Moving to the Cloud, What's the Issue?
Five essential steps to successfully manage your company's cloud adoption
Five essential steps to successfully manage your company's cloud adoption
This might sound familiar: An organization’s CEO makes a top-down investor-led mandate for more cloud. In response, the CTO proposes moving thousands of internal applications to the cloud within a time frame of two to five years. A couple of years later however, only a handful of token applications linger in the cloud. Yet the investors barely know the difference – they seem happy enough that the CEO even mentioned the word cloud in their quarterly report.
Such a common narrative highlights how easy it is to fake the world into thinking you’re doing cloud. After all, and let’s be honest here: that’s what many of your competitors are doing right now.
But you and I know that’s not good enough. Nominal progress in the cloud doesn’t fundamentally transform IT. Deploying a small number of digital applications, using open source technologies on a public cloud that are still being supported by legacy systems, isn’t a transformation. You still have the same problems, you’re just using different solutions.
So, how can you bring true cloud improvements to enterprise IT today that will make a real difference to your bottom line? Here are five suggestions we have seen in the industry that have more meat on the bones than what many of your competitors have been doing.
The first important step is to look at your business processes and see where manual activities are still required. Such processes normally span across several departments and applications, and while some might already be automated, this will likely only account for a small percentage.
Orchestrating across these different applications is the first crucial activity for an enterprise looking to become agile and enable itself for cloud-adoption. If you take an application and move it to the cloud without orchestration of the overall business process, you will not gain any competitive advantage, as you are still performing the same activities – it just happens to be on another cloud-based application.
Furthermore, when it comes to IT, processes around spinning up new environments for testing purposes might take weeks in a larger enterprise. If you orchestrate this process, (which includes IP address management, service management systems, storage systems, firewalls, hypervisors and so on), you can instead provision new environments within minutes. This can then be further improved by moving different elements to the cloud.
Orchestrating the pipeline is a good start and lays the groundwork, but today we see most cutting-edge businesses moving beyond this to full self-service portals for the development and deployment of applications across the enterprise.
Having everything available as self-service and guaranteeing that the services are delivered in a fully automated and orchestrated way, decreases the idle time for any business end users. They can be sure that if they need special service – such as a certain infrastructure for performance testing – it can be delivered within minutes, so that the test can be run almost immediately. Cases that require a version of an application with specific underlying infrastructure, can also be made available via self-services if orchestrated properly first.
Integrating self-service portals unlocks a magnitude of improvements on top of just orchestrating the current pipeline. Indeed, the biggest gains in speeding up the application lifecycle can be found here.
Enterprises that are really utilizing cloud often do so from multiple vendors. For developers, it doesn’t matter if a compute resource runs on Amazon's AWS cloud, the Microsoft Azure cloud or somewhere else. An end-user just needs the resource, and does not care where it runs.
However, from an enterprise perspective this can become a huge headache very quickly. Each cloud vendor has different expense models, prices and performance. Each one also has different security profiles and varying levels of maturity in the tools used to manage the infrastructure.
A proper orchestration layer enables you to leverage different providers on a plug and play basis. It offers integration to the different providers, while still orchestrating the business process. This means that if you decide to switch from one vendor to another, you just have to exchange a single part of the overall process, and the orchestration layer will still make sure that the business users are unaffected.
This enables you to leverage any cloud provider of your choice, but avoid vendor lock-in. An enterprise will therefore always be enabled to exchange providers whenever another offers better service level agreements.
As we dig deeper into the activities of the most advanced cloud adoption, we start to see the emerging trend of algorithmic IT decision making. The best analogy here is that humans have been driving cars for a hundred years, but we are now finally at the point where computers can make their own judgement and take over the steering, acceleration and geer changes as required. Similar advancements inside of cloud are allowing more and more decisions to be made for IT as well.
The next step is letting the orchestration layer decide on which cloud platform certain workloads should run. You can model rules that decide based on different criteria the activities each cloud provider should be utilized for. Different rule criteria could apply here – one could of course be the cost of certain resources, but others might include the location of data-centers and network connection, the expected runtime of the workload or any number of other factors.
Although this is still an emerging trend, it is one you should take very seriously, as it is one of the most likely to change the entire landscape of IT over the next 5-10 years. But it can’t be accomplished without orchestration. It is only in the orchestration layer that you can evaluate rules based on the complete business process running across multiple providers. The rules, which eventually will be written by algorithms themselves, will define the best options for you.
You see the results of companies that don’t evolve every day: Toys R Us have gone bankrupt, while Sears, Kmart, JC Penney’s and Macy’s are all closing hundreds of stores per year. Sure these are retailers, but don’t think for a moment that this is the only sector to get hit. That’s like the movie industry thinking that the music industry revolution was a fluke.
Companies that pay lip service to the cloud will end up like Toys R Us. Companies that fully embrace the changes happening in the cloud today will be the market leaders tomorrow.
By implementing the steps mentioned above, you can transform your enterprise. You can become truly cloud agnostic, leveraging any number of providers depending on what is best for the current business processes. The orchestration layer is crucial to abstract the underlying technologies. Ultimately, it enables you to plug into any cloud provider, while still ensuring that the rest of the process stays the same.
Furthermore, a cloud orchestration layer allows you to embrace the biggest change facing IT today: algorithmic decision making. This emerging AI-driven rule-based decision-making framework will enable enterprises to stay ahead of the technology curve instead of constantly falling behind. It will allow you to always make the best decisions, whether it be picking the most appropriate cloud for a given problem or avoiding vendor lock-in.
Could your approach to workload automation be holding you back? As the cloud and Big Data become more prevalent, EMA reports on the benefits the right WLA solution offers.