Updated: May 31, 2019
This week, as Chordia Consulting’s #CautionaryCanadian, I'll turn my attention to an interesting article from Datamation that summarizes the current state of Cloud technology and offers cloud strategy advice to the reader. I'll offer my own perspective about the article’s conclusions, suggesting some areas where agreement, but also a bit of skepticism may be in order. You can download and read the original article here. Note that Datamation will ask you to register (it's free) to access it.
This week’s article starts by identifying cloud computing as representing very significant opportunities and suggests if you are not already embracing cloud with a strategy to exploit it, you might be missing the boat. It’s hard to argue with this view given how dominant cloud computing has become in our industry. Even so, some of the perspectives and predictions made by the article’s author are suspect. A provocative statement can help to sell a point of view or a company’s services, but it often leaves the reader wanting more, being frustrated with a lack of practical advice as to how to take some important next steps.
I once was consulting to a very large European bank and tried to convince the CIO that he and his company should be exploiting cloud computing. I learned a lesson in the practicalities of running a large bank’s IT department and, being forced to get past the hype, grounded the debate in the attributes of this particular bank. These attributes were unique in many ways compared with other banks, let alone other companies in other industries. Details matter. One of the executives at the bank thanked me for my effort after he learned of my disappointment in not getting to an agreement to embrace cloud. “Things take a long time to change” he said “but your work has started that change.”
Contrary to the opinion expressed in our subject article, data centers are actually enjoying continued investment. New workloads that address digital transformation are driving investments both in the cloud and in traditional data centers. Data centers are home to “Systems of Record” and I do not see these workloads moving to the cloud anytime soon. To do so, there would have to be a good reason to go through the effort, cost, and risk of such a major undertaking. That time and effort is better spent enabling new workloads, some of which are destined for the cloud (for good reasons) while others may not lend themselves to cloud. Once again, the details matter. These new workloads are often what we think of as “Systems of Engagement.” When prototyping new SoE workloads, it makes sense to use cloud services. For integration and other reasons, however, some of these workloads may move back to the data center as they ramp up in production. If the data center infrastructure hosting these workloads is cloud-like, so much the better. This isn’t hybrid cloud: that is something else. But it is smart: putting the proven techniques of standardization, orchestration and automation to good use.
Many pundits (and our subject article’s author) believe you should move to the cloud for cost-reduction purposes. Ironically, I have seen use cases where moving to the cloud is less cost effective. Take, for example, any given workload in an environment and you can define the interfaces and services it requires to run as well as any integration points with other services that determine its end-to-end usefulness. Hosting this same service in the cloud means you need to have the same (or similar) interfaces present in the cloud service offering and you need to be able to interface between the workload’s new environment (the cloud service) and other workloads. There is almost always a gap between the services the cloud provider offers, given their need to be standardized to drive profitability, and the services your workload may require. You need to address these gaps, and, similarly, the integration issues to re-host this workload in the cloud. It turns out that there is not only a cost to move a workload, there is also a cost to support and integrate that workload so that it actually works. These costs are not always well understood and are often overlooked, causing the move to the cloud to be more problematic than expected. The details matter.
Our subject article describes the possible excitement about access to new technologies like AI, IOT and Blockchain in the cloud. For learning and experimenting with these new technologies, using a cloud service sounds like a good way to start. Be forewarned, though, that you will need expert skills and focused energy to learn what you need to learn and to figure out how to exploit these new technologies. This small detail is a large consideration. Without addressing it, the fact that these technologies are in the cloud is of little value to your company.
The points made by this article on the security reversal, agility, and vendor lock-in are all great points. Cloud security used to be an issue but has seen major investment over the past few years to address what were once legitimate concerns. Meanwhile, more and more companies have been in the news regarding their own systems’ security breaches. It was simply a matter of time before this security reversal happened. In terms of reasons for embracing cloud services, I have always felt agility was one of the most important and that agility, more than cost reduction, was a better source of cloud value. Finally, cloud service providers are all trying to create a higher level of service from their infrastructure in order to drive greater value for their clients and, therefore, drive more adoption. As long as they build these services using industry standard protocols, this will, by definition, avoid vendor lock-in. It is important to understand what the cloud providers are doing in this regard. The details matter.