6-7 people's time for 2 weeks and that all factors into ROI.
- -
We're already throwing around terrabytes of disk on SANS from machine to machine as required. Why not with CPUs and memory?
Is it just me or does a lot of what Sun is doing remind you of what SGI went through in the high-end visualization market a few years back?
Obviously things are not DIRECTLY equatable but I can't stop thinking about the comparison.
Couldnt you say that in both cases that their niche erroded due to low or no cost competitors?
Both had some great software. Could Sun having Solaris and Java be somewhat equateble to SGI's OpenGL and Irix?
Both companies had hardware at the heart of their business models at one point.
Sun seems to be doing what SGI did in trying to do a bunch of different things to pull itself out while in the process losing focus and STILL having hardware at the heart of the business model.
SGI is obviously still around. If you look at their website now, you can see they are targetting a much smaller niche than they used to (supercomputers). The day of thinking that an o2 will be on the desk of every college student has long passed. I'm sure SGI never thought they would be promoting Linux-based supercomputers on their homepage 5 years ago - lord only knows what Sun will have on theirs homepage 5 years hence.
-_-