No Private Cloud Plans? Well, You'll End Up in the Public Cloud...
Enterprise IT undergoes an evolutionary phase every 20 years or so. IT research firm IDC labels these evolutionary phases as platforms. IDC identifies the first platform as the mainframe, the second as client/server and the third as the cloud. While every evolution diminishes the importance and reliance on its predecessor, no platform has yet to completely replace another. To this day, IBM makes significant money with its AS/400 mainframe.
Today, there is a lot of talk about cloud and cloud services. Even then, many in enterprise IT have yet to jump on board, trying to ignore the hype and thereby downplaying the importance of the cloud and how it will affect their businesses into the future. I believe those companies are still transitioning from the first to the second platforms, and have yet to even consider the third.
Yes, there's still growth in the second platform. Many IT research firms tout 30 percent growth year over year with traditional IT spending. What enterprise IT decision-makers fail to notice or ignore is that the third platform -- cloud services -- is growing by about 300 percent year over year. Most of them see these numbers as hype, but it will change the face of IT suddenly and without notice.
Enterprise IT does not realize how many cloud services they already are consuming today, whether it is ERP and CRM systems or even applications that were just until yesterday a pillar of the data center. In the enterprise there is significant adoption of Office 365, or EMC Atmos and many others. While traditional IT spend grows at 30 yearly as enterprises go about their business building system the traditional way, what will happen in the coming months and years is that this growth will screech to a halt. I see 30 to 15 percent and then even quicker than that as the end approaches. So, what will be the reason then?
For one answer, let me share with you my thoughts based on endless customer interactions. Today when I talk to customers, I silo them into two categories. The first goes those who are content with the existing environment, but want to make it better, virtualize more and make it a lot more streamlined and efficient. The environments are typically static -- not a lot of changes occur, not a lot of VM requests, no need for self-service portal or elasticity, and so on. The environment is very predictable.
In the second customer silo, the environment is very dynamic, Those customers have adopted converged infrastructure but are starting to complain about converged infrastructure sprawl and their ability to manage it all more efficiently. The latter silo is a perfect candidate for a private cloud deployment and are either looking at a serious deployment or are already deploying it.
The first silo of customers that are striving to be 100 percent virtualized don't realize it yet, but they are headed to the public cloud. Right now, they're in no hurry to do this because either their hardware refreshes are not imminent yet, or culturally they are not ready yet. But without a doubt, this silo of customers will at some point leverage an IaaS offering to host their already 100 percent virtualized environment. Some of you might be thinking, "Elias, we have large amounts of data and storage on premises that would cost a lot to move to the cloud," or you may bring up some security or privacy concerns. Let's tackle both of these obstacles.
Storage capacity is a valid concern, but technology will overcome this easily in a number of ways. We may see technology that is able to bring up a working set of data and keep it as close as possible to the infrastructure, while replicating back to the on-premises storage. It's one way of not moving everything to the cloud.
The other option is that cloud storage will get to a point where it really rivals on-premises storage on cost. Today, the argument can be made that cloud storage in the long run is not cheaper than on-premises storage and that may be true, but I think that will not be the case soon. I also think that network and communications link are continuously improving and therefore moving these large sets of data will not be as challenging as most think.
As far as privacy and security are concerned, I have my own views that privacy is dead and I have written about it here on Virtualization Review many times. I have also discussed security and the fact that I believe that cloud service providers have better security than enterprise IT can ever have (also something I've discussed here before).
That being said, from a compliance, security, privacy perspective, we are starting to see the rise of the "verticalized" clouds that are intended to address healthcare requirements, financial requirements, educational requirements, and farther down the line.
For all these reasons, I believe that enterprise IT that is striving for 100 percent virtualization will soon find itself asking the question or find itself being asked by the business or a new savvy CIO: If we're 100 percent virtualized, why are we still spending CapEx money on IT infrastructure? Does this mean the on-premises or collocated data center is dead? No, it means its footprint will be significantly reduced and just like today you virtualize first. So, tomorrow you will put the workload in the public cloud first and you will have to justify moving it on-premises where the cost of hosting will be higher. Eventually, many IT organizations won't own a data center, and many IT departments who were once known as builders for infrastructure will switch it up and brokers of cloud services.
Posted by Elias Khnaser on 02/24/2014 at 4:59 PM