More Clouds Gather on the Horizon

Just like SOA? Just like XML? Just like CORBA?


Sorry, but to me, it still looks like this is being pushed by vendors with not quite the level of real demand by potential customers that the buzz would indicate. "If you build it they will come" was a line from a cheesy movie and doesn't always make for a sound business strategy.


I think that the biggest challenge presented for cloud computing is the same as it was for SOA and CORBA... so many vendors piling in at once they create confusion in the market. There isn't "a cloud", but "several different skies each with their own clouds".

Jim, I don't think cloud computing compares with Corba or XML, nor even SOA. The fact that some technologies were trumpeted as the "future" and they failed to achieve that status, it does not mean another technology or computing paradigm won't be the "future". Who believed cars would replace horses? Few did. Who saw the phenomenal result of the transistor invention? Few did. The PC, the Internet, the email, social sites, or Twitter?



I consider cloud computing as the future of "enterprise" computing, not the future of computing in general. And that's because it really makes sense. Who is going to spend huge amounts of money upfront on equipment and software, then pay lots of money for IT personnel when all is going to be available much cheaper and on a monthly basis? This is not going to happen tomorrow because all enterprises have already invested hugely in their equipment, but they will certainly start to move gradually to the cloud, be it a private or a public one.

I'm not buying that line either.

Clouds are the 'killer app' for SOA.



Forget all that same-old-never-ending-cycle of vendor BS about 'business analysts creating processes in BPEL editors then deploying them', forget about the 'reuse of coarse-grained services', forget about 'more agile projects' (with WS-* ? yeah, right !). There are no real cost savings in SOA as it stands, it's just good old fashioned strong cohesion / loose coupling over a new technology stack.



Moving to the cloud is going to save every organisation big $$, and you have to have SOA in place to get to Cloud 9, so the COO is going to light a fire under your sorry hackerz until they get SOA running into the virtualized skies. At last, the latest cycle of technology hype finds a realistsic business plan.



Even governments are dreaming of the G-Cloud (para 27)



Mik

The fundamental problem is that no one bothers to even define "cloud" or agree on what it is.



First, vendors such as Amazon release a hosting option (very similar to what ISPs have offered for years), and called it "cloud". While EC2 is pretty cool stuff, does it mean that "cloud" means "hosting"?



Then, other vendors provide applications on the web, and call it "cloud". While Salesforce.com is great stuff, it's been around since '99 and we used to just call it an "application". Does "cloud" mean hosted application?



I think "cloud" means "investor opportunity" ;-) .. but that's not to say that great stuff isn't coming out of the latest buzzword.



Peace,



Cameron Purdy | Oracle Coherence

coherence.oracle.com/

Hi Cameron,

cloud means Oracle DB 11g hosting :) Just kidding, but I believe we will see that in the future.

Cloud means a delivery platform for IT services. There are two types: platform providers (PaaS) and application providers (SaaS).


Peace!

From my point of view, there's a *huge* difference between what ISPs have been doing for years and what Amazon is doing with EC2.

For once, I actually agree with Cameron. Or probably Cloud means something bad for Coherence;) You like the little ;) Cameron ;)?

But what are we really dealing with here?

What is the difference between cloud and hosting? I guess scale-out is what is being pimped right now?

And I completely disagree with the "future of enterprise" line. The enterprise already has fixed infrastructure labor costs that dwarf hardware costs. So to me, it seems just the opposite. A small shop like ours would roll something out, and if it goes big then we don't have to worry about scaling out.

In any case, it sounds more of the Larry Ellison and Scott McNealy chants of "network computing" of the mid-late 90s that failed.

Why is it "beyond hosting"?

Finally a good discussion about this topic. I am pondering for a while if/how my unit in the company I work for could benefit from the cloud and see the same problem. Completely innovative models, greenfield stuff, yes. My company still runs stuff on VMS. Not that we like it, several times they attempted to replace it, the business case did not work out. And not because of the hardware costs. It looks like what some "cloud vendors" are trying to address with "private clouds" is essentially an outsourcing of the infrastructure (in-house PaaS).
I know a case where they bought an "on-demand" version of a software and then because of security and integration reasons they decided to host it in-house. So much of the savings by using an on-demand model...

In a once hot but now seems long forgotten book "Innovator's Dilemma" by Christensen, the term "disruptive innovation" was coined:


"A disruptive technology or disruptive innovation is an innovation that improves a product or service in ways that the market does not expect, typically by being lower priced or designed for a different set of consumers" -- Wikipedia - Disruptive technology


That is, a key factor to categorize new technologies / trends is PRICE (other factors include convenience, smaller size, lower power consumption ...). Moving to cloud computing paradigm will most likely save cost vs. traditional approaches. Lots web startups leverage EC2 / S3 because of lower costs + convenience vs. in-house / self-run hosting... That is, cloud is likely a disruptive innovation.


But when it comes to SOA, XML, COBRA, WS*-, etc., we hardly find any assertions on definitive cost-saving or lower-price. Instead, we hear buzzwords, visions, stories, wishes, expectations ... or unproven statements like XYZ will "dramatically" save (development/operation/maintenance) cost (if you invest a ton now!) Eventually you can draw the conclusion that those stuff are NOT at all disruptive.


Simply put, whatever the buzz, if it can save you $ over alternatives (which can be easily figured out), it will more likely to thrive in the long run.

I agree with the cost factor (be it monetary or otherwise) to judge the "disruptiveness" of an innovation.
I just don't see the comparison between SOA and Cloud Computing in such a direct way (except the hype). After listening a lot and experiencing a lot, I tend to see SOA in the enterprise as an enabler to "refactor" the legacy IT applications and to provide a blueprint for building up the new ones. The idea of assigning Domains responsible for certain data and processes with clearly defined services at the boundaries is important in big organisations and having some methodology to deal with it is a good thing. Naturally the hype and vendor hype was too big.