Fixing IT

Shallow thoughts on the future of tech

I've been writing a few articles on the topic of "tech" strategy recently, and the disdain in which I hold the idea of everyone getting techie. I think this will be the last one in this series, so you can all rest easy for a bit.

I don't believe that everyone should be into software, or development (and widely misused term "coding"). My previous articles talk about why. What do I believe we should do?

Without wanting to get into product endorsement, it is clear to me that while businesses need to leverage technology, they need to do so by leveraging technology services from others. Unlike the SAP ERP idea, I don't believe in one provider taking on all the technology supporting your business. If a business is to innovate in what they do using technology, they need to be able to combine technologies in ways that makes sense for them, not how a service provider thinks they might need it.

At a high level we call this integration. Typically integration is a technology intensive exercise but the situation is starting to improve. There are products and services hitting the market that enable integration with less effort. Most of these offerings are new, both in terms of their technical characteristics and sometimes their commercial propositions. Here are some them.

Simple Integration Frameworks

There are consumer or low end integration frameworks, typically offered as a SaaS, that allow users to link up business functions to respond to events, import data and run simple tasks. IFTTT or Microsoft Flow are examples of these.

They are generally fairly limited what they can be integrated with out of the box, the volume of transactions they can process, and with the logic that can be applied to the integration points or the data flowing through them. As such, their use is generally restricted to areas where there is no complex policy or logic decision required.

They are maintenance free, in as much as integration occurs over the internet with the SaaS provider's APIs.

Some can be extended by the user though, and as they become more capable and complex, they become...

Complex Integration Frameworks

Conceptually similar to simple integration frameworks, but with more capability and less rigid structures, these are also services and tools that allow businesses to integrate large numbers of components, apply complex logic and policy to high volumes of data. Mulesoft and Apigee are products that can perform in this role. There are heaps of others. Just don't say ESB.

As a tradeoff, these frameworks are harder to configure and will probably require software development of some sort. They are available both on-premises and cloud based, so low-maintenance options are available.

Serverless computing

These offerings a relatively new to the market and essentially allow an organisation to run their code in a technical environment that is entirely managed by the provider. AWS, Google and Microsoft all have products that are adjacent to their X-aaS offerings.

Yes, you still need to write software in the same languages you would otherwise use, but the operating environment is no longer your concern. With a well designed architecture, an application or business function running in a serverless environment can become a very low maintenance proposition.

I still need code?

I'm afraid so. However, as the industry tends to move towards more cloud based offerings the need to carry a lot of expertise in your organisation is reducing. IaaS and PaaS both assume some level of ownership of the infrastructure for the user. Cloud based integration services and serverless technologies go a step beyond that and allow an organisation to focus on developing the business logic required and ignore the supporting technologies.

Over time, we should see solutions that offer these "low infra" or "serverless" options become more numerous and more sophisticated, such that we see the transition as being from on-premises big-iron data-centers, to the consumption of cloud based logic. Where we currently are, with private or public cloud based VMs and containers is merely a way point in the journey.

Where am I going to get my code?

From Tradies. Meet your new dev team. Rawr!

What are Tradies?

For the unAustralians among you, a Tradie is a tradesperson, say a carpenter or plumber or electrician.

I'm not saying your plumber is going to write your code, but I think the engagement model for IT will change as the industry matures. If IT can become mature enough and the IT products on the market are presented properly, we should be able to get to a point where a business does not need to maintain a significant ongoing technical capacity. In the same way that a business in an office building does not maintain a full time aircon engineer, nor should it maintain a full time developer.

Because there will be no need for the ongoing operation of IT infrastructure the market should be able to move to a model where IT changes are small discrete pieces of work, some of which are able to be handled through configuration rather than code changes. We can see this now in the application of microservices architecture.

Enterprises would still be dependent on service providers (say, the provider of payroll systems) for complex changes that are peculiar to a particular subject, but the integrations that are particular to the enterprise are small and well contained.

However, if an enterprise wanted to do something innovative with its payroll data (like feed it to an ML forecasting engine for arguments sake) a short, sharp piece of development effort would be required to enable that data flow. This might enable the realisation of significant innovation and value to the enterprise, but little software development. This is the proper state of the world, as the intrinsic value is in the idea, not the effort required to realise it.

So, then we see the need more for short term engagements to deliver discrete pieces of work, perhaps with a maintenance arrangement applied. This looks a lot more like a maintaining an aircon system. a lift or unblocking a drain, rather than a team of devs on the lower floors of an office block.

This is my Tradie Model for IT. Compared to today's situation, it's not really pretty for IT workers. But it is an example shown to us by mature industries like building construction. They have architects, engineers, builders, maintenance. However, most business do not retain the services of these people as part of the business of innovation. This is what a mature IT industry looks like.