What is Hyperscale Computing?

What is Hyperscale Computing?

The term ‘Hyperscale Computing', is used to describe the process of efficiently scaling up distributed computing environments, from a few servers to thousands of them, and the provisioning and facilities required in order for this task to be completed. 

In comparison to conventional computing, the structural design of hyperscale computing can be quite different. Conventional computing environments usually employ high-grade constructs, for example, the ones found in blade systems. These constructs are usually abandoned in Hyperscale computing, in place of a more stripped-down design, which makes it very cost effective. Hence, the funding of a system’s software requirements is far easier, due to the vastly reduced hardware investment that’s required. 


Eliminating Design Elements 

The efficiency of Hyperscale Computing results in the ability to discontinue certain elements that are used in conventional computing: network-connected and locally connected storage, replace superior array storage networks. Virtual LANs replace dedicated computing, management & storage networks. Commodity network elements replace network switching and commodity computing components replace blade systems. Prudently designed applications and other software programs take the place of hardware devices meant for supervision and tracking. Outmoded power supplies are no longer used and efficient hardware configuration substitutes hot-swappable devices intended for high availability. 



The architecture for these systems can be obtained in single unit form, which utilises converged networking, or management used in a modest form factor, or a combination of local storage and network attached storage. 


The Benefits 

Hyperscale computing solutions involve an extremely low financial investment and the system requires minimal configuration. It can run in a private or dedicated system, with a base level of virtual machines. The architecture also works very well in highly upscaled implementations, where 1000’s of virtual machines, are being used. 

High throughput, improved performance and redundancy, intended for high availability and fault tolerance, is achieved through horizontal scalability. Overall, efficient hyperscale architecture, when combined with intelligently designed applications, offers a powerful tool that makes the control of an agile business, far easier and attainable than conventional methods.

Date: 20 September 2018, 10:09 am
Das könnte Sie ebenfalls interessieren: