Thursday Mar 28, 2013


The big SPARC launch took place earlier this week. As usual with these events, it was a lot of fun to attend, even if no gimongous tents were involved. I was standing in the back, next to the control booth; as the music went up and the lights went down, next to me, the technical director quietly says into her mike, "Go." And this happens:

And that was pretty cool. But was that the "boom!" I'm referring to?  Nope. That looked more like this:

(Note: Subtitle does not refer to me)

So that's quite a "boom." With the launch of the new SPARC T5 and M5 series of servers, we've set over a dozen new performance records, and shown that back in 2010 Oracle did indeed establish a SPARC roadmap that it could execute on.

Needless to say, IBM has kicked up quite a fuss about this. I would too, if I were them; they thought they had this thing licked. And a pretty funny tweet related to that:

But--the thing I noticed was that as much as IBM tried to talk down the performance specs we demonstrated, they were noticeably silent about what, if any, advantage they might have in the software space. And of course, that includes the OS. When I talk to customers, the truism is almost always that when they use IBM systems, they don't do so because of AIX, but despite AIX.  (Remember "Watson"? Do you know what IBM operating system IBM doesn't run on it?)

Meanwhile, Oracle Solaris remains the gold standard for enterprise OSes.  "If it has to run, it runs on Solaris" is the catchphrase in data centers all over the world. We were flattered when IBM embarked on what some called "Project Copy Solaris," a few years after Solaris 10 came out, but couldn't help but notice they didn't quite hit the mark.  And now Oracle Solaris 11.1 is the current release, and we've done a lot more that no one else seems to be finding the recipe for.

And of course, we've got a big advantage, that we've encapsulated as: Hardware and Software, Engineered to Work Together. Over the last few years, we've been able to demonstrate more and more what that means: at Sun, obviously we did a lot of work to make sure SPARC and Solaris were in sync, but meetings with the Oracle software teams were special events, held only a few times a year, and preceded by phalanxes of lawyers engaging to make sure everyone knew just what could and couldn't be discussed.

That's been replaced by something called: "going to work together every day." And it's paying off. And we can prove it.

We've got some new pages that go into more detail about what's new with Oracle Solaris. Maybe this will be a handy "cheat sheet" for OS developers at other companies. The first site talks about how we've optimized Oracle Solaris to get the most out of this new generation of SPARC processors; the other talks about how the continuous co-engineering process between Oracle Solaris and Oracle database works to the benefit of both.

Tuesday Mar 26, 2013

There Were Giant Flying Businessmen in Those Days

One of the side effects of choosing the name "Solaris" for our OS is that we share it with several other companies in unrelated fields: sunglasses, snack bars, roofing shingles and the like--and the book, of course. I hear there were even movies.

Just for fun, I collected a few images of these on our Facebook page in the album "Sorry, Wrong Solaris," and other people have sent in additional ones over time.

Last night, my cohort Rick Ramsey, he of BigAdmin-which-is-now-OTN-Garage, sent me another one: a Brazilian construction equipment leasing company. "Huh," says I. "Perhaps not the same strong connection to operating systems as, say, a bus company, but let's take a look."

Hmm, what's this?

It turns out this particular "Solaris" company was responsible for doing the site preparations for Cirque du Soleil as they toured South America. And, that tent... there's a SPARC/Solaris connection, there.

Enter Ultra Man

Here we are, hours away from a major SPARC/Solaris-based product event. Those of you who have been kicking around the industry for a while may remember that way back in November 1995, Sun had another event: the launch of Sun's first 64-bit SPARC processor, UltraSPARC.  And especially back then, events like that needed an interesting venue.

Someone's great idea: what if we held it in a really big tent?  And who's got a big tent they don't happen to be using at the moment?

...Cirque du Soleil did! This is one of their tents, set up for the launch event in the parking lot of the then-brand-new Sun Menlo Park campus (alumni will recognize the fitness center in the background). The story about how this event was set up and executed is fascinating, at least to a computer industry/show biz geek like me; leveling out an area and setting up a huge circus tent to hold a one-hour press event with 1,500 people in attendance, then tearing it down again immediately. And the industry hadn't even made it to dot-com madness yet.

The logo for the event was "Ultra Man" (not to be confused with "Ultraman"), "a man, stylized, bespectacled, business suited, legs tapering into flames of rocket fire as if he were launched from a cannon."

That would be he.

There was even a guitarist opening the event, a former Sun employee who knocked it out of the park. What was frustrating to me personally was that this all happened literally days before I moved to Northern California, so I had to settle for watching it via closed circuit TV. Webcast? Yeah, didn't really have those back then, so we instead had satellite downlinks to various Sun offices, which hosted local viewings of the event. That did make it more of a social thing.

As I add this to the Facebook album, the final bit of serendipity, as many of you may already be thinking, is that the site of this 1995 event is today in fact the Facebook headquarters campus.

Today's event isn't going to be quite so elaborate; we've already got a perfectly useful conference center to hold it in. But even if we don't have giant flying bespectacled businessmen waiting in the wings (that I know of), the product news should be very interesting. I don't know if either Larry or John play guitar.

Tuesday Mar 19, 2013

More on the March 26 Webcast: "New SPARC Servers with the World's Fastest Microprocessor"

Let's be real.

If, in March 2010, Oracle had made an announcement about the "world's fastest microprocessor," you might have been left scratching your head.

It turns out that SPARC T3 did come out later that year, and it was a significant boost over the existing UltraSPARC T2 --"a 2x increase in performance over the previous generation of SPARC T-Series systems" -- but no one was saying it was the industry's fastest. However, what was notable was that its release date and performance lined up with the SPARC roadmap issued at the time of the Sun acquisition, which isn't always the norm in this business.

If, in March 2011,  Oracle had made an announcement about the "world's fastest microprocessor," you still might have been skeptical.

But yet again, just a few months later, we announced the SPARC T4, which did set a world record in Java performance.  And it not only kept to the 2010 SPARC roadmap, it exceeded it, smashing the performance levels of the T3. Hmm.

But now it's March 2013.  And the title of the event is "Announcing New SPARC Servers with the World’s Fastest Microprocessor."

That's pretty bold... but given what's been happening in recent SPARC history, is it as hard to believe as it might have been three years ago?

Tune in and find out.

Oracle Live Event: Announcing New SPARC Servers with the World's Fastest Microprocessor
Tuesday March 26, 2013
1:00pm PT / 4pm ET



UPDATE: There will be all sorts of people live-tweeting the event, using hashtag #SPARC. Got questions as you watch?  Send us a tweet, including that hashtag.

Friday Mar 15, 2013

"uptime: 3737 days..."

What were you doing in late 2002?  Well, these guys were booting up this SPARC/Solaris server. They finally had to shut it down for the first time in over 10 years -- to move it.

Some thoughts:

1. This was Solaris 9. More uptime goodness added in 10 and 11... and wait till you see what we've got up our collective sleeve.

2. "But what about all those critical kernel fixes they missed?", you might ask.  Yes, that's a challenge. See the last part of #1.

3. Glad we fixed the lbolt thing.

4. They brought it down just to move it? Would George Costanza have accepted that as a reason?

5. Hmm; over 10 years ago. Wasn't that about when Steve Ballmer proudly announced the new Windows Server reliability feature that let you schedule automatic nightly reboots? (Unix people might have called that exciting new feature "cron.")

Thursday Mar 14, 2013

It's That Time Again: Share Your Great Solaris Story at Oracle OpenWorld SF

The Oracle OpenWorld Call for Proposals is now open, through April 12th. What does this mean to you, the Solaris person? This is a chance for you to share your experience with your peers, and to interact with some of the top Oracle Solaris experts in the world.

Oracle OpenWorld San Francisco is a unique experience in a unique city, and if your proposal is accepted, you'll get a complimentary pass to the show (you can also submit proposals for JavaOne and MySQL Connect).  OOW is September 22nd-26th this year.

Learn more about Oracle OpenWorld 2013

Tuesday Mar 05, 2013

March 26 Webcast: "Breakthrough with Extreme Performance"

Save the date -- on Tuesday, March 26th, join Larry Ellison and John Fowler as Oracle unveils a new standard for extreme performance. There will be a live webcast at 1:00 PM PT, or if you happen to be in the San Francisco Bay Area that day, you can register to attend in person at the Oracle Conference Center in Redwood Shores.

Oracle Live Event: Breakthrough with Extreme Performance
Tuesday March 26, 2013
1:00pm PT / 4pm ET




Security. Speed. Simplicity.
An efficient, open, affordable cloud platform for SPARC and x86 systems.


« March 2013 »