Dimension Data > Cloud > Why cloud costs have become an issue – and what to do about it

Why cloud costs have become an issue – and what to do about it

TwitterFacebookGoogle+LinkedIn

Peter Prowse: SVP Strategic Alliances Dimension Data

Muneyb Minhazuddin: Vice President Solutions Product Marketing VMWare

1.1.       Initial expectations of cloud were unrealistic

It’s common today to hear that more than 70% of workloads will run in the public cloud by 2021; but that public cloud hasn’t delivered the expected savings, and isn’t the best place to run all workloads. The dichotomy exists because few companies entered into the cloud with their eyes open. In the first rush (some call it Cloud 1.0) organisations were seduced by potential cost savings. They believed the hype that everything could run in the cloud, and didn’t recognise that workloads have different characteristics.

People initially viewed cloud as infrastructure-centric, and loved the ease of access for Dev/Test workloads, and/or greenfield applications. Most, if not all organisations, had existing application estates and didn’t take into account the complexity of migration, networking, security, compliance, back-up, operational management, and services integration.

While it was easy enough to spin up a new app in the cloud, they underestimated the complexity of migrating an existing enterprise application, and there’s more than one tale of an organisation publicly announcing they were moving their whole organisation into the cloud only to find – a couple of years and millions of dollars later – that it wasn’t feasible, or cost much more than they were spending today.

The complexity of migrating brownfield environments has driven a need for migration tools, such as those developed by VMware, to help organisations move traditional 3-tier mission critical applications like ERP, HR and database from within the corporate firewall to an external provider.

To be fair, some early adopters of hybrid cloud who had realistic expectations, and took a considered path to the public cloud, have been very successful.

1.2.       Applications are now driving hybrid cloud decisions

These days (Cloud 2.0) people are taking a more mature approach, asking themselves ‘Where is the best place to run this particular workload?’. The needs of the application are rightly beginning to determine the infrastructure decision, rather than the other way round.

Organisations are finding that, in many cases, what the app needs isn’t simply all on-premise, or all in the cloud, but a hybrid cloud, or hybrid IT deployment where the core IP and database stay in the data centre, and the front-end customer experience app sits in a public cloud.

What constitutes the optimum venue depends on the nature of the workload. Stable production workloads may be best left on dedicated infrastructure, or a private cloud which offers some variability of cost based on load. On the other hand, digital experiments, IoT projects, or digital customer experience apps may be better placed in a public cloud that you can turn up or down in real time.

1.3.       The key to cloud is to balance cost, control, and speed

The challenge uppermost in the minds of CIOs is cost containment, and they often focus too much on cloud’s ability to help with just this. The priority of their stakeholders in the business is speed of execution, and if they have their own budget they’ll go out and buy a shadow public cloud or SaaS service to satisfy that need.

Both approaches neglect the importance of keeping control of compliance, security, and data sovereignty. In Cloud 1.0 some financial services organisations spent millions of dollars on public cloud and lost track of where their data was and whether it was compliant and secure.

Public cloud provided speed of delivery and people thought it would save costs too, but found on its own it didn’t. Private infrastructure offered better control, but less speed and higher costs. Hybrid cloud does have the potential to balance all three – speed, cost, and control – but only if you take steps to achieve services integration.

To make hybrid cloud work you have to balance cost, control, and speed

What we advise organisations to do is keep all three sides of this triangle in balance. Ignore one side and you will run into unintended consequences. The focus today has to be on delivering what the business needs, from an environment spanning on-premise legacy, on-premise private cloud, purpose-built clouds, and hyperscalers, in a way that balances the need for cost, control, and speed.

1.4.       Virtualise your data centre to drive services integration

What companies are now coming to realise is that hybrid cloud isn’t really about infrastructure; it’s all about services integration.

One way of achieving services integration is to virtualise your whole environment (compute, network, storage, and security), so that all its components (on-premise, data centre, co-located, and cloud) work in a consistent manner. This makes your workloads hardware independent and portable. It’s also an essential step towards automation of the application lifecycle.

It’s very important to our clients to achieve this portability. Over the last few years we’ve seen clients voting with their feet – moving out of the data centre to co-location facilities to balance cost, or moving workloads to public cloud and then back onto their premises because they need dedicated infrastructure.

1.5.       The right hybrid architecture can help you get business value out of your data

What organisations really care about is their data. They want to analyse it to make business decisions in real time. Applications and infrastructure are only the life-support system supporting the data.

In the past, data resided in the data centre. But with IoT it’s being created by new kinds of devices at the edge. It makes sense that the apps that process this data should reside at the edge too.

Hybrid cloud environments that feature services integration have the potential to allow organisations to get insight from their data in a speedy, compliant, and secure manner, regardless of its location.

This is why organisations are investing with companies like VMware and Dimension Data – to build the capability to abstract their workloads from the underlying infrastructure, and provide the automation to manage them across a diverse IT eco-system.

We help clients to employ containerisation to make their applications portable, so that the needs of the workload can determine the infrastructure. We deliver services integration to break down the silos of cloud that have been created, and make on-premise and cloud work the same way.

View the 2018 Hybrid Cloud IT Trends here