Window of opportunity

In South Africa, cloud adoption by service providers and customers has been slow or even tentative. There are good reasons for this. Local bandwidth is still unreliable and expensive. There are major challenges in terms of compliance and governance. South Africa has strict legislation that prevents storage of sensitive information on off-shore servers that fall under other jurisdictions. Local companies, especially at enterprise level, are usually cautious in adopting new technology.

Public cloud services can be added in a step-by-step process.

Andrea Lodolo is CTO of CA Southern Africa.

While deploying a private cloud is a compelling choice, the focus now is on hybrid and public cloud services. Private cloud is typically an inevitable extension of virtualisation projects and does not have as many risk factors to be taken into account.

Despite all the risk factors above, there is a window of opportunity right now for service providers to offer cloud support and for customers to adopt public cloud offerings. There is a careful decision required: wait and go local or wait too long and the lucrative niche market in this country could be dominated by overseas service providers like Google and Amazon.

Surveys show that South Africa is three to five years away from having everything in place to offer complete cloud solutions. The same surveys show that local customers are only one to two years away from expecting to have these solutions in place.

This is the window of opportunity. If local companies fail to close that gap between availability and expectations, global players will probably move to fill the vacuum.

For the smaller business, right down to SME level, the situation is less complicated. Such companies are more agile and usually have lower risk issues. There are offerings already in the market that are basically “cloud in a box” solutions, which smaller businesses can deploy with good results.

Larger businesses are the ones that need to examine all the risk factors in fine detail. Beyond compliance and governance issues, there are critical concerns with security, availability, service delivery and, finally, very careful choices to be made regarding what applications and processes are suitable and “safe” to be moved to the cloud.

A careful balance must be struck. Failure to adopt cloud solutions within a reasonable timeframe means competitors might get ahead in terms of gaining the benefits, which are generally cost savings and improved service delivery. Poorly managed, planned and hasty cloud adoption carries many risks, however. These can include failure to achieve the hoped-for savings, no significant improvement in service, and exposure to potentially fatal risks of disrupted business processes or data leakage.

Avoiding the risks entails really detailed planning and having the technical resources in place to manage a cloud solution and make sure there are accurately measured results.

The first step, as with any network project, is knowing what the current situation really is. What applications are running, what workloads are they handling, what business processes depend on these? Cloud solutions, like virtualisation, offer an opportunity to rationalise, consolidate and optimise if the scoping and planning is well executed.

A useful approach is to begin cloud deployment with greenfields projects, rather than core processes in the business that will require a heavy investment in change management, and of course, significant risk.

Companies that have successfully established a private cloud solution have the advantage of a proven foundation from which to work. Building on that, public cloud services can be added in a step-by-step process.

One critical point that affects any proposed cloud solution is redundancy planning. At any point in the deployment, it is necessary to be able to backtrack to a previously stable condition. Depending on the terms of the SLA with a service provider, such redundancy abilities might be permanently required. While disaster recovery is usually addressed in contracts, it is something that needs to be fail-safe.

It is just simple common sense that a cloud solution should be built in baby steps.

Unfortunately, none of the steps outlined above are reliable or meaningful unless a company has the technology tools, policies, procedures and systems in place to check the results.

These measurements are not once-off. They have to be run continuously to monitor and manage the solution. This is where cloud solutions are really no different from any other network challenges.

Firstly, an accurate assessment of the starting point is needed. Then the results need to be measured as each individual change is made. Finally, the whole solution has to be monitored and managed to achieve optimal results and ensure that changing usage or business conditions have not affected the initially positive results.

Studies on virtualisation, which is a closely related matter, have shown that automation is a key requirement for achieving positive outcomes.

It is not surprising, then, that successful cloud solutions are equally dependent on the technology that manages them - and the careful planning that minimises risk exposure. So, finally, a company either trusts its service provider or it makes sure to monitor them closely.