Although organizations are moving more applications and services into the cloud, certain workloads still require the control, reliability and security of on-premises infrastructure. Hyper-convergence has emerged as a key technology for effectively balancing cloud and on-premises resources.
Hyper-converged systems reduce the cost and complexity of on-premises infrastructure by tightly integrating core infrastructure resources into a single platform running on industry-standard hardware. They also provide the foundation for private cloud platforms that deliver some of the speed and scalability of the public cloud, but with greater security and control.
Private clouds offer a compelling mix of scalability and security, but they are notoriously difficult and expensive to implement correctly. It involves properly integrating virtualization, compute, storage and networking components, along with security, management and orchestration functionality. In 2015, a Gartner survey found that roughly 95 percent of private cloud initiatives failed.
Hyper-converged platforms remove many of those challenges through the pre-integration of compute, storage and networking. Combined with a native hypervisor, these resources enable cloud-like levels of automation, scalability, agility and flexibility — but with lower costs, better application performance and better security.
IDC analysts say that IT departments are increasingly adopting hyper-converged systems for the express purpose of building private clouds. The research firm says that the hyper-converged infrastructure market is growing faster than 150 percent a year, and could be worth $6.4 billion by 2020.
Hyper-convergence has evolved from the converged infrastructures developed in recent years to remedy data center bloat. Converged infrastructures consist of pre-racked and cabled compute, storage and networking components integrated into a unified system based upon a validated reference architecture. This approach shortens deployment time, improves management and delivers one-throat-to-choke support.
There are drawbacks, however. Converged infrastructure solutions are made up of distinct hardware components that have been engineered to work together, which can lead to issues with vendor lock-in. In addition, most converged infrastructure products come in standard form factors with rigid configuration rules, severely limiting provisioning and expansion options.
Hyper-converged systems, in contrast, cannot be separated into their component parts. Software-defined technology delivers both compute and storage functions using the same x86 server resources. There’s just one, highly automated platform to manage. However, a key feature of hyper-convergence is its scale-out architecture, which makes it possible to increase capacity by adding modules. This design approach creates a more agile and efficient IT infrastructure for organizations seeking to implement a private cloud.
In fact, hyper-converged systems share several of the characteristics of the cloud, including:
Traditionally, IT infrastructure has been designed, configured and implemented on a box-by-box basis, creating a complex, siloed environment that is inflexible and difficult to scale. The cloud provides greater agility but is not suitable for all workloads. Hyper-convergence enables organizations to gain cloud-like simplicity without sacrificing the security and reliability of on-premises infrastructure.