The Top 5 Things ITSMers Get Wrong about Cloud Migration

Migration to Cloud

As an IT service management (ITSM) professional, which public cloud “camp” are you in: Cloud is great? Cloud is a disaster? Couldn’t care less about cloud? (Hopefully none of you are still at “what’s cloud?”) These and many other cloud “attitudes” tend to be present, all at the same time, in a typical enterprise. And, worryingly for ITSM professionals, it can lead to doing the wrong thing such as lift-and-shift cloud migrations.

A lift-and-shift migration is the seemingly simple and innocuous process of moving a workload from on-premise to the public cloud – changing little and with the same operational model. Lift-and-shift is perceived as a low-risk approach to getting into the cloud because of the assumption that “if you don’t change anything, then surely it’s kind of the same as before the move.” That moving a workload to the cloud is like moving from one server type to another, after all cloud is just running your workload on someone else’s servers, right?!

No, cloud is not “just someone else’s servers.” Public cloud is a different operating and business model to on-premise and it’s why ITSM professionals need to be involved. But before they do, they need to be able to avoid the five common misconceptions about migrating to the public cloud.

1. Lift-and-Shift to Cloud is a Good Idea

This is an appealing idea among people who are risk-averse and/or are not familiar with cloud. The people who want change but don’t want to change.

Non-cloud, on-premise workloads can be flaky and have been described as being like fragile snowflakes and laden with technical debt. It’s not unheard of for servers to never be patched or rebooted because of the fear that they will never restart, or will require hundreds of patches to restart. Would you lift-and-shift these snowflakes into the public cloud?

From an ITSM professional’s point of view, lift-and-shift is even more complex beyond the technology. An ITSM professional also needs to understand the operational model of cloud, characterized by the “shared responsibility” models of cloud service providers and having to adjust your operational model to theirs.

If you have snowflakes and no cloud-operations understanding, then lift-and-shift is not a good idea.

2. Lift-and-Shift to Cloud is Easy

However, if you still think lift-and-shift is a good idea for your company, then your next thoughts need to turn into actions and “how do we lift-and-shift?” With an element of cloud naivety (hopefully this term is polite enough), someone might think:

“VMs and data? They’re just files, right?! I can zip ‘em up and copy them to the cloud!
Hit the launch button, and we’re off!”

Just like pointing a rocket at the moon with your one good eye, lighting the touch paper, and then putting your fingers in your ear. What on earth could go wrong?

Sadly, the two things that will put you “off-trajectory” are cloud architecture and operations:

  1. Cloud architecture is not like non-cloud, on-premise. And the network and security, are different.
  2. Cloud operations is now shared responsibility and this means that you have to “bend in the cloud wind” to work with the cloud service provider’s operational model.

The cloud service provider is now doing some of the stuff that you used to do, so now you have to do things their way in terms of architecture and operations.

3. Cloud is Just Rent-a-VM

With cloud naivety, someone might think: 

“Cloud migrations are just about virtual infrastructure, because cloud is just renting a VM on someone else’s server.”

But this is “cloudy thinking,” not cloud thinking, because the cloud has many more services beyond just renting a VM.

If you’re an ITSM professional who works with enterprise IT and the infrastructure team, you may have the perspective that cloud is all about VMs. As an example of this perception bias, a lot of total cost of ownership calculations (TCOs) only compare the cost of an on-premise server to a cloud instance. This is a woefully inadequate comparison, as the cloud instance is a fully loaded cost including everything the cloud service provider does for you, and not just a VM hardware/software cost.

An ITSM pro needs to understand the full range of cloud services beyond virtual infrastructure. With the leading cloud service providers having more than fifty cloud services covering directories, databases as a service, and new paradigms such as serverless computing.

4. Pick Any Cloud: They’re All Similar Utilities

There was once industry talk of cloud service providers being like electricity utility companies and it would be as simple to consume cloud as putting a plug in the wall socket. But anyone who’s tried to understand the complexity of utility provider bills and charges knows that it’s anything but simple.

Similarly, cloud service provider billing models are different and hard to compare: per-minute versus per-hour; instances are different sizes; pricing is different; cloud services have different names and features; and all services are accessed through incompatible APIs and consoles.

ITSM professionals need to fully understand one specific cloud service first and then work to align operational models to get billing, security, service delivery, and operational differences. To help, leading cloud service providers, such as AWS, have guides for ITSM professionals such as the Cloud Adoption Framework.

5. Multi-Cloud Saves Us from Cloud Lock-In

The statement: “We don’t want to be beholden to one cloud provider and be locked-in” isn’t new or specific to cloud. Multi-vendor strategies have been around for years, for those companies that can afford them.

Small enterprises don’t have the luxury of being able to afford the cost and complexity of using multiple cloud service providers at the same time. Many large enterprises can, but still shouldn’t. Beauty pageants of multiple vendors competing for different parts of a corporate IT landscape are often exercises in vanity not sanity. Spending months and thousands of dollars on meetings and decisions to save a fraction of that on list price discounts is “penny wise, dollar foolish.”

Many successful businesses have “gone all-in” with one cloud service provider because they can focus their scarce resources on getting the best out of that provider. Some use two cloud providers to hedge their bets for specific features like disaster recovery, making sure that all their IT eggs aren’t in one basket. It’s fine, and it’s not lock in.

Companies that want a single pain (sic) of glass across multiple-clouds to avoid lock-in have more expenses (cloud broker software, cloud consultants, etc.) and they are unable to access new advanced features only available in one cloud e.g. AWS Lambda or Azure Functions.

A cloud lift-and-shift migration might initially feel like the “best” approach to cloud from an ITSM point of view. That nothing much needs to change, and it’s like flicking a switch to turn the cloud on. However, the reality is that much is different in the cloud. And if you don’t invest the time to understand the new cloud operational model, your lift-and-shift might initially be more crash-and-burn. So please try to learn from the mistakes of others rather than learning from your own.

  • Torsten

    Thanks for sharing this. I fully agree with your points. Before starting your cloud adventure, you should consider a few things, like understanding pricing and licensing models.

    I have outlined a few points on my blog (in German only):

    • Thanks, Torsten. I’m sure my German friends will take a look at your points :)