As with most technology trends, cloud has proven to be neither a one-and-done nor a straight shot. Companies are finding that the cloud makes sense for some use cases, but not all. When it does make sense, not all cloud deployments or providers are created equal. The answer to ‘Which is best?’ can change over time. And yes, managing it all is complicated.
This is the fourth in our End User Summit Top-of-Mind Blog Series, which reflects the top issues discussed during the Infrastructure Masons October 2019 End User Summit in San Jose. Attendees included senior-level end user leaders from across the globe, IM Foundation Partner executives, and members of the Diversity & Inclusion (D&I) Committee.
End users at the Summit self-selected into one of four discussion groups: data center, network, hardware, and infrastructure management. In all four, leaders talked about cloud, but in the infrastructure management group in particular, cloud was the focus.
The future is hybrid – on-prem and multicloud
The way of the future is hybrid: a combination of on-prem and cloud. And not a single cloud – that is, a single type of deployment with a single service provider – but rather, a multicloud combination of SaaS, IaaS, and PaaS from a variety of providers. As one end user explained it, “Depending on what the use case is, you would use different parts of the different clouds or even your own enterprise servers.”
Getting from here to there is less straightforward than it might seem. One end user shared her example: “My company said, ‘Okay, we're going to put everything in the cloud because it's cheaper. Wait a minute. We're going to retain some back. We want to take advantage of the cloud, but it's too costly to go to the cloud.’ I don't think we're very mature as an industry about how we create an optimal balance of on-prem and cloud. That has to be addressed from a cost and capability perspective.”
“Depending on what the use case is you would use different parts of the different clouds or even your own enterprise servers.” – Click to tweet
It’s all about the use case
Creating the balance starts with understanding the use case – and bottom line, whether it’s more cost effective to go to the cloud or go on-prem. Sometimes, for example, a company will go to the cloud for burst capacity, or capacity in new regions, because demand isn’t predictable and/or large enough to justify new on-premises capacity. Sometimes a company will stay on-prem when they’ve designed their own environments for optimal efficiency delivering on very particular use cases.
One end user explained his use case for going to the cloud – or not: “We owned the datacenter, the hardware, the network, the software that ran it, and the shared platform teams. People’s immediate reaction was ‘We can go to cloud and it’ll be faster and cheaper.’ I said, ‘If it’s good for the business, amen we should do that.’ But we had to do the data analysis. We spent a lot of time modeling the cloud enough to do an apples-to-apples comparison. What it worked out to be is the majority of our core infrastructure is on-prem. We leverage cloud in regions we’re expanding into but when it gets to a certain size we bring it on-premises.”
Another end user explained her use case for staying on-prem: “We design our own servers to get rid of all the things we don't need. If we go into the cloud, we have to adopt what it is. What you see is what you get. But I don't need all that shit. There's this very specific stuff I want. I'm going to design my own and then I don't have to pay for all that other stuff in the cloud.”
“When you go into the cloud, you're actually abstracting away from the underlying architecture,” explained another end user. “And when you do that, there's a missed opportunity for driving efficiency and using the architecture as it's architected with the applications and the use cases that you're running. And that's why with the emergence of AI and some of these other things you see people building their own specialized infrastructure so that they can drive much more efficiency, better performance, and so on.”
“Balancing on-prem and multicloud starts with understanding the use case – and bottom line, whether it’s more cost effective to go to the cloud or go on-prem.” – Click to tweet
Movement to the cloud will ebb and flow
As use cases and cloud capabilities change, companies will ebb and flow into and out of the cloud. “As technology evolves and the price constraints change, you're always going to see companies evaluating what they want in the cloud and what they don't want in the cloud,” explained one partner. “I don't think there's a solid trend line. It’s going to go up and down.”
“The movement of businesses into the cloud could actually go through some oscillations along the way,” agreed one end user. “The reason could be because of business drivers – and that's just which is cheaper. Or the reason could be efficiency. In some use cases it'll make more sense to just own it all and hyper optimize the whole stack on-prem.”
“As technology evolves and the price constraints change, companies will ebb and flow into and out of the cloud.” – Click to tweet
Managing across hybrid multicloud environments is more complicated
The role of infrastructure management becomes more complicated in a hybrid, multicloud world. It requires orchestration across environments for different use cases.
As one end user explained, “When you go to the cloud typically it's not one cloud. It's multiple SaaS clouds. It's multiple IaaS or PaaS clouds. But we still have to control our destiny by being able to orchestrate across those clouds.” And across use cases. “Not just to provision network, compute, and storage, but also to provision endpoints using orchestrating technologies and capabilities, provision badges, provision bikes or scooters, or building offices in a click.”
“None of the individual cloud vendors are going to do it for you,” the end user added. “Leveraging the continuous integration/continuous delivery and deployment (CI/CD) pipeline to control our movement across hybrid and multicloud has to be part of our destiny. But we don’t want to get locked into proprietary black box tools. We need to have the flexibility of choices to create our pipelines, orchestrate across clouds.”
Figuring out how to develop infrastructure management capabilities using infrastructure as code will help managers optimize hybrid multicloud deployments. (Infrastructure as code is the management of IT infrastructure using configuration files rather than physical hardware configuration.) It could also help resolve some of the scale challengesSummit attendees discussed. But it requires team members who understand the various cloud structures to assess which cloud or on-prem is best for which use case.
“Infrastructure management in a hybrid, multicloud world requires orchestration across environments for different use cases.” – Click to tweet
The cloud offers tremendous benefits in terms of flexibility. Lots of enterprises – clearly – are taking advantage of it. But they’re also thinking more strategically about all of the options, from on-prem to any of the various cloud services from any of the various cloud providers. And they’re making decisions based on individual use cases, taking into account the very real management complexity that comes with spreading workloads across multiple environments.
Check out these posts in the End User Summit Top-of-Mind Blog Series: