Cloud Live: The seven sins of multi-cloud

Kris Saxton of Automation Logic divided the cardinal sins of multi-cloud into abstraction, ambition and approach

Multi- and hybrid cloud are relatively new terms, still misunderstood and misused. Kris Saxton spoke about the seven sins of multi cloud that he's observed in his work as co-founder of professional services consultancy Automation Logic at Computing's Cloud & Infrastructure Live event this week.

Abstraction

Sins one and two were ‘too much' and ‘not enough' abstraction. There is still a lot of confusion around the multi- and hybrid cloud terms, and that is a problem because too much abstraction can lead to unmet expectations.

"It's quite common in a multi-cloud engagement to hear senior stakeholders talking about how they're going to ‘play the cloud providers off against each other'" and make choices and place workloads based on tiny price differences. "That is absolutely not going to happen," said Saxton.

In an instance like this, too much abstraction could lead to frustration and poor investment decisions: clarity about what multi-cloud is and does is vital.

Saxton defined under-abstraction as missing the importance of software, which will slow down business:

"‘Software is eating the world', as the saying goes - and infrastructure is absolutely no exception to that. Failure to recognise that has been creeping into infrastructure, it's quite easy to miss that you can do almost a complete multi-cloud implementation in software now.

"If you fail to recognise that, you're going to miss all of the opportunities and benefits that come with doing things in software, related to speed, quality and cost."

Many large firms have some form of governance that chooses the tools that the infrastructure team will use. They must have someone that understands "the developer mindset" in those governance layers to ensure the correct level of abstraction.

Ambition

Most of the sins fall under ambition - because IT teams naturally want to reach as high and as far as they can.

The counter to all of the sins of ambition is a strong product owner, who can take a principles-led stance on architecture, discourage shared services and move towards codified patterns of use and focus efforts where they are needed.

Sin three is trying to be an IaaS, instead of just consuming it. There is no longer the need for persistent infrastructure that there was in the pre-IaaS and -PaaS days in most cases; failing to recognise that will lead to oversized teams because you are managing pointless infrastructure.

The next sin is trying to build too much too soon, specifically when it comes to compliance and governance, which are a necessary part of multi-cloud implementations in large firms. Delivery teams need to decide whether to do this up-front or at the end of the process.

There is an argument to be made for both approaches. Building infrastructure without these compliance layers is "relatively easy" - but, at the same time, you need to have something usable and useful to show to stakeholders as soon as possible.

Having seen both, Saxton favours the latter: "Focusing on building useful things...is more likely to bring you success than spending a long time building highly-compliant small units of value."

Sin number five is ‘overcomplicating'. "Engineers are not devoid of sin," said Saxton; they tend to overcomplicate systems because it's exciting and interesting - and "because they can." "That will lead to awesome infrastructure, but also awesome operating costs."

If a system is very complex and advanced, it is often difficult to hand it back to a client and they end up with a dependence on you.

The final sin of ambition is ‘faux agile'. Every delivery now has ‘agile' tagged on to it, but very few are actually doing it properly. The problem here is in a mismatch of expectations and management styles between engineering teams - who might have been practising agile for decades - and the rest of the business. This can lead to chaos and low morale.

To combat faux agile, there needs to be investment in retraining and education for middle management, to teach them about the foundations of lean and agile thinking.

Approach

Saxton called the seventh and final sin ‘organisational transformation', and is closely related to ‘faux agile'. IT and engineering teams are often experienced and comfortable with product- or service-oriented reorganisations, but the surrounding organisation is left largely unchanged, leading to a disconnect between managers and managed, which creates stress and disengagement.

Much like the sixth sin, the way to get past this is to invest in or restructure middle management to align them with agile and lean practises.

The virtues of multi-cloud

There are seven sins, but only three virtues: the things that you must do to avoid sinning. They are: