The pandemic-fuelled boom in enterprise cloud adoption is still going gangbusters.
In its latest forecast, Gartner expects worldwide end-user spending on public cloud services to reach $494.7 billion this year (a significant increase of 20 per cent from 2021) and nearly $600 billion next year.
Gartner vice-president Paul Delory recently offered up a glimpse of the hot-button issues organizations may encounter as they boost their cloud budgets. Here are highlights from Delory’s webinar presentation on the top six trends in enterprise cloud.
1. Cloud teams will optimize for business outcomes, not technical implementation.
Delory said that while business value has always been the primary goal of cloud deployment, “in talking to clients, I do think a lot of them have lost sight of that.”
He issued a gentle reminder to cloud teams: focus on business outcomes to avoid getting swept away by the undertow of complicated IT minutiae.
“You can’t be a gatekeeper. Remember, your end-users are trying to create business value. They’re trying to innovate, they’re trying to do new things. You don’t want to get in the way of that.”
Here are some steps Delory suggested to ensure your cloud strategy stays aligned with business goals:
- map CIO-level priorities to cloud architecture initiatives
- build a reusable cloud adoption framework to onboard new services quickly with best practices learned over time
- create cloud advisory councils spanning IT and business
- set up cloud communities of practice (“clubs” of non-IT employees interested in cloud)
Delory said cloud advisory councils and cloud communities of practice help “make sure that users have a voice, (so) we as engineers are building what the users need.”
2. Hybrid and multi-cloud adoption will increase operational complexity and cost.
“Multi-cloud is hard,” Delory acknowledged. “(It’s) making your architecture more complex and expensive.”
Delory pointed out that expanding beyond a single cloud also complicates matters such as data management and governance, data dispersion, IT integration, cybersecurity, asset tracking, vendor relationships and employee skill requirements.
“Multi-cloud means a lot of changes for people, policy and process(es) as well,” he said.
Here are Delory’s tips for managing the cost and complexity of hybrid and multi-cloud environments:
- prioritize a primary strategic cloud provider
- only add a second cloud provider when you have business needs that can’t be met by your primary provider
- place workloads based mainly on their integration and data affinities
3. Business resilience will be built into the application architecture.
Two factors are driving this push for more resilient IT architecture, according to Delory: ransomware and the rise of cloud-native apps.
To recover from ransomware, he suggested building “an immutable data vault” that is read-only and completely separate from production, development and testing. He said this area can act as “a secure, isolated recovery environment … where you can restore your data, where there’s no chance it can get re-infected.”
Similarly, he said, “cloud-native applications really need to be treated differently (than traditional apps) when it comes to backup and data recovery.”
There’s a movement afoot to design systems “where replication, redundancy, performance (and) management, those are features of the application code itself (so) we’re not needing to bolt that on at the infrastructure layer,” Delory said.
Delory’s recommendations for enterprise cloud IT resilience:
- start by building resilience into cloud-native applications
- prioritize ransomware defence and recovery, especially in secure, isolated recovery environments
- create KPIs/metrics for IT resilience
4. Distributed cloud will displace private and hybrid cloud initiatives.
Delory defined distributed cloud as “public cloud services that are running outside the provider,” either inside your data center (wait, what?) or pushed out to the edge.
“Essentially, you have a substation of a cloud provider in your data center. You’re going to have a little baby Azure, for example, that runs in your data center and is running the actual cloud provider services on-prem. It’s on-prem infrastructure but it’s tethered to a cloud provider in some way,” Delory said.
In what type of scenario would an enterprise deploy distributed cloud?
“If you have a workload that can’t move to public cloud but you still want cloud services,” he said. “AWS, Microsoft or Google, they’ve already built the (public cloud) service. Now they’re going to give you some hardware and a way to run the service on-prem. (Their) sales pitch is don’t try to build a private cloud, you can buy it from us.”
Delory’s suggestions for adopting this distributed cloud model:
- evaluate distributed cloud substations offered by major cloud providers
- compare these to hyperconverged infrastructure (HCI) products from major data centre vendors
- replace “homebrew” private and hybrid clouds with distributed cloud offerings and/or HCI
5. Containers and serverless will become a foundation for application platforms.
Delory clarified right away that it’s not really serverless technology, but more like containers-as-a-service (CaaS). In a nutshell, the cloud manages your servers so you don’t have to. As an example, he said Kubernetes could be used to orchestrate containers, making workloads portable and enabling them to work with hybrid or multi-cloud architectures.
“The idea behind a serverless container architecture is that you package up whatever code you want to run in a container, then you just feed it into this platform and that container just runs. You don’t need to worry about setting up the underlying infrastructure; you don’t manage that. This is something that the cloud provider can take care of for you,” Delory said.
Some of Gartner’s tips for CaaS and serverless architecture:
- optimize apps for cloud-native architectures, including CaaS and serverless
- introduce serverless container architectures into DevOps workflows, then expand other use cases from there
- use public cloud Kubernetes and CaaS services unless you have specific compliance or functional requirements they can’t meet
6. The crisis-level skills gap will compromise cloud innovation.
“I really do think this is a crisis,” Delory emphasized. “(The skills gap) is a barrier in a lot of cases to going cloud-native. It’s not the tools, it’s that you don’t have the people who know how to use the tools.”
His suggestions for coping with the skills shortage:
- budget time and money for staff to learn new skills
- use skills growth as a KPI for teams and individuals
- tie skills growth into each employee’s annual review and bonus pay
- prioritize Kubernetes and DevOps skills