The Amazon Web Services Transformation Day was held in Auckland last week and, as always, it was a popular event packed with geeks from across the country.

As you'd expect, there was plenty of chest-thumping about how great AWS is. Jeff Bezos' cloud computing outfit remains ahead of the pack with a third market share globally, followed by Microsoft's Azure, Google's Cloud Platform, IBM and Alibaba.

Even though it's 2018, and cloud computing is well-established, much of the discussion at Transformation Day centred around the holdouts who "prefer to schedule in downtime for businesses to patch their servers".

Which is curious because it seems that the arguments for running your own infrastructure, as opposed to renting it and letting it be someone else's capex and management headache, are worn thin by now.


In the past, there was some validity to the claim that cloud simply meant running your code on someone else's computer.

As time went on, the cloud quickly evolved into a new and, please don't hurt me for using the word, paradigm for computing.

Now, most of the techie people I talk to avoid setting up their infrastructure and simply spin up virtual machines on one of the cloud platforms, or use lightweight microservices like Docker containers for their projects.

Being able to access scalable computing and storage that's resilient and near users in almost every part of the world is a big attraction to New Zealand companies intent on world domination from the get-go.

In fact, it's hard to see how they could achieve that without the cloud, and of course, our rather good and now affordable network infrastructure.

Add to that the ability to quickly get new features (AWS boasted about how it released 1430 last year, with nine out of 10 being customer requested ones), much faster time to market for new service offerings (hours, not months), and the ability to right-size a solution to fit business needs and regular cost reductions. What are the reasons organisations still stick with their own infrastructure?

IDC senior analyst Doug Casement pointed out one possible reason. Some 60 per cent of local IT projects are fizzers.

With a failure rate that high, it's no wonder an organisation would hesitate to embark on a drastic move like shifting everything to the cloud — or even a hybrid platform that's partly on-premises, partly in the cloud.


Other reasons I've come across include:

• The existing solution was so painful to build that management would rather resign en masse than contemplate touching it for the next five years.
• Legacy apps that run core business processes and which cost a small fortune to buy that aren't happy in the cloud for various reasons.
• Billing and licensing complexity for cloud services, especially when an organisation grows with more advanced needs.
• Data sovereignty; sometimes it's not right using clouds in other jurisdictions, but there are local providers as well, like Datacom and Catalyst.
• It makes sense financially, but nobody in the organisation knows how to move things to the cloud.

The first three reasons point to organisations that aren't doing IT right, and the last three are relatively easy to fix — there's even a local outfit, Auldhouse, to provide AWS training now.

Are there other reasons why the cloud is wrong for your organisation? Drop me an email please and explain.