The innovations from hyperscale infrastructure providers like Amazon, Microsoft, Google, and IBM are going to impact enterprise data center. This trend will enable large enterprises to take advantages of the efficiency and performance achieved by cloud providers. With this trend, the line between on-premises infrastructure and public cloud will get blurred.
Here are five trends of modern infrastructure to watch out for in 2019:
The rise of containers & Kubernetes in the enterprise data center
Enterprises are embracing containers at a much faster pace than expected. The thriving open source community combined with the availability of mature commercial offerings has given the confidence to enterprise customers.
Kubernetes has become the front and center of enterprise container platform. From traditional OS vendors to modern PaaS providers, every major platform vendor has a commercial Kubernetes offering making it the new operating system of the data center.
In 2019, expect to see a meteoric rise of Kubernetes in the enterprise. Containers will redefine the hybrid cloud by bridging the gap between legacy and modern applications and on-premises and public cloud infrastructure.
Google has announced the preview of on-prem GKE that comes with tight integration of its proven Containers as a Service (CaaS) in the public cloud – Google Kubernetes Engine. Microsoft is moving fast in bringing Azure Kubernetes Service to Azure Stack private cloud platform. IBM is betting on IBM Cloud Private – a hybrid cloud platform based on Kubernetes.
Traditional PaaS has gradually transformed into a container management platform. PaaS industry leaders – Red Hat and Pivotal – have embraced Kubernetes as the foundation of their platforms. 2019 will witness the complete transformation of PaaS into CaaS.
The availability of specialized hardware for niche workloads
During the initial days of computing, the software was written for common hardware architecture based on Intel X86 or SPARC. Eventually, Intel became the lowest common denominator for OS vendors and application vendors.
Public cloud has changed the dynamics of hardware and CPU market. To drive the scale of economy and efficiency, hyperscale providers like Amazon, Google, Microsoft, and Facebook started to build custom hardware and processors that complement Intel CPUs.
With niche workloads such as machine learning and high-performance computing becoming the key drivers of public cloud adoption, hyperscale providers are turning to custom hardware to accelerate the performance of the applications.
In 2019, we will see the trend where custom chips will replace generic software. The heavy lifting done by software written for virtualization, graphics, and HPC will be moved to purpose-built hardware that delivers massive performance and efficiency.
Field programmable gate arrays (FPGA) used in Microsoft’s Project Brainwave and proprietary accelerators such as Google Cloud TPU will become more prominent in 2019. They may even become available to power enterprise data centers.
Multi-cloud will no more be a buzzword
Enterprises are investing in multiple cloud platforms to reduce the risk of vendor lock-in and single point of failure.
Large enterprise customers are using at least two cloud providers when running mission-critical workloads. The common pattern is the use of object storage for maintaining an additional layer of redundancy. Deploying the DR site in a different cloud platform is also a common use case of multi-cloud. But these patterns don't deliver the benefits of true multi-cloud adoption.
In 2019, we will witness a trend where customers will federate multiple cloud platforms to run a distributed workload. They may run a legacy transactional database in one cloud platform while running the data warehouse in the other.
