The Hyper-convergence of cloud computing infrastructure is now a reality. The growing success of OpenStack implementations in public and private clouds is testimony to the trend of Hyper-convergence.
Hyper-convergence simply means modular distributed systems running on commodity hardware that are by design massively scalable, fault tolerant and reliable.
Now a few questions come to the fore. How are Hyper-converged systems different from traditional converged systems? What is fostering Hyper-convergence trend in the age of multi-clouds (public,private and hybrid clouds)? What are the benefits of Hyper-converged systems? How is OpenStack enabling Hyper-convergence?
We will examine all these questions in this post.
Converged systems vs. Hyper-converged systems
Converged systems are a natural extension of traditional IT infrastructure, which previously existed and operated in silos. The converged systems combined two or more of these technologies to provide pre-engineered systems.
The pre-engineered converged systems are separate technology components engineered to work well when combined together and run on specialized or purpose built hardware. Hyper-converged systems on the other hand are distributed modular components designed to scale and run on commodity hardware.
A good example of a converged system provider is VCE. VCE provides best in class storage, compute and network converged system that is tightly coupled via software. In this case compute is VMware VSphere with hardware from Cisco UCS, the network also from Cisco and storage by EMC, hence the VCE acronym for the independent company.
Hyper-convergence takes convergence to the next level and transcends technologies and vendors across the IT infrastructure spectrum, beginning from networks, storage and compute to applications, databases and integration. The table below summarizes the characteristics of converged and hyper-converged systems.
Genesis of Hyper-converged systems
Converged systems resulted on account of technology vendors coming together to engineer extensions and combine technologies of each other products. As a consequence closed fragmented islands of convergence emerged and flourished. As with VCE there are other providers such as HP who offer converged systems.This is seen with network providers, database providers, virtualization , integration and in other technology domains.
These carefully stitched technology partnerships carved out by vendors in different market segments, are now under pressure or in some cases not even relevant any more. This is caused partly on account of the disruptions happening in their own backyards, as well as maturing of open-source initiatives like OpenStack that have provided robust distributed computing platform for software defined innovation.Further virtualization effects ushered by software defined storage (SDS) and software defined networks(SDN) have propelled Hyper-convergence to center stage of cloud computing.
OpenStack and Hyper-converged systems
OpenStack in its most simplest description is the new operating system of the cloud. OpenStack is the distributed software platform that ties and orchestrates storage, compute and networks on heterogeneous commodity hardware. The growing implementations of public, private and hybrid clouds are built on OpenStack platform.
The growing support from technology vendors and fast evolving OpenStack ecosystem is bringing innovations across networking, compute and storage domains, this is propelling hyper-convergence to the fore. OpenStack has enabled new public, private and hybrid clouds to be built on commodity hardware, thus eliminating the need for traditional converged systems that are complex and expensive. More importantly OpenStack eliminates lock-in with a specific converged system provider, making way for cost effective, distributed, massively scalable public, private and hybrid clouds.
In the growing list of enterprises adopting clouds, OpenStack has become the cloud platform of choice for applications supporting a variety of workloads from transactions to analytical workloads. To name a few, Apache foundation’s Hadoop, Spark for large analytical workloads run on OpenStack platform. It is the relative ease, economical and elasticity of distributed software offered by OpenStack, which is appealing for enterprises to run any type of workloads, without the need to buy expensive converged systems.
Clearly, Hyper-convergence brings to enterprises reduced complexity, simplified architecture of hyper-converged systems, with no vendor lock-in to give IT cloud computing, flexibility and agility to scale on economical commodity hardware. As a consequence, Hyper-convergence delivered with OpenStack gives enterprises the power to choose public, private or hybrid cloud to flexibly and economically run enterprise workloads.
Therefore it follows that IT strategies and IT organizations in the age of multi-clouds are well served, by incorporating Hyper-convergence with Open stack as part of enterprise IT’s core foundation.