The future of on-prem and the cloud

Any organization that shifts from an on-premises IT infrastructure to public cloud infrastructure-as-a-service (IaaS) will spend some time operating within a hybrid model. There is no magic switch you can flip to instantly migrate everything from your data centers to the cloud. But how long can (or should) they coexist?

If you’re planning on using a hybrid infrastructure for an extended period of time, my advice is this: don’t. While it’s impossible to avoid a hybrid setup during a transition period, most organizations are best served by committing to cloud IaaS completely (or as much as possible) and following a plan that can get you there incrementally over the course of two to three years.

Why businesses are moving to the cloud

One of the main drivers for undertaking a cloud IaaS migration is the existing talent pool, for a couple of reasons. First, as legacy on-premises hardware and networks continue to age, the pool of those with the expertise to properly maintain those devices and systems shrinks. On this level, it’s not uncommon for professionals to either retire or change careers, and there certainly isn’t much younger talent with experience working with older IBM or Sun Microsystems hardware, for example. Longstanding knowledge within an organization is very valuable, making it both expensive to replace and costly to lose.  

Similarly, with “younger” talent coming out of schools with more of a cloud focus, cloud IaaS is where an organization wants to be if it wants to attract and retain newer employees. The goal is to develop, grow, and (hopefully) retain talent, and that is becoming harder and harder to do if a company only offers on-premises infrastructure and related tools.

There are varying skill sets in play when looking at on-premises versus cloud infrastructure. For example, the toolsets used to manage and support on-premises hardware and network devices are typically different from those used in a cloud environment. This includes differences in monitoring, performance management, and implementation support. And I’m not talking just about differences in terminology around how these tools work; cloud IaaS management and security tools are typically quite different in function and use compared to on-premises tools.

Top 3 drawbacks of a long-term hybrid model

Operating under a hybrid setup forever is possible in theory, but unless you have an unlimited budget doing so over an extended period doesn’t make business sense. Here’s why:

• It requires additional administrative support. Under a hybrid model, you need professional systems administrators supporting on-premises and in the cloud. These teams handle things like patches, monitoring, failover, backup, and restores. This is more than just extra work; it involves extra knowledge – and probably extra sets of tools.
• Hard costs reach a tipping point. At some point, the physical footprint supporting your on-premises architecture — which was probably built years ago when it made economic sense — stops delivering the needed ROI. Imagine a couple still living in a big house after the kids move away. The house may be nice, but it’s not very efficient. You’re essentially paying for space you don’t need or use. Your needs have changed, but you still pay for the entire house. Eventually, overhead costs are spread out over a smaller base, so unit costs go up.
• Different policies. If an organization has both a cloud team and a non-cloud team, it has essentially told its people they’re either on the varsity or junior varsity team. I have spoken with too many clients who have unknowingly created that issue, which can lead to resentment on the team. If the company says, “we are cloud IaaS first,” expect everyone to want to be on the cloud team. If the organization is saying the future is in the cloud, but they want IT staff to keep managing on-premises, what does that staff do in three years when their peers are working in the public cloud? How do they get retooled and retrained? Organizations need to be aware of the problems this type of situation could cause over time.

Like most things, it comes down to people

Businesses should be thinking about what it takes to support their on-premises and cloud IaaS based on the talent that’s available. If you reach a certain level based on size and scale and need 24/7 coverage — essentially all businesses in this era require that level of support — how many engineers do you need to cover all your various systems for 365 days a year? 

There are certainly organizations that have recruited a team of smart people who stuck around over time, but this team eventually realizes it’s hard to also have a life outside of work. It’s not just the sheer hours but also the constant stress of waiting for that phone call. Planning activities is always an issue because you know you’re accountable; if something does happen, it could cost you a weekend. This realization ultimately affects innovation — you can’t expect people to be on call each week and then also implement that next software/hardware solution that drives the company forward.

In the end, a hybrid model is inevitable during a transition to cloud IaaS, so the idea is to make that transition as efficient and cost-effective as possible. With that in mind, here are three steps to get the ball rolling:

1. Bring all stakeholders into the discussion. Technical leaders should team up with CFOs and other business leaders to map out and explain why each step of a cloud migration makes business sense.
2. Perform a full TCO analysis. Analyzing the costs involved in a cloud migration requires much more diligence than only using the online calculators provided by the major cloud providers.
3. Build a three-year roadmap. Create a plan to migrate to the cloud incrementally based on business priorities, and make sure that plan continues to move forward.

Most would agree that migrating customer-facing and internal-facing systems are top of mind; that’s where you get the “big bang” and where ROI is generally made. Where it’s not made is in file systems or network devices. That said, if you run out of money or time before you complete your migration of these aging systems, you’ll be stuck with a sub-optimal solution and business could suffer.

Ultimately, there is no cookie-cutter solution that works for every business, and these are just some of the issues you need to consider. How you get to the public cloud may look different from a peer or competitor, but the fact remains that limiting the amount of time you spend operating under a hybrid model will almost always give you the best chance at success.

Michael Bathon is Vice President & Executive Advisor, IT at Rimini Street.