How can you change your IT culture and transform it to deliver services that usually takes weeks in a matter of hours? PwC’s head of cloud services Robert Nikolouzos explores whether businesses can deliver secure infrastructure and consistency at scale, while freeing up time for higher value activities, drawing on his experiences at PwC.
A few months ago I attended a session on cloud, automation and orchestration. Throughout the session I found myself mesmerised by the speaker and nodding my head at everything he was saying.
He started by talking about the benefits of cloud and then moved on to the following story: a company decided to move their servers from their local data centre to the cloud. The process of delivering a service into production used to take 15 days in their data centre. Now, in the modernised cloud it only took them… 15 days. A very disappointing outcome for such a huge investment in their strategy.
What went wrong? It soon became apparent that automation, and orchestration – the workflow of automated tasks grouped together to deliver a service without manual intervention – was not part of their strategy. Cloud was only part of the solution. By moving their focus to automation, changing their processes and introducing an orchestrator, the same environment was ready to be spun up and this time it took less than three hours. Now that is productivity.
I looked around me and most people were amazed, but some commented that this was too hard to accomplish. It seemed like IT nirvana. At PwC, however, we had already achieved this outcome and reached the phase of automating and orchestrating the management of operational tasks.
So how did this ‘too hard to accomplish’ task take place? Although this is not the only approach to a successful cloud transformation*, it’s an approach that worked for a large enterprise like PwC, so I’m confident this modernisation strategy can be adopted by any kind of organisation.
Although, as stated earlier, while cloud placement is only part of the solution to IT modernisation, it’s an integral part of the transformation. It will lay out the foundation to enable the use of different types of cloud services and it will allow you to look at IT infrastructure from a different perspective.
Here are some guidelines to successful cloud adoption and IT transformation:
- Architect for the future
Architect your solution with enterprise-grade security in mind. Encryption, audit, logging, firewalls and dedicated links are just an example. Don’t rush to add workloads manually but take time to automate every server build. The last thing you want is to create technical debt and start remediation activities just a few months later. This is where you need to make a decision. Are you choosing a multi cloud strategy, or are you definitely going with one cloud provider? This decision will help determine the choice of your software tools.
- Configuration management is key
Choose a configuration management system from day one. A configuration management system is used to automate the delivery and operation of infrastructure services. This is going to be the foundation of automation so focus on upskilling your current teams to treat your infrastructure as code. Don’t forget to budget for training when introducing a new tool.
- The right orchestrator
Choose an IT orchestration tool and create a self service strategy. The target audience is imperative in your decision. You will probably choose a different orchestrator if your users are the IT department and a different one if you’re planning to open it up to the rest of the business. Again, as with configuration management, upskilling is key, so invest in training.
- The cloud/automation team
Create a cloud/automation team, even if it’s a small one. Automation needs time to get momentum and realised benefits and most likely, your current teams are fully utilised and can’t take on building out the required capability and upskill, while keeping the lights on and delivering the current workload. It’s important to note that this team should not be isolated from any other team and should have knowledge of all current processes, especially change management.
- Assess your monitoring solution
Assess your current monitoring solution and choose a tool or set of tools that enable you to inject new services automatically with zero manual intervention. Automation methods could be API calls, remote commands and scripting and are just examples of what a combination toolset could look like.
- Assess your backup solution
You most likely have a backup solution in place. What is the capability of your current tool? Does it support API calls or remote commands? Can you create rules that will pick up new servers and understand when a deletion takes place? Does it have the ability to backup cloud native services? What is the level of complexity? After having answered those questions, you can now assess your solution against cloud native provider backups and create a backup strategy.
- Automated billing
Choose a cloud finance tool and design a robust tagging solution. Tags will help you define and identify resources in the cloud. I can’t stress enough the importance of reporting, knowing the indirect costs and avoiding untagged resources, especially if you’re in the complex world of recharging and cost reallocation. Things are only going to get more complicated in the public cloud so make sure you understand the power of API calls and choose your tool accordingly.
- Cost optimisation
Reserved instances, rightsizing and automated shutdown policies are going to be a hot topic once you start adding workloads to the cloud. Get on the front foot. Work close with your finance department, change management and leadership to introduce new processes.
- Embrace failure
Last but not least, promote a blameless culture. There’s massive change needed to apply all the above and change the IT direction. Your people should feel safe and encouraged to try new things, innovate, embrace failure and learn from it.
Moving to the cloud has clear benefits, can be cost effective and can help speed up the delivery of services, but for a successful transformation, automation and orchestration are critical components that cannot be overlooked. At the end of the day, cloud is not enough.
*The DevOps culture is not addressed in this article, not because it’s not the right path to a successful transformation, but because it’s too big a topic and would not get the justice it deserves.