Cloud is on its way to becoming a business necessity. Gartner has predicted that by 2028, more than 50% of enterprises will use industry cloud platforms to accelerate key business initiatives. While many organisations are already taking advantage of the cloud, navigating the plethora of solutions and service offerings that are seemingly launched on a daily basis, others are progressing at a slower pace with their digital transformation efforts.

Seelan Gopalan, product portfolio manager, cloud and communications, at NTT Data’s GTM division, says South Africa’s business environment is challenging; businesses are under cost control pressure, and loadshedding continues to be a business risk. Cloud adoption is now seen as a de-risking initiative, as the datacentre will take responsibility for the uptime SLA.

Gopalan compares the cloud modernisation journey to upgrading to a new car. There are a number of questions you need to ask yourself: has my family grown? Is my car being troublesome? Do I still have a warranty? What about a maintenance plan? Where is my car sitting on the security list in terms of hijacking? You’ll make a decision based on this and I think that’s very similar in the infrastructure cloud space,” he says. “Everyone’s on a different journey and there’s different trigger points that will make them do something.”

There are always going to be organisations that jump in head first with a full lift-andshift strategy. Darryl Govender, a director at Synthesis Software, suggests firms do an assessment of their current infrastructure. “What we recommend is first installing an agent so you can get a look at the estate more holistically and gather metrics around things like CPU utilisation, RAM, storage, network, bandwidth and so on to start to map specific exercises around lift-and-shift.”

Govender says that on-premises or co-located environments are often overprovisioned, because a business wants to make sure it has enough capacity to handle peak demand, unexpected workloads and future growth. “Hardware or space is paid for upfront but underutilised and if you were to take the exact same spec and liftand- shift to the cloud, it’s always more expensive. There’s no two ways around it,” he says.

Govender says businesses need to understand their demand. “Instead of having double the CPU, double the RAM, double the storage, we’ll get it to the right amount based on a four-week sample period to see what the environment is doing,” he says. He adds that the workloads that make the most sense for cloud are the ones that a business has control over – its own software. “When commercial, off-the- shelf solutions are hosted in the cloud, it becomes a little bit trickier because you don’t have control around being able to choose which services to use or where you would deploy them,” he says.

Govender’s recommendation for an infrastructure refresh is to do what he calls an “optimised” lift-and-shift, or looking at all the different moving parts to see how to best leverage the right regions, size of instance and type of service. “There’s a lot of nuance when it comes to engaging an on-premises installation of certain software versus the cloud,” he says. While this kind of approach will definitely help with lowering the costs of cloud over time, Govender says that cloud-native is the Holy Grail. When workloads are developed from the ground up to be cloud-native using cloud-specific services, organisations can fully realise the benefits of cost savings, business agility, operational efficiency and developer productivity.

“Ideally when you’re doing a refresh, I’d suggest that you look at the software you control. With your own development team, you can then actually go and deploy it with cloud-native technologies.”

For commercial, vendor-hosted applications, Govender suggests switching to a SaaS version of the software if it’s available. “They’ve made their own architectures that work well in the cloud and you just deploy it into your own environment.”

When considering a cloud-native strategy, containerisation has changed how applications are managed and deployed in the cloud. “It’s the consistency of the user experience,” says Eugene de Souza, Red Hat South Africa’s regional cloud business leader. “Whether it’s a developer or an IT operations person, containers are consistent across clouds, laptops and on-prem environments…but that does mean that you have to manage it yourself. You need to decide what core business you are in as a customer.”

Gopalan says virtualisation still plays a vital role in infrastructure optimisation. “Virtualisation is a key part of what we do and it’s how a lot of our clients transition,” he says. “If you look at the cloud journey, it will typically be on-prem into virtualisation and then virtualisation into Kubernetes and containerisation.”

Virtualisation brings benefits that range from flexibility to the better utilisation of physical resources. The ability to consolidate multiple workloads onto fewer physical servers or platforms is more efficient when compared to running each workload on its own dedicated physical server or environment. “There are a lot of different offerings within the virtualisation environments like Hyper-V, VMware or Nutanix,” says Gopalan. “It’s a journey. You have to get from A to B to C and what we are doing is enabling the first part. And if the client does eventually end up in a containerisation world, my job is done because I’ve kickstarted that journey.”

There are so many features and functionalities available in the public clouds and it’s as easy as ticking a box to add something even if you don’t need their entire toolkit. “There’s no point of having a Rolls-Royce to get from A to B when a Toyota Corolla will do the same,” says Gopalan. He suggests having some sort of cadence rhythm around reviewing workloads. Instead of doing it at the end of the financial year, increase the frequency “to make sure it’s the right use case for you, and then continually check that the right things are in the right place. The truth is, and this is hard for the hyperscalers to accept, is that sometimes it’s better to move back [on-prem] than it is to persist with a strategy of being in the cloud.” Moving back on-premises or to a colocation environment is a growing trend.

If the client does eventually end up in a containerisation world, my job is done because I’ve kick-started that journey.

Seelan Gopalan, NTT Data

According to a survey by IDC Research, 71% of the respondents said they were planning to move some or all of their workloads from public clouds back to private IT environments within the next two years. Only 13% expect to run all their workloads in the cloud. “You get businesses that are significantly disgruntled with not realising their cloud investment, but we’ve also seen cases where people have moved back specific workloads,” says Govender. “I would suggest a hybridised strategy across the cloud, and even across cloud providers. “This makes more sense than just having an all-in strategy.”

Effective cloud modernisation requires comprehensive planning and change management strategies to ensure successful adoption and integration. “There’s different parts that make the journey tricky in terms of realising the investment and it takes several years to get right,” Govender says, adding that one of the biggest mistakes organisations make is assuming that certifications are equal to skills. Cloud skills are built with deliberate practice over years and converting on-premises engineers to the cloud isn’t an overnight process. It’s likely that businesses will go through a series of missteps throughout the cloud journey. “If you don’t have the stomach for several years [of work], then that’s probably not a good move,” he says. “But to accelerate the journey using best practice, you need to learn from the right instruction book, the hyperscaler partners in South Africa, instead of trying to figure it out yourself.”