1 November 2018

Cloud-initiated digital transformation

Article by New Relic cloud architecture senior director Lee Atchison


When it comes to executing a successful cloud-migration, only organisations willing to fully embrace the agility and flexibility required – and not just use the cloud as a cost-effect way to store and host applications – will fully reap the rewards.


So, what’s holding businesses back from going all-in for the cloud?


Organisations have been reticent about giving up control of their data to a cloud provider and concerned about data security implications. These concerns are now beginning to reduce as organisations realise the benefits more than outweigh the risks which can be controlled and mitigated.


So, the question should no longer be whether businesses should move to the cloud but, instead, how do they maximise the move to the cloud?


Does the cloud make financial sense?


In some cases, using the cloud can be cheaper than owning your own data centres, for instance, while, in other cases, it can be much more expensive. What often goes unsaid is how expensive the cloud migration process is. The first few months can be very costly, depending on the method you use to execute your migration.


Let’s say you’re simply lifting and shifting your existing data centre applications and leaving them unchanged in the cloud, in this case the cloud would probably be more expensive to use than what you’re using now.


However, if your applications make use of cloud specific features that enable capabilities and efficiencies in your application – such as dynamic resource allocation – then the cloud would be the much cheaper option.


Something else to consider is what services you want to use from your cloud service provider. These higher value services provide great capabilities to make your migration process easier, but remember to calculate those costs on top of the standard cloud and data centre costs.


It’s all about data, data, data


Even though security concerns in the cloud are now more controllable, security should still be seriously considered, especially in light of the General Data Protection Regulation (GDPR) that came into force this year.


There’s a great deal of integral cloud cybersecurity that organisations can leverage. Cloud providers invest heavily in security because it’s critical to their business.


They have high-end security teams internally and provide a wide variety of tooling services themselves and via third-party vendors to help their customers build security in their systems as well.


Notwithstanding this it is still important that no matter whether an application is hosted on-premise or in the cloud, it is still up to every software team to build their application in such a way that it makes use of those security systems correctly.


This goes to the core of what GDPR does in imposing restrictions on companies on the way they manage their data. That is true whether or not they are using cloud services or do everything in-house.


The cloud can actually provide some of the tooling that will help companies follow these restrictions and enable better security. But the cloud, by itself, neither solves nor worsens your security needs.


Instead, it gives you a framework for providing a secure application and infrastructure over which you need clearest possible full visibility of the vulnerabilities, risks and application performance


The perfect migration


Unfortunately, there’s no such thing as a migration without challenges. Most companies tend to start with lofty expectations once they’ve made the decision to move to the cloud, but it isn’t necessarily as easy as we’d like it to be.


Businesses then start to blame different departments on any problems that occur during the migration and there is typically a pressure to declare” victory” before a team is ready.


How can a software team or organisation meet their promises to stakeholders? Most companies moving to the cloud struggle with this.


Doing cloud migrations right involves taking actions while you are planning, during the migration itself, and follow up after the migration is complete.


Central to this is instrumenting applications early in the process which will enable you to find problems quickly and help give you a clearer understanding of what is going on with your application during all phases of the migration.


In the planning phase this requires determining the type of metrics you want to measure and understanding why you want to measure them.


These Key Performance Indicators (KPIs) will be important metrics that you need to track during your migration in order to assure your migration is successful. Additionally, you need to create a resource inventory and establish their dependencies, so you understand how your app components work together and you determine the performance expectations of each component.


Having this level of visibility into an application and performance data will help determine which resources are in scope for migration and a reasonable order for migrating resources.


During the migration this will enable you to understand if the cloud resources are actually improving the operational capabilities of the application. And through rigorous acceptance testing you can prove the goals were met or iterate until confirmed successful.


Accelerating the change


Gaining confidence and proving results is critical because the cloud-based infrastructure you are running on is now out of your control. Understanding the impact cloud technologies have on applications and customer experience is essential before taking advantage of more advanced capabilities and services the cloud offers.


As you move your production applications to the cloud you must address many concerns including if the performance impact is what you were expecting.


Dynamic technologies, serverless, and distributed architectures all offer compelling capabilities to innovate and scale, but you need to be able to understand if the performance impacts, positive or negative, are appropriate for the application and will meet your business needs.


It’s also important to consider when advancing with new technologies to identify where it is in the hype curve, because that technology is likely being used for more things than it should be used for.


The danger of going “all-in” on a new service without instrumentation is being left with an application that’s running on a sub-par infrastructure.


Using technology early in its hype cycle is dangerous unless you understand the value you are getting from that technology, and you cannot understand that value without having visibility into how the technology is impacting your application.


Embracing the cloud


Overall, moving to the cloud provides numerous opportunities for those businesses willing to seize the opportunity and are willing to experiment with new ideas which are made possible by the cloud technologies.


While industry trends and market challenges might differ across companies, the constant for businesses is they must keep the customer experience at the centre of their decisions and must show greater flexibility and nimbleness to ever-changing customer expectations.


The cloud offers this path to both venerable institutions as well as to entrepreneurial start-ups.