Saturday, September 5, 2009

TechnoFads: XML, Clouds, Snake Oil and Elvis

After reading about the cloud services hoo-hah lately, especially eminating from VMWorld and from the various press outlets following up on the Google Gmail outage, I thought about the 50,000 foot level view of this a bit.  This whole cloud thing is following the same hype-vs-reality curve that XML followed.  Indeed, many past buzzword frenzies have followed this curve.  AI, neural networks, 3D-everything, virtualization, universal authentication, and so on.

Remember when XML first started flowing out of the press years ago?  It was going to save the world.  Everything was going XML.  Even database engines and operating systems were going to be completely based upon, and built with XML.  Yeah, sure.  Yes, I know about XPath, so STFU please.  What happens with the hype curve is that there is an immediate latching-on to the craze and excitement, followed by a self-fueling frenzy to be the first to stake out new territory where the new wunderkind stuff will make a dent.  In most cases it doesn’t take hold, but the rush to get there first is where the frenzy comes from.

In the 1990’s it was all about “object-oriented”.  Everything was going to become and “object”, whether it liked it or not.  “Behold, you have become an object.  Shed your procedural ways, you heathen!”  This overzealous knee-jerking has two sides to it: good and bad. 

On the good: It fuels enthusiasm, which fuels interest and effort to rapidly explore and therefore mature the technology sooner.

On the bad: It gives non-techies reason to laugh, mock and scoff at the next great idea.  Knowing that it too will likely never pan out.

The good side always falls squarely within the tech/geek world, which makes sense.  But the latter side.  The bad side.  That rests mostly in the financialy, and marketing world.  Those are the folks holding the checkbooks, if you recall.  So every time the techies cry wolf about the next “great idea”, and it doesn’t ultimately pan out as originally promised, it simply tightens the grip on those checkbooks.  Your next funding request just became harder to push through.

It’s happening again with “cloud services”, especially cloud storage and BI services.

Remember, XML was going to do it all.  Be all things to everyone at all times for all reasons and needs.  Obviously, it wasn’t.  It was overhyped and everyone who had actually worked with it, knew it.  But we weren’t the ones with the microphone at the time.  The folks that saw Elvis at Burger King had the microphone and they ran with it, like crazed stoned-out hippies at a Woodstock mudfest.  “Get yer free shrooms!!!”  oh boy.

Sure, XML has finally found its niches and done rather well.  There are still some areas where it may not fit as well as initially thought, but overall, it’s doing pretty well and making sense where it’s being used.  But the idea that because something is “new” automatically makes it “better” is completely illogical. Just like humans.

So, back to “cloud services”… There are two basic needs for which it is being aimed at but only one really fits at this point.  The two are “online” and “offline”.  Nobody wants to admit that however.  Which is rather odd, to me at least.  It’s all being touted as “internal” vs “external” and “private cloud services”, blah blah blah.  I really wish marketing people would have to have worked in the related field for at least 5 years before they’re allowed to market it.  Just think how different the world would be if that were true.

It’s not Internal vs Exernal.  It’s Online vs Offline

“Online” refers to systems and services which are directly used by business operations.  “Offline” are those systems which are not directly used, but which reinforce or provide backup for the “online” systems.  Pretty simple, huh?

Some examples of “potential” online cloud processing would be things like SalesForce.com, Live Mesh and Google Docs.  But more than just data, there’s the application aspect to this.  Think of how something like running AutoCAD over the Internet might play out for an engineering firm with 50 users.  What happens when that “system” goes down?  I’m sure local caching failure techniques will be considered to help mitigate such events, but you have to come back to the “why?” question.  Why do it in the first place?

After repeated failures of cloud services like Yahoo, Google, Twitter, Facebook, and many others, nobody even bats an eye at the notion of moving their online business needs into the clouds.  Complete insanity!  Anyone that ignores those factors is putting a loaded gun up to their forehead while smiling and reassuring everyone around them that all is well.

Vendors love the idea.  They gain greater control over the applications (because they host them, duh), and the licensing (because you don’t get to install anything) and they don’t have to pay for auditing and compliance headaches.  Much better for them.  Much more risky for the customers however.

Just as XML found its way into common sense places, so to will cloud services.  One obvious niche being filled is in the area of data backup.  Think Mozy, Carbonite, Amazon S3, and so on.  Running your daily business processes directly from cloud services is just too risky.  Not just at “this point” in time, but anytime.  It’s just not a good idea from a reliability and risk standpoint.  However, using it to maintain offsite, offline backups of critical business data makes perfectly good sense.  Sure, cost will play a huge role in the decision process, but that points the gun at the CFO not the CTO, so be careful when that argument comes up.  Eventually, it will come up. 

Vendors like Microsoft, Amazon, Google, EMC, Cisco and IBM will push much harder to squeeze every hype-ridden drop of revenue out of their cloud service offerings by laying it on thick with your company’s golf-playing suit-wearing MBA folks.  And just like they do when getting off their air trip, they bring up some interesting article from the skymagazine about why they should replace their entire computing platform with a different one, they will get that Kool-Aid drink of cloud crap as well.

Cloud services make good sense for augmenting or outright replacing offline systems, but not for online systems.  At least not for medium-to-large (enterprise) customer environments.  For individual and small businesses it might work well enough to be practical.

One last thought:  What would happen if the ISP’s (Cox, ComCast, etc.) decided they were tired of being nothing more than a conduit and felt like jumping in to capture some of this market as well?  Some already are getting into remote backup and storage.  It could get interesting for other cloud services.  Let’s just hope the hype curve is normalized soon.

No comments: