Stephen nails it as usual.
PHP isn’t suddenly “enterprise ready” because major vendors are supporting it. One the contrary, major vendors are supporting it, and other P scripting languages, because these technologies are already being used… and often by the smartest developers. Vendors can now make a buck supporting these environments… and offering “certified stacks” that “reduce risk”… because a market has been created already.
Its about what’s under the waterline? How can a vendor sail around the iceberg, without making like the Titanic?
Its not like nobody used Linux before IBM started selling boxes that ran the OS. Or Windows for that matter – remember when it was shareware, and flew under the radar screens of purchasing departments and IT departments? Yes we should all have the tee-shirt…
All that said, IBM’s is still a king-maker for enterprise technology. When IBM says a technology is ready for the enterprise then, by definition, it is. That is the Big Blue imprimatur. Its amazing what 100 years of keeping promises can do for your trust rating. Some recent IBM stamps then:
And today’s big news? “Blogging is enterprise ready“. Its not just BusinessWeek readers that need to think about blogs – so do IT staffs. Go IBM. I am looking forward to the nightmare of a bunch of new cool feeds i need to read. dagnabbit…
IBM was not first. But why should it be? If the initiative becomes popular IBM will have more bloggers than all the other firms in tech put together, in pretty short order. This is going to be very interesting to watch, especially from the perspective of somone that has been known to claim they know IBM as well as, if not better than, anyone else out there… The volume of IBM information I parse is already stupidly big. Well now, i need to step up a gear or two. IBM blogging as a personal challenge. Will this madness never end?
Jaime Cardoso says:
May 16, 2005 at 6:43 pm
It’s the old “Is this a fashionable business” issue again. “Some of us” (that means me but I’m pretty sure there must be more people out there) still think it’s the technical virtues of a product that make it “Enterprise Ready” and, PHP still has some shortages that will cost you extra in the long run (A web presence whithout strong HA features?)
sogrady says:
May 16, 2005 at 8:04 pm
i dunno, Jaime. i don’t think it’s a matter of fashion. i think it’s more a recognition that many web sites – SMB, departmental, divisional and otherwise – won’t require HA features. they just need to get built quickly and easily. for that, PHP is a better solution than J2EE and other HA options.
Jaime Cardoso says:
May 16, 2005 at 9:04 pm
Well, I’d try to avoid anything I couldn’t evolve into an HA configuration. I’ve been in to many interventions made during business hours (who cares if this stays down for a couple of hours) just to discover that hell had broke lose because the entire organization was using that “stupid useless service” (the most famous of those incidents was when email was starting to be used).
A service is not critical until it goes down and someone calls us at 2AM. The first time, it can’t be helped but, the second, ….
Dennis Howlett says:
May 18, 2005 at 7:48 pm
Jamie’s bang on the money but I think this can go a lot further. It’s just that ‘we’ haven’t quite figured it out…yet.
James Governor says:
May 18, 2005 at 11:48 pm
hello Dennis. welcome aboard. I think Stephen and Jaime both make excellent points. Perhaps we need to segment between “enterprise” or “B2C” apps, which *potentially* will need to scale, and those that are more like the old “departmental” application. I am certainly not arguing that PHP won’t scale. But does *everything* now need to be HA? That sort of uber redundancy worries me a little. Whatever happened to capacity planning, one of Jaime’s hobby horses. Just as Jaime argued Sun needed smarter models for storage sales based on how apps are going to be used, so similar considerations should colour platform choices. I don’t remember anyone saying they needed 5 9s reliability for Access-based applications, but would you really claim they by definition aren’t mission critical? Suffice to say it will be interesting to see this play out. With reference to Stephen’s recent post on C++ for example. And questions like why does flickr have problems with availability and what is Yahoo going to do about it?
Nigel Fortlage says:
May 19, 2005 at 5:32 am
Gentlemen, I think James is correct when he states that not everything needs to be HA and Jamie, if everyone did have HA and it worked as well as we all would want it, wouldn’t that take away from some of the work you get paid to do? Not that is a justification, but rather a business assumption of risk. My example to put it in perspective: In doing systems evaluation for consolidation we looked at every machine and service we used. To the executive and owner of our firm, Instant Messaging was not mission critical. But when we take down IM for any number of reasons my phone rings off the hook with 45 (50% of staff) associates who yell at me because I took away their life line inside the organization.
So as a business risk that system does not need to be HA, but to the users they can’t live without it. Which one is right?
No matter what angle you take in the end the business wins and it is not an application that should be on a HA solution. Why, because the business risk is low.
Sorry I did want to comment on PHP because we have used it for a few years and no HA is not a big issue for us on those apps, nor cost effective, so it is a great, thin, functional environment for what we need.
IBM is just following through on their business plan (read: gain revenue) by adding it to their list of supported environments.