9 common pitfalls of building an image lifecycle management solution in-house

Post • 5 min read

Building an image lifecycle management solution in-house is tough. Most of the time, in-house architects lack the required cloud skills, or are spread across different areas, so can’t give the level of attention needed. 

And, even if they are able to build something, it’s often a case of: ‘Is it working? Yes. OK, let’s move on, we’ve got other fires to fight.’ 

Many companies underestimate what they’re letting themselves in for when they choose to take this in-house. These are some of the issues many will face.

1. Architects are spread too thinly

Often architects working in-house aren’t able to dedicate the time to deliver quality image lifecycle management solutions that are well maintained. They will build, but then have to move to another project. We’ve seen this many times, where we’ve picked up some script in a pipeline that was created of very bad quality, or poorly maintained.

Even if the scripts are good, they’re only good for a short time – perhaps a few months. When cloud providers release updates or new services, those scripts are not maintainable, so you can’t really develop or evolve them.

Don’t blame the architect, they’ve got other things on their plate. But with a fully-managed approach, solutions are built efficiently, and any issues are solved in less than a day. We’ll handle it. Then, the architects can focus on the more valuable tasks within their organisation.

2. The costs of keeping it in-house start to snowball

Companies often underestimate the resources, skillset and time needed to build and maintain an image lifecycle management solution. It’s a significant investment to deliver in-house, even for capable architects. For many, you might need to train staff, then dedicate three or four full-time architects for six months or more. This can be a drain on resources that might be better used elsewhere. And, if it’s not managed efficiently, costs can snowball.

3. A lack of multi-cloud expertise undermines solutions

Often when companies are trying to take image lifecycle management in-house, one of the reasons they fail is a lack of multi-cloud expertise. The architect will be tasked to develop a singular solution to work across any public cloud.

To achieve this, an overarching vision is critical. Those involved need an understanding of how to approach the different cloud environments. They don’t need to be an absolute expert in every single cloud, but it sure helps.

And at the very least, you need a level of expertise across all clouds to understand how things differ on Azure, GCP and AWS, and how to deliver the image so that the end user has a similar experience on each of them. Working with a cloud specialist that lives and breathes public cloud is a safe bet.

4. DevOps skills are overlooked

Also, it’s quite common that cloud architects, particularly in large companies, don’t prioritise DevOps skills. They’re great at designing solutions and creating the high-level concept, but often they aren’t up-to-speed with the tools working under the hood.

This means they don’t know how to effectively convert the vision of having a streamlined process working in the cloud, into actual technical solutions that can be developed.

5. The capabilities of open source tools are overestimated

It’s true that most of what’s needed can be achieved with the open source tools available on the market. However, these are really just the starting point, solving around 10-20% of the requirements. That leaves 80% of the job left to do. To get to 100%, things like automation, streamlining processes, securing the image, and more are required.

And often the team tasked with building an image lifecycle management solution using open source tools lack some of the programming skills. They might have some solid scripting skills, but can only create a bank of scripts, not a product that is well maintained and reliable. So the quality of the solution would not be on the same level as fully-maintained SaaS product like IBM Multicloud Machine Image Toolkit.

6. A lack of management or long-term processes hamstring efforts

The project needs to be managed as you would a long-term software project, not just a quick fix. This means defining clear requirements, going through a full delivery process, including all the stakeholders and refining sprints to make sure what’s been delivered is sufficient.

It’s not something that can be just scripted and left to live on its own. You need to really develop, adopt and embed the best practices for the full software development lifecycle. If the company doesn’t adopt these best practices – which can be hard if it’s not used to working in this way – it won’t work. Continuous improvement is key.

7. Many mistake it as ‘just another development project’

Some of the companies we work with are accustomed to developing and delivering software for their customers, and think they can apply the same principles. But this is often a very different approach. Those projects can be maintained in a totally different way, quite often using the waterfall method, with projects set in stone, extremely sequential and inflexible.

Image lifecycle management solutions need to be constantly refined. As things change, you cannot predict the solution you had in mind at the beginning, will actually work in real life. You very often have to work with the customer, the end user, refine, create new solutions, learn from mistakes,  adapt to cloud providers changing things, and deploy something new.

These are basic agile ways of working, but it’s fundamental to the success of a project like this.

8. Stakeholders end up working in silos

In larger companies, there are multiple stakeholders invested in this kind of solution. Those who manage the different cloud providers, for instance, plus security teams. It can be difficult to ensure collaboration across the board, particularly without a dedicated resource or taskforce.

9. Keeping up with compliance becomes a burden

Many large companies end up spending a lot of money on audits, verifying compliance. In some cases, they’ll take a reactive approach to creating images manually, on a case-by-case basis, to fulfill compliance needs. This isn’t really an efficient way of working.

For ISO and other audits, we can provide companies a baseline image that’s given to users to ensure every security control is implemented in the correct way, the software needed is installed, and other compliance issues are fulfilled. Once an image lifecycle management solution – like IBM Multicloud Machine Image Toolkit – is implemented, it can be shown during an audit, and tick all the checkboxes an auditor has. 

In some cases, companies might have five or more audits a year, and this is a far more efficient way of ensuring compliance and not sweating each time.

Next steps

So, if a company is serious about building an image lifecycle management solution in-house, they need to understand it requires a few dedicated, full-time resources that have the full set of skills for the job. And they need to embrace the continuous improvement principles to keep things compliant.

For some companies, this might be feasible. But if you’d like some help, check out IBM Multicloud Machine Image Toolkit, or get in touch with us to find out more.

CloudCloud

Get in Touch.

Let’s discuss how we can help with your cloud journey. Our experts are standing by to talk about your migration, modernisation, development and skills challenges.

Ilja Summala
Ilja Summala LinkedIn
CTO
Ilja’s passion and tech knowledge help customers transform how they manage infrastructure and develop apps in cloud.