DevOps: Quicker, Cheaper, Better

Harold Petersen discusses critical factors in creating a DevOps environment that delivers a trail of business value.

©iStock.com/ArtyFree

©iStock.com/ArtyFree

The jury is in: DevOps is a better, faster and cheaper way of deploying IT solutions. You no longer have to compromise by choosing just one or two of these when kicking off a development initiative.

But what’s needed to achieve all three?

In a previous article, Bring down the wall between Dev and Ops, I said that DevOps is not just a matter of investing in one of the readily available DevOps tools. 

I’ve come across several organisations that recognised the promise of DevOps and have gone so far as to invest in establishing the technology required for automated provisioning of platforms – as well as automated building, testing, deploying of software releases - yet nothing has changed in terms of actual results and business value.

Sometimes this is due to a lack of leadership vision, direction and governance; other times it’s because, deep down, people across the ICT organisation are too set in their ways to enable DevOps to work.

As a result, after investing real dollars in DevOps tools and triggering great expectations, no improvement in quality of applications, or speed and frequency of releases is achieved. 

However don’t blame the DevOps principles! An organisation needs more enablers than just tools to achieve quicker, cheaper, better solution deployment and realisation of value.

Certainly, DevOps tools are a must for creating an optimised end-to-end delivery solution, but as the old saying goes, ‘A fool with a tool is still a fool.’

Certainly, DevOps tools are a must for creating an optimised end-to-end delivery solution, but as the old saying goes, ‘A fool with a tool is still a fool.’ 

So what else do you need to get started  ? The following are some of the critical enablers that – together with automation – underpin a holistic DevOps capability:

  1. The right organisational structure
  2. People with the right skills
  3. A culture that wants to work together to deliver   

Let’s take a quick look at each.

1. The organisational structure to make DevOps possible

DevOps calls for a less hierarchical and less segregated structure that devolves responsibility to expert individuals within teams that between them master the breadth of business requirements gathering, development, integration, testing, procurement, deployment and enabling of operational production in a compliant manner. 

In the new world of DevOps, these ‘business system teams’ need to be able to leverage reliable, scalable and secure IaaS and PaaS self-service solutions – without the need for complex engagement with infrastructure and network teams to obtain production resources.

Their focus needs to be on the end-to-end business solution, its functionality and its service levels (availability, capacity, security and contingency targets). This is fundamentally different from the traditional, siloed structure, where each group tosses its work ‘over the wall’ to the next one.

The Born-Digital Way

Spotify, the music streaming app, has adopted an interesting structure that enables DevOps by organising its people into Squads, Tribes, Chapters and Guilds. Instead of traditional ICT siloes, Spotify has formed Business System Teams called Squads, which are similar to Agile Scrum teams and apply Lean Start-up principles – using the slogan ‘Think it, build it, ship it, tweak it’ to support the entire life cycle of particular services.

Tribes are collections of squads that work in related areas, together looking after the life cycle and continual improvement of one aspect of Spotify’s service delivery, such as its music player, the mobile app, the store or search functionality. Tribes   behave like a start-up company and have a degree of freedom and autonomy. Each takes end-to-end responsibility for innovating, designing, building, hosting, testing, releasing then beginning again – with multiple projects in play at any given time.

One of the keys to ‘quicker, cheaper, better’ is a fundamental re-focussing of your business solution  teams to a horizontal rather than vertical structure.

The teams need to be enabled to focus on end-to-end business solutions rather than having to engage application development teams, infrastructure teams, test teams, security teams etc. for each of their contributions to the business solution.

Break down the silos and think ‘end-to-end’.

As an aside, we’re noticing growing interest in the Technology Business Management (TBM) approach to creating cost transparency and optimising the ‘Running’ as well as investments in ‘Changing’ enterprise ICT service solutions, delivery and support.   

TBM provides the tools and techniques required to manage, measure and communicate the value of IT services; It enables you to become the CEO of your technology business.

If technology is managed like a business, it:

  • ‘manufactures and provides’ business technology to the rest of the enterprise, 
  • speaks business language in terms of reliable financial transparency, service functionality, services levels and
  • has a track record of optimising costs and delivering demonstrable business value. 

A DevOps capability certainly meets the need under this strategic model to optimise the ‘changing’  of technology business solutions and enable the ‘running’ of technology services: quicker, cheaper and better.  

2. The people to make DevOps happen

In the traditional model there are project planners, Business Analysts to define the requirements of a new solution, developers, testers – then a whole team of Ops folk whose mantra was ‘keep the lights on, minimise issues and never take risks’.

Seek T-shaped skills

There is an increased need for ICT people (particularly within DevOps business system teams) with so called ‘T-shaped skills’.

Aside from their own deep expertise, they need to understand DevOps enablers such as processes and tools, as well as the importance of other people’s subject matter expertise and of course the business requirements and value expectations.

The term ‘T-shaped skills’ isn’t new: it’s been in use since the 1990s in consulting and technical companies.

The term is defined as “The vertical bar on the T represents the depth of skills and expertise in a single field, whereas the horizontal bar is the ability to collaborate across disciplines with experts in other areas and to apply knowledge in areas of expertise other than one's own.”

DevOps requires that you crunch all these skills and people together within, cohesive, end-to-end business system teams with the same objectives: to optimise value to the business and deliver quicker, better and cheaper. To realise the benefits of DevOps, you need the people with the right attributes – and it should definitely inform your strategy for acquiring new hires.

If your existing employee’s skills are developed and nurtured to support the change, then employees can be empowered to deliver effective change.

However, even an empowered employee may not be intrinsically motivated to deliver effective change.

The correct job and team design, or more broadly, the right organisational structure as discussed in the previous section, are critical to developing motivated people with the right skills and goals aligned with DevOps principles.

This might involve re-training and convincing them of the personal and professional benefits of working the DevOps way. People in the habit of holding on, “because no-one else can be trusted with it and I don’t want it going pear-shaped”, need to learn to let go and trust. But importantly, you need to provide an environment that fosters this trust.

3. The culture to make DevOps work

So, DevOps demands a culture of co-operation and collaboration: shared desire, attitude and behaviour to make the concept of end-to-end service delivery work.

DevOps isn’t about free-range tech-heroes, tiger-teams trying to outscore other teams or inviolate fiefdoms. It’s about communal, concerted effort towards a common cause – which requires instilling a sense of shared responsibility within the entire team.

DevOps isn’t about free-range tech-heroes, tiger-teams trying to outscore other teams or inviolate fiefdoms. It’s about communal, concerted effort towards a common cause – which requires instilling a sense of shared responsibility within the entire team.

Meanwhile, a flatter organisational hierarchy and the establishment of business system teams don’t deny the need for leadership. DevOps calls for true servant-leadership and lots of people skills to break down the divides between the traditionally opposed cultures of Dev and Ops. 

On the one hand the teams need to be served and facilitated to self-organise and be the engine for quicker, cheaper and better delivery. On the other, leadership is needed to establish shared vision, governance and an environment in which the teams and their culture can flourish.

Where to start?

Unfortunately, there’s no magic DevOps wand. You can’t say, “As of 1 January, we’ll all be a single, end-to-end team and all our many processes will be streamlined and automated!”  

But you can pick a particular (new) development/business system and apply the DevOps approach to it. This will give you the chance to test the tools and principles – giving you and your people (and the business!) the confidence to take it on to other business systems.

Applying DevOps to updates or extensions of legacy applications may not work anywhere near as easily, because of entrenched processes, non-virtualised hosting and traditional thinking. 

So two-speed or bi-modal IT is the way to go – at least in the short-to-medium term. As we’ve said previously on the subject, “The important thing is to get faster and better at the things that matter to the success of the business. That’s usually the 20% that delivers the most value for your customers and your organisation. The rest of the organisation then needs to be tuned so that the speed they operate at doesn’t inhibit progress.”

Another way to work towards ‘faster,  cheaper, better’ is to set your team small monthly goals towards DevOps. Quarterly or annual is simply too long, too overwhelming and you’re unlikely to achieve them in the press of everything else you have to accomplish.

Importantly, rather than just implementing DevOps tools and virtualised environments or just sending your people to a DevOps training course, focus on the holistic blend of enablers that underpin the value of DevOps. 

At CSC Consulting we’ve developed a proven model covering eight key capabilities for getting DevOps happening within even the most complex IT Organisation. This article is somewhat of a simplification, but I hope it inspires you to start thinking of the possibilities – or even take the DevOps plunge. 

Related Reading:

Related Services:

About the Author

Based in Canberra, Harold Petersen is Director Advisory for CSC Consulting, responsible for overseeing ten National Practices in a consulting force of over 700. Specialising in SIAM, PPM, ITSM, Lean IT, ICT Strategy & Governance implementation projects and programs, his 25+ years’ industry experience includes Director for CSC Consulting in Singapore and Malaysia, IT Service Delivery Manager for a major Australian bank and IT Service Executive for Asia Pacific accounts within a global outsourced services provider.