Cloud computing is entering its teen years


The cloud has entered the “awkward teenage” phase, as James Staten at Forrester Research recently noted. Luckily, it’s not overrun with rowdy adolescents the way other technology paradigms have been (think early social media). It’s been maturing, and so have best practices for its use and adoption, along with the related security and application aspects.

In fact, it could be argued that IT management and security have kept a tight rein on all cloud activities. Many of us did what we could to help it grow while making sure we could prevent serious damage in the process. First as a research analyst, and now as the vice president of strategic solutions at the software vendor CA Technologies, I have written multiple research reports and white papers on responsible ways to use a cloud (and when not to).

Cloud computing is growing up, and it’s time for us in IT management to loosen our grip. It won’t be an easy transition, and each environment requires its own solution. Here are some top-line suggestions to consider.

  • Allow for discovery: Many organizations are still reluctant to take full advantage of the potential of cloud computing and deploy pilot projects. And every time there are “rogue” activities, it sets off alarm bells. Let’s be realistic — no system is foolproof or completely immune to problems. There’s no question that a complete shutdown is sometimes justified, but it’s also vital to have adequate remedial measures in place ahead of time. Cloud installations need to offer strategic visibility, thorough transparency and predictive intelligence. That’s the best way to implement truly transformative cloud initiatives that offer the greatest benefits.
  • Provide modeling and simulation capabilities: Caution is still justified — it’s hard to make a full-on investment when there’s a lot at stake and the shift to the cloud is still relatively young. Rather than throw the cloud into real-world applications immediately, it’s best to set up test runs. Fortunately, companies considering a move to the cloud can simulate a service and determine performance, capacity requirements and cost structure before making the commitment. Similarly, parallel development processes make it possible to test those services before they’re completed.
  • Help develop credible capabilities: Implementing applications that are perfectly appropriate in other environments, but aren’t a good fit for the cloud, is a ticket to losing internal support. To avoid such mishaps, it’s best to consider turnkey cloud computing platforms that enable corporations to create a truly analytical model of the services they need, and then transform the application into an on-demand business service. IT management needs to offer insights into current and future capital and operational costs by using predictive analytics, managing service levels and costs, and providing visibility regarding who is using each service, how often they are using them, and how those services are performing.
  • Provide “grown-up” financial management: Wishful thinking aside, it isn’t always easy to map financial benefits to IT investments. In the evolving world of cloud computing, this can be a major obstacle. That said, there are lifecycle and asset management capabilities available from companies such as CA Technologies (where I work), Nimsoft and other IT management vendors. These account for the physical, financial and contractual aspects of hardware and software assets on-site, virtually and in the cloud. With a software asset manager, companies can get an integrated view of the infrastructure base, and analytic tools can help identify real asset costs and avoid unnecessary capital expenditures and lease penalties.
  • Provide automation and assurance: Skip the guessing game. Current solutions from CA Technologies, Nimsoft, HP and VMware for cloud service assurance automatically link applications and transactions with the underlying infrastructure. Whether they reside on- or off-site, these solutions deliver real-time visibility into the performance of critical cloud services. Even in the most dynamic and heterogeneous environments, available technologies help corporations keep pace with rapid change and confidently adapt new computing models. This is the best way to ensure the key business services run smoothly and to optimize the business user’s support experience.

I’m not suggesting that any of this is a magic bullet. And again, as with any major technology shift — mainframe, distributed, desktop — there will be singular problems and unique solutions. The undisputable reality is that cloud computing is growing up, running mission-critical workloads and delivering on its potential. It’s the responsibility of those of us with a history in different (and evolving) IT disciplines to provide the right mix of guidance, management, freedom and supervision.

Andi Mann is vice president of strategic solutions at CA Technologies. Mann has more than 20 years of experience in enterprise software on cloud, mainframe, midrange, server and desktop systems. He co-wrote the handbook, “Visible Ops – Private Cloud.” He blogs at Andi Mann – Übergeek and tweets as @AndiMann.

Image courtesy of Flickr user Erin Purcell.


Dan Farfan

“consider turnkey cloud computing platforms that enable corporations to create a [thorough, robust] analytical model of the services they need, and then transform the [model] into an on-demand business service.”

Excellent! Assuming my few word changes are still what you meant. :-) Most of my career as a computer scientist has been in telecom, so the metaphor I use for the “promise of cloud” is that of the OSI 7 layer model from the 70’s – the bible of sorts for telecom devices of all types to interoperate. Trouble is, at layer 7 (“Application”) it becomes the old cartoon of the Einstein-like character standing in front of a board full of equations that end in a final bubble “Miracle Happens Here.” The telecom folks literally threw up their hands and said, “We don’t know jack about what ya’ll do, but Software dudes, go for it. Right here. Lucky number 7. Groove on.”

I predict, the cloud vendor who first best re-imagines, implements and offers a computing services layered model will dominate this space with almost the ubiquity the OSI 7 layer model did (Heard anyone challenge it since the 80’s? Besides my 7+2 layer improvement, that is :-)

IMO, none of the “XYZ as a Service” folks are even in the right forest much less near the right tree to be barking up. That’s why the field is still so wide open for any cloud vendor to make what will be known as the next leap in computer science, easily as profound and transformational as the operating system itself.

So far, as I see it (and my startup does use a cloud vendor, so I’ve studied most of them) the cartoon is still “Miracle Happens Here” but with mostly a blank chalk board waiting for a bunch of programmers to fill in the magic “..Software dudes, go for it. Right here. Lucky number 7. Groove on.”

(psst.. if a 40 year old call to action is still your call to action, especially while you’re pretending to be new and shiny, there’s a good chance there’s opportunity to step back, rethink, create a new and blow everyone out of the water.)


Andi Mann

Hey Dan, great comment! I think you are really onto something here.

Certainly agree that the focus still is far too deep into the stack, and there is not really a comprehensive description of the entire service, top to bottom.

So I think there is a real opportunity to figure out a more complete service model – along the lines of the OSI model, if you will – from the hardware on up to the UI, with everything in between.

I know CA is doing something a lot like this – our CA AppLogic solution is all about designing and assembling a repeatable service model, so you can deploy it anywhere, anytime, without having to redesign it. Not exactly the full service layer model you are describing, but given what we are doing I can certainly see the value in your idea.

Thanks for posting – great food for thought :)


Great discussion here. There is definite value in Dan’s ideas for creating a analytic model for services and transform it into an on demand service.

The software industry with SOA and Cloud have made significant progress towards standardization of the end to end stack and made development of a XML descriptive model possible. However, vendor focus to differentiate for revenue and customer lock in takes us away from the goal each time we think we have come close.

However, its a hard reality that we have to live with multiple architecture models at the app level for a while. But with cloud based automation and ubiquity of deployment, this becomes manageable to a certain extent.

I have worked on middleware platforms at Commerce One (B2B) and SAP that are model driven. Extending that work at Gravitant, we have created a cloud broker platform that is driven by an model that combines predictive analytics with app and cloud Infra architecture models to broker(design, source, provision, govern) services between providers and consumers. This approach will enable IT management to support multiple app architectures that can be deployed anywhere from one platform and control is for cost, resources and SLA’s.

Its not exactly a full service layer model but a significant step in that direction.

Thanks Andi and Dan for the post and discussion.

Dan Farfan

My pleasure :-) I admit I haven’t kept up with the progress AppLogic has made. I recall liking what I saw (although I’m not sure when I first read about it).
CA is certainly one of the plausible homes for my work.



It’s a case of doing a cost benefit analysis isn’t it? As Andi says cloud needs to be managed to make it what it needs to be. This involves understanding all the benefits – both actual and potential – as well as the possible problems

Andi Mann

Absolutely, great point. This is something a lot of zealots on both sides seem to miss. Yes, there is risk, but so there is in a lot of business decisions. You just need to balance the risk against the return. This is not new, and shoul dnot be scary. Indeed, it should be SOP.

Thanks for the comment.


Andi Mann

Interesting graphic, thanks for the link, Mitesh.

I wonder at the difference between ‘perceived risk’ per your link, and ‘actual risk’.

Or maybe it does not matter – maybe perception *is* reality.

Thanks for contributing.

Steve Caniano

As a father of teens I finally get why most cloud vendors I know have all the answers, expect to change the world, and hope to someday make a buck! ;-)

Comments are closed.