The move to digital business platforms is compelling infrastructure and operations (I&O) leaders to rethink their data center strategies. Digital business platforms like AI, which encompasses machine learning (ML), deep neural networks (DNN) and the IoT, are driving requirements for agile and scalable compute infrastructures.
In 2018 I&O leaders should focus on enabling increased agility and effective ecosystems for emerging digital business initiatives by deploying serverless architectures, container ecosystems and three-tier environments.
These three predictions represent fundamental changes that will impact data center infrastructures through 2020.
By 2020, 30 percent of data centers that fail to effectively apply artificial intelligence to support enterprise business will not be operationally and economically viable.
With the advent of AI and ML, I&O leaders have the opportunity to balance and reduce system complexity and create a new paradigm of “self-organizing systems. Under this model, I&O leaders can expect a broadened and enhanced role for AI, either as a platform or as a service.
I&O leaders who fail to invest in ecosystem and platform intelligence such as artificial intelligence for IT Operations platforms risk becoming irrelevant and ultimately jeopardizing their enterprise’s ability to compete as a business. This is especially true as their skills and tooling fall behind growing operational complexity and the increasing demand for proactive, personal and dynamic services.
By 2020, 90 percent of serverless deployments will occur outside the purview of I&O organizations when supporting general-use patterns.
Since the launch of AWS Lambda — arguably the first serverless computing service — interest in harnessing serverless technologies has exploded among the developer community in leading IT organizations. Serverless computing offers three primary benefits to developers:
- It supports running code without having to operate the infrastructure. This enhances developer productivity and allows them to focus on code development without having to worry about the underlying infrastructure.
- It can enable easier horizontal scaling, due to the auto-scaling properties of the back-end resources. Scalability now becomes a software design issue.
- Public cloud hosted infrastructure as a service (IaaS) serverless frameworks allow for a truly on-demand consumption, because there are no idle resources or orphaned VMs or containers.
Enterprise adoption of serverless computing is still quite nascent due to the immaturity of the technology for general purpose enterprise workloads, as well as the fact that a majority of workloads today are “request driven” rather than “event driven.” However, event driven workloads will grow in importance as next generation front-ends are driven by new technologies.
By 2020, more than 50 percent of enterprises will run mission-critical containerized cloud-native applications in production, up from less than five percent today.
Container adoption has been growing at a viral pace for software development and testing use cases due to containers’ ability to bring environmental parity to the software development life cycle and to enable continuous development and deployment of application software. Understandably, organizations want to extend those benefits into production environments to fully reap the value of the agility and efficiency they have achieved in the development and testing phase. While developers are primarily driving tool adoption around containers, I&O leaders need to be prepared to support these containerized applications in production. Crucially, they must also ensure that business SLAs around security, performance, data persistence and resiliency are met.
As well as improving developer productivity I&O leaders can expect additional benefits from this technology. As they can run on a bare-metal infrastructure, containers can be operated more efficiently than VMs on single tenant server infrastructure. Because of their smaller resource footprint, containers can also enable a much higher tenant density on a host. Containerized applications can be managed more effectively with less configuration drift, as it is possible to more easily redeploy services and automate their lifecycle management.