Choosing between hyperconverged and hyperscale architectures

By Eric Carter | | Hyperconvergence

If you’ve been following our blog for the last few months, you may have seen this primer on software-defined storage (SDS) and its applicability to hyperconverged and hyperscale architectures. It’s essential reading if you’re looking to get a handle on the differences between the two. However, when trying to choose a new storage system, sometimes it’s important to see the similarities and differences side by side. We’ve put together the following table that highlights each storage architecture and what to look for when purchasing your next storage solution.

What is hyperconverged?: A hyperconverged system integrates application compute and storage in one unit designed to scale out by adding more units.

How does hyperconvergence work?: Hyperconvergence typically leverages commodity “whitebox” x86 servers and integrates software-defined compute along with software-defined storage. Compute and storage scale linearly as IT incorporates additional server hosts. Data and workloads are auto-balanced to new nodes added to the system. Hyperconverged systems can be architected using software and a bring-your-own-server approach, or companies can buy solutions that pre-assemble hardware and software under a single SKU.

How does hyperconvergence work?: Hyperconvergence typically leverages commodity “whitebox” x86 servers and integrates software-defined compute along with software-defined storage. Compute and storage scale linearly as IT incorporates additional server hosts. Data and workloads are auto-balanced to new nodes added to the system. Hyperconverged systems can be architected using software and a bring-your-own-server approach, or companies can buy solutions that pre-assemble hardware and software under a single SKU.

How does hyperconvergence work?: Hyperconvergence typically leverages commodity “whitebox” x86 servers and integrates software-defined compute along with software-defined storage. Compute and storage scale linearly as IT incorporates additional server hosts. Data and workloads are auto-balanced to new nodes added to the system. Hyperconverged systems can be architected using software and a bring-your-own-server approach, or companies can buy solutions that pre-assemble hardware and software under a single SKU.

How does hyperconvergence work?: Hyperconvergence typically leverages commodity “whitebox” x86 servers and integrates software-defined compute along with software-defined storage. Compute and storage scale linearly as IT incorporates additional server hosts. Data and workloads are auto-balanced to new nodes added to the system. Hyperconverged systems can be architected using software and a bring-your-own-server approach, or companies can buy solutions that pre-assemble hardware and software under a single SKU.

What is hyperscale?: A hyperscale environment decouples application compute and storage enabling each to scale independently.

How does hyperscale work?: Hyperscale leverages commodity servers and a software-defined approach, scaling the resources needed for application compute and storage separately. As storage needs grow, companies can add servers running SDS software to the storage tier to expand capacity independent of the application compute tier. Data is auto-balanced to new nodes added to the system. Conversely, as compute needs grow, they can also add servers to expand CPU power, independent of the storage tier.

Hyperscale – what’s under the hood?: With hyperscale, scale-out storage software installs on servers – referred to as cluster nodes – that network together to form a storage resource pool. The compute tier, which may include both bare-metal as well as hypervisor-based systems, accesses the storage pool via a software client or proxy – or directly via an API. Each tier scales independently. Hyperscale is the architecture of choice for companies such as Facebook and Google who want the flexibility to meet fluctuating data demands in compute or storage capacity independently, at will.

What are hyperscale use cases?: A hyperscale deployment is also suitable for general purpose server virtualization but because of its unique decoupled nature it also supports non-virtualized applications effectively. Hyperscale excels in supporting large and growing datasets. The architecture is ideal for big data applications and for underpinning OpenStack cloud environments.

What is the hyperscale business advantage?: A hyperscale approach to datacenter and cloud infrastructure offers a high level of elasticity helping organizations rapidly respond to changing application and data storage needs. Hyperscale lowers TCO by taking advantage of low-cost hardware, enabling the flexibility and to scale compute or storage resources as needed, and driving automation to reduce the amount of human interaction required to operate a data center – even at high-scale.

In today’s diverse computing environments across companies of various sizes, both hyperconverged and hyperscale options can be useful. Companies that are modernizing their IT infrastructure with a goal of being more agile and “cloud-like” should investigate whether hyperscale, hyperconverged, or a mixture of both systems provides the right level of support for their storage needs.

At Hedvig, we support the deployment of our software-defined storage solution in both scenarios. This gives data center architects that have fluctuating application needs the ability to mix and match hyperconverged and hyperscale deployments with all systems working together under a single management system. Our goal is to ensure a high degree of flexibility for what you need now and for your future data requirements.

Interested in finding out more? Download our hyperconverged vs. hyperscale infographic that further explains the difference and provides a decision tree to determine which is best for you.

Download