The original cloud
A mainframe is the original cloud
Marc Benioff, chairman and CEO of Salesforce, has fired back at Larry Ellison and his beloved Exadata for cloud computing, saying that the server fails the cloud test and that Oracle lacks a genuine vision. Working — and walking — a crowd of 14,000 at his annual Dreamforce conference in San Francisco, Benioff championed a world …
A mainframe is the original cloud
...but passes the 'runs wildly successful very wealthy company' test quite well. I wonder what he's doing wrong?
Man, talk about a decent game of buzz-word bingo. It was like listening to three web designers try to explain what Web 2.0 means. I really don't get what all the folderol is; you have server(s), you have clients, you have data flowing between them. As someone who's still using the same login name I was using for bbs'ing back in the early 80's, I'm not seeing a whole lot of difference. Sure, there's some new shiny and clients are getting thinner (physically) and more mobile but, um, d'uh! That's how things have been going for the last 40 years.
/beer, 'cause I like beer.
But Benioff is gonna lose this argument. Ellison's pitch simply boils down to big IT shops - banks, governments, whatever - seeing the benefit in the whole quick-provision-of-a-small-slice-of-our-overall-computing-capacity model, but done privately for security reasons, ie not done under someone else's ultimate ownership.
Pretty simple, really. It's amazing that Benioff doesn't get it, given how simple he appears to be.
Given security and legal considerations, I'm not surprised that many companies are not enthusiastic about dumping their databases on the cloud.
But that Oracle is thinking like a mainframe company? Yes - that seemed obvious to me back when they bought Sun. What on Earth would they want with the dying SPARC architecture (from the viewpoint that everything except x86 is dying, except for architectures like MIPS and ARM, presumed suitable for small-scale and embedded computing only) - except to have their very own proprietary boxes with which to compete head-to-head with IBM, so that they can make their profits on the hardware instead of the software (helps stop piracy, makes things look more reasonable to accountants, makes it easier to compare an Oracle bid with an IBM bid)?
If one thinks that IBM is not long for this world, Larry Ellison is wrong. If one things that the cloud may be just a passing fad, Larry Ellison is right. I'd be inclined to bet on Larry Ellison being right.
Youngdog said on Thursday 1st January 1970 00:00 GMT
Submitted at Wednesday 8th December 2010 07:06 GMT
..Red Cube acquisitions are bad but 'build software that looks like Facebook'.....
I predicted Facebook BEFORE I WAS EVEN BORN!
How's THAT for innovation!
..Red Cube acquisitions are bad but 'build software that looks like Facebook' is good?
Neither of these chumps is going to find real innovation without first pulling their heads out of their arse...
Public clouds aren't for everyone. Private & public clouds will stay.
And SPARC, AMD and any other alternative choice is necessary. Without it, Intel will stagnate.
So, when possible, I always choose something different.
Supporting diversity helps everybody.
* Intel copies SPARC with their multi-core processors,
* Intel was last, who stopped talking about MHzs???
OMFG, I need those like I need a hole in the head.
FFS, Benioff, I want to decide what data I'm given (as least at a high level). Not have "The Cloud" decide it for me.
Otherwise, I know what I'll be getting.
Spamvertising. And we know how well that went down.
'coz these guys need to go back to skool.
Nobody wants push services. It's the web equivalent of cold calling: It's intrusive, open to abuse and allows big business a degree of control over the channel.
Sales people love it because it's intrusive, open to abuse and allows big business a degree of control over the channel.
" * Intel copies SPARC with their multi-core processors, "
Err, multicore design was created by IBM back in 2001 used in POWER4 chips.
It depends on how you define multicore. If multcore means multiple cores o one piece of silicone then Sun was there before IBM. If you define it by multiple cores in a socket then IBM got there first. You gotta admit that Sun took it further than the rest though.