Monday, August 24, 2020

Hidden cloud migration gotchas — and how to avoid them

From a lack of visibility into infrastructure to licensing and governance issues, frustrating surprises await your migration to the cloud. Here’s a real-world look at how traditional companies are navigating the cloud’s less-documented challenges.

Building a technology stack from scratch in the cloud can be a dream -- if you’re a startup. But if you’re an established company steeped in on-premises solutions, shifting to the cloud can mean unexpected hurdles and headaches beyond belief.

Companies looking to make good on the benefits of moving to the cloud need to go in with eyes wide open. To be sure, the much discussed benefits of the cloud -- in terms of time to market, cost savings, the ability to scale resources as needed, and so on are real. But enterprise cloud adopters say they have been surprised by some of the lesser-documented challenges migrating to the cloud brings, such as the difficulty in changing traditional mindsets, the lack of visibility into the new infrastructure, the cost of data transfers, governance issues and how licensing agreements need to be revised or new ones negotiated.

Here, we take a look at hard-earned lessons from real-world cloud migrations and how traditional companies are navigating the hidden hurdles involved in moving to the cloud rightscale careers.

Making the mental shift

Financial and actuarial models are the lifeblood of Pacific Life. Running those calculations requires a lot of high performance computing, but only for a few hours a week, making the cloud an ideal solution for the Fortune 500 insurance company.

So about four years ago, rather than purchase additional hardware, Pacific Life’s IT staff began renting compute capacity and moved the workloads to Amazon Web Services.

“A lot was driven by the fact that we have a very expensive data center to operate here in Newport Beach [California], which as you can imagine, is not the best place to run one,’’ says Reza Salari, manager of cloud security and service transformation at Pacific Life. The insurance company also has a data center in Nebraska that it uses for disaster recovery.

Once the company saw the benefits of being able to scale up and down in the cloud it generated increased internal interest. More recently, Pacific Life began hosting its first application on AWS, data visualization software from Tableau, so that employees can generate custom reports from anywhere, globally. Currently, Pacific Life is working on migrating 28 websites to AWS. But as the company moves deeper into the cloud, along the way there have been some surprises.

Tableau required “a big mindset shift to conceptualize how the pieces come together,’’ says Salari. “First was conceptualizing the architecture and rethinking our tools; we couldn’t just move our on-premises firewall and put it there.”

Both the hosting of Tableau and the migration of the websites is requiring “a lot of security rethinking,’’ as they bring in a new ecosystem of vendors and products Pacific Life hadn’t worked with before, adds Jason Vigil, a solutions architecture consultant at Pacific Life.

“We always assumed that moving to cloud was a simple matter of, ‘OK, we’ll teach our engineers to do Amazon and poof, we’ll have cloud engineering,’’’ says Salari. “We didn’t really understand the depth of the migration and the people organization challenges. We thought we’d have Amazon come in and do a couple of classes and we’d be a cloud organization. That was a huge shift change.”

Dealing with endless capacity

Today, some four years after Pacific Life launched its cloud journey, there are only a handful of people who have caught onto the “cloud fire,” says Salari, and there are those within the two 280-person IT organizations who perceive the cloud as hosted infrastructure that requires a simple “lift and shift” to move something off premises.

But it is far more complicated than that, experts say, stressing that organizations should prepare for cloud migrations well before starting the actual transition, especially when it comes to expected consumption and usage.

“You have to plan for capacity differently than when it was in your data center,” says Mindy Cancila, a research vice president at Gartner. “Most organizations overbought their capacity [when first moving to the cloud] because what you didn’t want to have happen is not ask for enough, and then have to go back and ask for more infrastructure and not have it approved.” That approach ends up being quite a bit more expensive, she adds. “So it’s a paradigm shift for organizations to think about their consumption and usage, which is very different.”

Yet, Cancila finds that most of her conversations with clients are often reactionary and she councils them on the need to put better controls around governance, cost, and usage. “I often will share with clients that it’s not that anyone’s doing anything wrong, it’s the inherent nature of cloud: on demand, self-service, pay as you go; so organizations have unique challenges around putting a governance structure around that endless capacity.”

Evaluating and rearchitecting workloads should be an active part of a migration plan, notes Deepak Mohan, a research director at IDC. “If you don’t change, then what happens is you use the same resources and all you’ve done is shift from a PC I’ve paid off versus one I pay for on a monthly basis.”

No comments:

Post a Comment

Server management systems

Enterprises receive the services and functions they need (databases, e-mail, website hosting, work applications, etc.) for their corporate I...