By Kabelo Makwane, managing director of Accenture South Africa’s Cloud First business

‘Lift-and-shift’ – the classic approach to cloud migration – means this: take all your existing IT structures, procedures and applications and transfer them to the cloud.

On paper, such an approach seems sensible. Yet, while cloud migration may mean the benefit of eliminating physical hardware infrastructure on the IaaS layer, transferring needless complexity elsewhere can in fact cause a business’s total cost of ownership on the cloud to escalate. More broadly, the speed and effectiveness of any cloud migration is highly ecosystem-dependent; the application landscape forms only one part.

The reason is the amplifying effect of the cloud. Poorly designed or inefficient IT operating models have their flaws emphasised by the as-a-serviceparadigm. On one hand, this means that if businesses haven’t designed an efficient service management model or built their application sets with an eye on leanness, they may end up spending more to maintain those functions in the cloud. Other ecosystem areas – including governance and management – also require an expert guiding hand when it comes to cloud pivots.

All businesses are likely to witness a degree of change in their IT operating models when undertaking a migration. For one, cloud changes the way people operate within an IT environment. Because of the high degree of automation within the cloud, for example, people are both freed up and required to work at a higher order, both in terms of IT ops and administration.

Given this consideration and others, enterprises need to assess how fast and to what extent they want to rotate to the cloud. Will it be full-on ERP in the cloud? Full CRM in the cloud with mobility? Partial? The growing impetus for enterprises to forward integrate using cloud services is a related consideration. Indeed, the trend is coming to demand increasing attention.

Consider the near-ubiquity of self-service banking apps. Today, we’d say that a financial services provider without one belongs in the stone age. And yet banking apps have been a relatively recent development. Healthcare is seeing similar developments with the rise of wearable tech. The industry is likely to witness the same level of forward integration, requiring consumers to take some degree of responsibility in terms of managing their own health and wellness. 

When it comes to the application-level changes cloud-minded businesses undergo during migration, natively cloud apps will be those requiring the least remediation. Others will likely require more rationalisation and consolidation. Hearkening back to the concept of leanness, enterprises often have a large real estate of custom applications spun out by IT over time. Performing analysis and due diligence on the application landscape before a migration is key. Migrations present a good opportunity to do a little spring cleaning.

Integration layers – what we used to call ‘middleware’ – are key here, yet the area has also witnessed something of an overhaul. There are still middleware players – they haven’t gone away; they have just reinvented themselves. When it comes to modernising applications, containerisation allows for application migration without full reengineering. From a DevOps perspective, tools such as Kitkat and Cucumber can help with rewriting applications so that they become cloud ready. 

Another key aspect of the migration ecosystem is governance. The sphere may be easy to overlook, yet business need to ask themselves key questions: ‘Do we have a governance model for operating in the cloud? Does it extend to security, risk and identity management?’ In this line, a concept known as shared responsibility is gaining traction. The approach means that enterprises cannot negate their responsibility to secure their applications and data, even if the cloud is more secure than their existing on-premise environment. It’s about a joint effort.

A final ecosystem component is that of cloud service brokerage. Businesses’ habits of consuming federated IT infrastructures and application environments haven’t gone away. Interest in engaging multiple providers is still driven, in some cases, by organisations shying away from vendor lock-in. Alternatively, enterprises may buy different applications for different purposes, informed by their strategy.

When it comes to consuming federated or disparate environments, businesses often require some level of service brokerage. So there is an emerged discipline – within hyperscale cloud particularly – wherein specific boutique outfits are now building brokerage capabilities. The result is that organisations can consume from multiple cloud types on their own terms.

At the broadest level, to be able to fully realise the promise of cloud in terms of speed, efficiency, cost effectiveness, scale and optimisation, capable oversight is required. The result is that cloud management platforms have emerged – many hyperscale cloud providers are now building management capabilities onto their offerings.

AWS’s EC2 platform for example has seen ongoing investment to natively embed functions around identity management and application integration in the cloud and Microsoft Azure Cloud also sees heavy investment in this area as well in order to build native capability. By embedding such tools, it becomes far easier for businesses to deploy cloud applications as extensive knowledge of middleware and integration tools are no longer required.

In sum, effective, cost-effective cloud migration depends on thoughtful deployment on a host of levels – applications, governance, management and brokerage among them. In short, it’s all about leveraging the ecosystem.