By Anthea Nadin
Traditionally, enterprise IT infrastructures were designed around Storage Area Networks (SAN) and Network-attached storage (NAS) on dedicated networking equipment, using separate, dedicated Ethernet switches and large scale-up compute systems to provide network connectivity. These systems were difficult to manage and required an onsite specialist to oversee the systems.
Over the last decade, technological advances such as hypervisors and the cloud have changed the way the data centre is built and operates, and today, we have composable infrastructure which can run any application and deploy resources extremely fast.
Virtualisation was the first big game changer, allowing for the abstraction of compute resources from the underlying hardware, bringing consolidation, efficiency, flexibility and availability improvements to the data centre. As virtualisation evolved, the rest of the infrastructure evolved, introducing Converged and Hyper-Converged offerings. By leveraging software-defined technologies like Software-Defined Storage (SDS) and Software-Defined Networking (SDN), IT is slowly embracing and building the software-defined data centre vision by abstracting the logical data centre resources virtually from the underlying DC hardware.
Despite the evolution of traditional IT workloads and the introduction of more fluid next-generation services, many IT administrators still have to manually provision virtual machines, and still have some workloads on bare metal infrastructure. This has created an environment where they need to dynamically respond to changing requirements while having to manually perform routine, repeatable tasks to manage the existing infrastructure.
Speed, agility and efficiency – guaranteed
Composable infrastructure is the answer to these challenges, using automation based on Artificial Intelligence to offload some daily tasks and reduce the risk associated with human intervention. Composable infrastructure uses code to drive a software-defined environment and perform the deployment, configuration, adjustment and retirement of compute, network, storage, OS and application resources (composing resources to service workloads). This helps IT operate like a cloud provider to lines of business, maximising the speed, agility, and efficiency of data centre infrastructure and operations to meet SLAs consistentl,y while providing predictable performance to support all workloads.
Three specific technology architectures are used to create a composable infrastructure: Fluid pools of infrastructure resources, which include computing, storage, and fabric resources; intelligent software; and a single, unified API. For an architecture to be truly composable, each of these three layers must be involved in the execution and operation of the infrastructure.
These allow resources to be individually changed, aggregated, disaggregated, and composed, based on an application’s precise needs. The beauty of composable infrastructure is that it provides a single scalable environment which can span across one, some or all of the following use cases: Virtualisation, containers, bare metal, as well as private or hybrid cloud, DevOps, enterprise VDI, and data analytics.
What you need, when you need it
Essentially, composable infrastructure ties the promise of the cloud together with the reality of today’s hybrid environments. While most companies are using numerous applications and services in the cloud, a great deal of their infrastructure remains on-premises. Composable infrastructure bridges the gap, providing the ability to execute both traditional and cloud-style agile operations from a single infrastructure.
Based on the report “Hybrid IT Strategy Insights: How Composable Infrastructure Can Fuel Breakthrough Business Innovation And Performance” released by Forrester in February 2018, Two-thirds of IT decision makers ended up with hybrid-IT by accident. As a result, they weren’t able to reap the benefits they expected as they had not intentionally set up their hybrid environments in a common and compatible way to integrate their on-premises portions optimally with what they had available to them in the public cloud. Many had also not updated business governance practices, opening them up to further complications.
These challenges are easily avoided by approaching experienced partners in the design stages, with the intention of planning and building this hybrid environment from the ground up. Building workflows requires a high level of organisational maturity, and may not be compatible with all applications. Although composability simplifies IT operations and management, it needs proper engagement with experienced consultants to understand the specific requirements and use cases and then to appropriately size, design and deploy the infrastructure, which will form the foundation and offer the flexibility required in the future.
Watching the evolution of IT
As businesses evolve and grow beyond hyperconvergence, composable infrastructure helps IT cut costs, increase storage for all workloads, and improve networking, all while accelerating and simplifying all elements of a company’s infrastructure. Leading the field is HPE, which is evolving its solutions alongside business needs. The latest generation of HPE composable infrastructure solutions have expanded their foundational elements of fluid resources pools, software-defined intelligence, and unified API, extending their value proposition beyond their initial composable modular frames to include a wider choice of workload-optimised form-factors which now includes discrete rackmount servers.
In fact, HPE Synergy was the first platform built from the ground up for composable infrastructure, and remains the flagship composable solution. Synergy provides composability for any mid- to large- scale enterprise that prefers a modular form factor and operational model. HPE Composable Rack, on the other hand, provides a composable solution for customers who have a preference for deploying discrete rackmount servers and want to simplify and automate compute, SDS, HCI and software-defined fabric at rack-row scale. Optionally, HPE GreenLake provides both these solutions. HPE GreenLake is delivered as a managed service on-premises, with a pay-per-use consumption model.
These solutions have been designed from the ground up for new deployments, but there are also solutions for many existing rackmount deployments. For example, HPE Composable Fabric and/or HPE OneView offer composability to existing environments, with no need to rip and replace.