Cloud migration and data – can you avoid the wrong path?

null
Image Credit: Shutterstock

If you’re not digital first now then you’ve already lost. At least, that’s the message that CIOs and CTOs are being fed now and it’s driving many down the cloud migration track at breakneck speed. In Accenture’s latest technology report, 94% of the 6,600 business and IT executives that were surveyed said the pace of technology innovation in their organisations had either accelerated or significantly accelerated over the past three years.

Such is the prevalence of high profile case studies of digital transformation projects, it’s a common misconception that these enterprises have moved entirely to the cloud or achieved this feat overnight. This is untrue. Capital One, for example, is a US bank known for riding the bleeding edge of innovation and adopting new technology early. The bank, however, would openly view cloud migration as a journey, which requires a broad effort to be successful at such a scale and for Capital One, this began back in 2013. On this journey, Capital One has reduced the number of datacentres it has from eight in 2014 down to three in 2018. While the migration process is complete, the bank continues to optimise refactoring traditional platforms to use serverless and even has a dedicated agile team to enable software engineers with the process. 

The reality is that responsible innovation often takes time and cloud migration at scale is complex. Enterprises can’t always justify the expense of re-architecting every single application that they have, they are more likely to deploy a mix of on-premise, SaaS and public cloud infrastructure, all running alongside each other and creating data on their performance. 

Over time, the mix of IT will move from on-premise to hybrid cloud and public cloud deployments, and those deployments will become more complex. For example, our own research has found that the number of companies running multi-cloud has doubled in the past twelve months from 2017 to 2018.

Converting pressure into progress

Moving to digital is seen as the price of admission for doing business today and being competitive in your chosen market. Migrating to the cloud has to be viewed as part of an organisation-wide transformation that will take place over time. Alongside the technology change, this level of transformation is a people issue that requires more collaboration between cross-functional teams. It involves changing the mindset of your employees and helping them to upskill. Structures will need to adapt. 

For CIOs and CTOs considering what technologies they need to adopt as they pursue a cloud-first approach, it can help to see what others are experiencing in their migrations. We regularly analyse technology adoption among enterprises that run massive applications, and how they are making use of public cloud services like AWS, Azure and Google Cloud Platform.

Based on the latest data for the beginning of March, 68.3% of users continue to adopt AWS alone. Most companies have never switched cloud providers, but a multi-cloud approach is emerging. This is sitting at 10.5% of organisations currently, which has doubled in the past twelve months. Behind this growth is a bid to avoid vendor lock-in and make the most of different cloud services. 

Image Credit: Pixabay

Image Credit: Pixabay

(Image: © Image Credit: Pixabay)

Whether you are in the midst of a cloud migration, or looking at multi-cloud, it’s important to consider the overall journey rather than seeing cloud as a destination. Approaching multi-cloud isn’t often a case of “lifting and shifting” the whole company from one provider to another, it’s a question of selecting where you are going to achieve the most impact. Multi-cloud isn’t going to be a one size fits all solution.

One approach is to create a “spear tip” team to implement any new technologies that are being introduced and make sure they can deliver success. Whether you are looking at new projects on one cloud — or at running a new implementation on a different public cloud platform — this should help keep the project agile enough to work through initial findings and learn from them while day-to-day operations carry on alongside. From these initial projects, companies can use these lessons to educate other teams on the potential opportunities. By continuously improving your approach over time, you can continue to make the most of your experiences.

Without this approach to cloud migration, it may not yield the results you expect. According to Computing and KPMG, cloud migration outcomes don’t always match the expectations for some organisations. Enhanced business agility was the highest expected business critical benefit of cloud migration at 63%, but the reality is that it delivered less than expected at 51%. 

Understandably, cost was the key focus for KPMG’s research, but it pulled out a notable statistic: only 15% of the organisations were auditing their cloud workloads continuously, which led to only 23% believing that they knew where most or all of their cloud spend went. This highlights that legacy analytics tools cannot deliver the visibility needed to support the investment customers are making in modern architectures at cloud scale.

To anyone considering cloud migration, the KPMG results might be a little dispiriting but they indicate that planning the correct path for cloud migration, whether full or partial, is vital for success. To achieve this, you need data on what you are using, how well you are using it, and you have to be able to compare how that spend resulted in value for the organisation.

Spear tip approach

Making the move to becoming a cloud-first business is not just a technological change, but a cultural one. It’s a move from a top-down, command and control environment to a more agile, team-driven structure that uses data to lead decisions. Most companies would be advised to take a ‘tip of the spear’ approach with a small, innovative DevOps team to provide a proof-of-concept for the rest of the organisation.

There are some common approaches to the cloud migration processes that can be considered. They each have their own merits, and it’s important to pick the right one based on your current needs and experience.

Where a flagship product will need to be re-architected for a cloud-native approach and segmented into microservices, a less critical application may only need re-platforming, new database infrastructure and basic changes to be ready for cloud use. In contrast, a ‘retain and retire’ approach will apply to low revenue-generating and end-of-life applications, where an on-premise solution will make the most financial sense.

The first is based on lifting existing applications as they are into the cloud. This means carrying out a thorough initial assessment, where technology assets are catalogued along with application code and business processes. Once this is complete, the right cloud implementation is then started.

The second approach focuses on the destination and adopting more cloud-native technologies, rather than a straight migration. This involves looking at your application, what service it provides, and how to deliver the same result. Working backwards in this way can help you redesign the application to take advantage of microservices and containers. 

Image Credit: Pixabay

Image Credit: Pixabay

In order to decide which of these approaches best suits you, it’s worth looking at your existing data. Machine data created by applications as they run can provide valuable insight into what is taking place every day. This treasure trove of insights can help you choose the right path and avoid issues further down the line in any cloud migration. Machine data will, for instance, reveal a comprehensive inventory of applications, their connections to each other and any cross-independencies they have, along with their servers and supporting technologies. It also provides accurate data transfer volumes, insight into performance challenges and shows any hidden latencies.  

Establishing a machine data-driven baseline before a migration takes place is a vital way to protect your enterprise as it will supply meaningful metrics and thresholds that you can evaluate in the current on-premise configuration. Once any initial migration is completed, you can assess again and see where you can further optimise your approach. Using this data will enable you to establish new goals and compare performance over time.

Similarly, machine data can prove very useful for teams that are involved in supporting ongoing business processes or security. Getting a continuous stream of data back from your operations can support security teams in seeing that nothing untoward is happening, while the same data can be re-cut and contextualised for business teams to see how they are performing too. 

Image Credit: Shutterstock

Image Credit: Shutterstock

(Image: © Image Credit: Alexskopje / Shutterstock)

Interpret your data to take the right approach

The key to success in cloud migration is to be realistic and focus on moving the company forward. The aim should be to squeeze as much value from your cloud migration as possible and continue to build on smaller teams that are empowered to innovate and take an entrepreneurial approach. Establishing continuous intelligence from the outset that provides real-time, visible insights into operations, security, customer experience and business results will be critical to any migration project and inform future business decisions. 

The big success stories are out there. Channel 4, the public service broadcaster, has embraced machine data analytics for DevOps modernisation initiatives across its full application stack and cloud infrastructure to elevate its on-demand viewing experience. Other good examples of innovative companies include Pitney Bowes, USA TODAY and Discovery Networks, where they have adopted new technologies to power their digital transformation projects. Making use of data can inform your cloud migration as it goes on, helping you take advantage of new options when they are suitable and when they can make a difference.

After all, it’s important to remember that Capital One’s cloud deployment wasn’t built in a day.

Colin Fernandes, EMEA Product Marketing Director at Sumo Logic