Machine learning and cloud are set to continue as ‘hot topics' in 2019. As technology evolves, organisations that don't take the reins in terms of machine learning and cloud adoption may see themselves left behind. You only have to look as far as the Australian market to see numerous initiatives and great projects taking place. However, to be successful, they must have data at the centre of their attention.
According to Gartner, machine learning promises to transform business processes; it will not only reconfigure the workforce, optimise infrastructure behaviour, but also blend industries through rapidly improved decision making and process optimisation.
Moreover, the exciting part is that we are just at the beginning of the enterprise machine learning transformation. Machine learning will increasingly become a core business and analytic component. Its advancements and capabilities will allow organisations to automate pattern detection, prediction and decision making which will drive transformational efficiency improvement, competitive differentiation and growth.
Early adopters of machine learning that have had time to experiment with its capabilities will move from a proof-of-concept stage to production deployment of multiple use-cases. This will allow for an emergence of technologies and best practices aimed at enhancing operation, scale and ultimately achieve full transformational value.
Infrastructure and tooling will continue to evolve around efforts to streamline and automate the process of building and deploying machine learning apps at enterprise scale. In particular, machine learning workload containerisation and Kubernetes orchestration will provide organisations a direct path to efficiently building, deploying and scaling apps in public and private clouds.
We will see continued growth in the automated machine learning (AutoML) tools ecosystem, as vendors capitalise on opportunities to speed up time-consuming, repeatable chunks of the machine learning workflow, from data prep and feature engineering to model lifecycle management.
Streamlining and scaling machine learning workflows from research to production will also drive new requirements for DevOps as well as corporate IT, security and compliance, as data science teams place increasing demands on infrastructure, continuous integration/continuous deployment (CI/CD) pipelines, cross-team collaboration capabilities, and corporate security and compliance to govern hundreds of machine learning models, not just one or five, deployed in production.
Beyond technology, we'll see continuing demand for expert guidance and best practice approaches to scaling organisational strategy, skills and continuous learning in order to achieve the long term goal of embedding machine learning in every business product, process and service.
A recent report by advocacy group StartupAUS in collaboration with Microsoft revealed Australia is facing a critical shortage of coders — including full stack developers, front-end, back-end and mobile — user experience designers and start-up-focused sales roles such as business development managers. Visionary adopters will seek to build an investment portfolio of differentiated machine learning capabilities and optimise their people, skills and technology capabilities to best support it.
In 2019, we expect to see the best practice approaches to scaling organisational strategy alongside continuous learning of machine learning.
Cloud will become significantly more important for organisations in Australia with Gartner predicting 30 percent of organisations will use object storage as a data repository on-premises, bringing cloud architecture to the data center by 2019.
As organisations begin to understand and appreciate the value of adding cloud to their existing infrastructure and applications, a mix of public cloud and on premise cloud will become increasingly important, offering organisations the flexibility and delivering a solution that best fits their needs.
Any vendor that only offers one option and “locks in” a company will find their customers will be at a disadvantage. With this choice of deployment options, the need for a consistent framework that ensures security, governance, and metadata management will become even more crucial.
As a result, companies will be able to simplify the development and deployment of applications, regardless of where data is stored and applications are run, also ensuring that they can use a variety of machine learning and analytic capabilities, working in concert with data from different sources into a single coherent picture, without the associated complexity.
With these benefits in mind, companies will be more likely to move to a hybrid cloud model, enabling them to have workloads and data running in private cloud and/or public cloud based on their needs. Bursting, especially with large amounts of data, is time consuming and not an optimal use of hybrid cloud. Instead, specific use cases such as running transient workloads in the public cloud and persistent workloads in private cloud provide a “best of both worlds” deployment.
Hybrid cloud will also open doors for organisations to utilise a platform as a service (PaaS), an application development platform for developers to write custom applications without provisioning the underlying infrastructure they need to run.
Most major PaaS software can run on customer's premises, hosted in a private environment or natively in the major IaaS public cloud. The PaaS automatically configures infrastructure resources across these environments, making them a platform for hybrid cloud.
Even though this is a market with great potential to grow, the hybrid model is a challenge for public cloud as well as private cloud only vendors. In order to prepare for this move, vendors are making acquisitions such as the recent one by IBM buying Red Hat.
The reality is that in 2019 we'll certainly see more of them, also mergers among vendors to broaden their product offerings for hybrid cloud deployments. It's safe to say the race to drive hybrid cloud is only starting.