CIO, Cloud, Data Center

The Cloud Is A Journey, Not a Destination

You’ve all heard the mantra that enterprise IT workloads are marching slowly (or quickly, in some cases) but surely to the cloud.  In this context, the assumption is that workloads are exiting the local data center and making their way to public cloud providers such as Amazon and Microsoft.  While this is certainly true for some applications, it’s absolutely false for others.  That said, even for those applications that may not be easily migrated to the cloud, organizations are still looking for similar benefits.

With that in mind, and before we get too much deeper, let’s consider a few items. First of all, consider why organizations are migrating their workloads to public cloud providers.  For most, the decision ultimately boils down to economics.  A public cloud provider is able to amortize their fixed costs across dozens, hundreds, or thousands of clients.  They are able to enjoy significant economies of scale that individual organizations may not be able to achieve.  Additionally, in most cases, deploying a new application in the cloud requires very little or no initial capital expenditure.  If you think about how we generally buy data center infrastructure, this desire begins to make a bit more sense.  Rather than tying up a bunch of capital for which a return may be months or years away – or is really nebulous – why not move to a solely OpEx-based computing model in which you basically rent someone else’s infrastructure?  And then, as you need more computing power, you simply request it from the provider.

But the move to the cloud goes well beyond simple direct economics.  In many cases, there are potential operational benefits to be had.  For example, standing up a new workload becomes a really easy task. Deploying a new virtual machine in Azure is very simple. In fact, it’s so simple that your end users can – are probably are – doing it.  That’s not a statement intended to denigrate the intelligence of end users.  Rather, the intent of that statement is to demonstrate that those with deep technical skills in an organization are not necessarily the technology gatekeepers, as was the case in the past.  For these kinds of super-simple processes to work, however, cloud providers – which include public cloud providers and even SaaS-based tools, of which there are thousands – have spent considerable time automating backend processes and orchestrating the activities that take place when a user makes a request.

Boiling It Down

So, quickly summing up the expected outcomes from cloud adoption reveals that there are four key computing characteristics that organizations are seeking:

  • Economies of scale
  • OpEx-based economics
  • Easy and granular scale
  • Operational efficiency

My view is that organizations really don’t care where their workloads operate – whether that’s in an on-premises environment or in the cloud – but they do care about the characteristics of the environment that is supporting their workloads.  Please don’t misinterpret what I’m saying here.  Obviously, there are cases in which workload location – particularly geography – does come into consideration.  My argument here revolves solely around public cloud vs. on-premises.

Let’s consider each of these characteristics in a both more depth to try to understand what’s really happening.

Economies of Scale

Non-IT companies really don’t want to hire hordes of technical IT specialists to manage key business systems.  People are expensive and can be difficult to manage.  However, with traditional deployment models, a lot of companies find themselves hiring network administrators, storage admins, end user specialists, virtualization administrators, database administrators, technical business analysts, and a whole lot more.  And then, for redundancy, there is cross-training or hiring additional people.

Since it’s the business they’re in, cloud providers are well-suited to staffing many of the infrastructure-related functions identified above.  And, as stated before, because they can amortize these costs over their full customer base, they can provide such services to individual customers at a lower rate.  Best of all, the customer doesn’t necessarily need to have full-time people in those local roles.

Of course, not all of these jobs can simply transfer to a cloud provider. Even if you’re entire data center resides in AWS, you’re still going to need local network admins to maintain connectivity, and your virtualization administrators will probably become your cloud admins.  As for the others, you may be able to redirect those staffing lines to more business-facing technology roles, or potentially eliminate them.  Of course, most companies aren’t operating 100% in the cloud, so simply eliminating critical roles in not an option.

Keeping it Private

You may wonder just how you get to any semblance of economies of scale when you want to keep your data center local.  This desire to achieve economies of scale may be a key driver for cloud adoption, but often because companies are finding other options potentially lacking.  Having a lot of customers over which to spread the cost is just one reason that cloud providers enjoy economies of scale.  The other characteristics of cloud environments that enable economies of scale are automation and orchestration.  While it can take some time to get these kinds of services in place, once they’re there, fewer people can manage the environment.  I used to work for someone that referred to this as “making the routine things routine.”  It’s a mantra that has stuck with me and is, I believe, more important than ever today.

It’s not likely that you’ll see the same staffing advantages as a cloud provider, but you can get some of the way there by adopting some of the characteristics of cloud, including automation and orchestration.

OpEx-Based Economics

With someone else handling the infrastructure, you’re basically renting it in exchange for a monthly fee.  No upfront costs means no CapEx spend and everything becomes an OpEx play.  For CFOs, this can be a dream come true since ROI calculations become far easier; you just pay for what you use and nothing more.

Here’s why this has become a key driver for cloud: as budgets constrict, CFOs want to wring as much value as humanly possible out of everything they buy and they don’t want to wait years for that ROI.  Consider the data center hardware lifecycle.  Most of your equipment is probably on a three- to five-year cycle and, at the beginning of a new cycle, you procure equipment that will get you through that replacement cycle.  You buy what you need for today and enough overhead to grow into over a period of years.

You end up with a significant amount of infrastructure for which you don’t have any ROI for a long time and you might never see an ROI on some of it if you don’t grow into it.

Keeping it Private

If you try to keep your infrastructure in-house, there are some ways that you can reduce CapEx costs and even shift some of the expense to an OpEx budget line.  For example, you could place all of your infrastructure under an operational lease, the payments for which would come from your OpEx budget.  Or, if you’re OK with a CapEx spend but want to try to minimize the ROI time horizon, consider one of the emerging data center infrastructures – such as hyperconverged infrastructure – that provides a pay-as-you-grow operation.  Buy what you need for today and then grow as your business needs change.

Easy and Granular Scale

This one is easy and is associated with the previous point.  With cloud, as you need to expand, you just do it and pay the delta cost between what you have now and what you’re adding.  There is no need to wait weeks for new expansions to arrive and be integrated into what you already have.  Just as importantly, there is little to no effort involved when you need to scale.

Keeping it Private

It’s unlikely that you’ll find an on-premises infrastructure that is as easy to grow as the public cloud.  You’ll always need to add hardware of some kind when you need to grow, but that doesn’t mean that the expansion process needs to be difficult.  With a lot of the hardware options that are available on the market today, you can scale the environment with ease and without downtime.  Moreover, you’re generally able to grow in incremental units that make sense for your business, but the exact step size is dependent on the vendor you choose.

Operational Efficiency

Many cloud services are really easy to use.  End users can quickly and easily deploy new services.  Here, though, we’re focused on services such as AWS and Azure.  These services, for IT pros, are really easy to work with, especially considered that they’re no hardware and virtualization deployment to deal with.

As has been mentioned previously, a lot of the operational efficiency has come about because cloud providers need all of the economies of scale they can get to maintain their pricing level and to bring them down over time as competition heats up.

Keeping it Private

If you want to keep your data center local, you need to find ways to improve your overall operational efficiency. In fact, the term operational efficiency is becoming one of the most used terms in marketing.  Companies across the enterprise IT vendor spectrum are touting wares that massively simplify the data center paradigm.  These solutions are a combination of hardware and software solutions that enable easy growth and that are easy to manage.

In addition, there are automation and orchestration tools available that enable you to bring cloud-like management capabilities to your private data center. The typical term for the kind of environment you need to build is private cloud.  Many people equate high levels of virtualization with having deployed a private cloud.  This is an inaccurate statement.  In such cases, you’ve simply virtualized a lot.  Until you’ve deployed tools that enable self-service and that automate and orchestrate processes, you’re not a private cloud.

It’s Characteristics That Count, Not Destination

The whole point of this post is to help you understand that cloud is really a journey, not a destination.  If you can adopt the characteristics that make cloud so compelling, it may be possible to retain infrastructure locally.  For many, moving out of an on-premises environment isn’t really feasible, but they’re still looking for ways to maximize their IT investment, shorten the ROI for their IT acquisitions, and help end users better meet business needs.  With the right local environment – hardware and software – you can become cloud-like in your operations.  While you may not get to the same depth as a cloud provider in all areas, you may be able to get “close enough” to meet the needs of your needs.