The term “hyperconvergence” is a relatively new term to the IT scene. It is an industry that went from basically nonexistent in 2011, to a predicted $5 billion, or 24% of the market by 2019. The idea of hyperconvergence combines compute (virtualization), network, and storage into commodity boxes that allow for a scale-out design.
Behind the Drive
Some of the major forces behind the hyperconvergence drive is the need to quickly spin up more virtual machines for distributed applications, the adoption of agile software development, or other scenarios where hyperconvergence makes better business sense over traditional infrastructure, such as remote office/branch office.
How is this different than its cousin, converged systems? Converged systems aim to combine compute, storage, and server into a single SKU. Converged systems themselves have gone through evolutions throughout the years, offering best-of-breed services like data protection, performance, and global management. A hyperconverged infrastructure builds on top of these ideas, while utilizing software-defined storage instead of a traditional SAN infrastructure.
Management is almost always built into native management tools like the vSphere Web Client, leaving behind separate storage management software. It is true management convergence of your datacenter: virtual compute, virtual networking, and virtual storage.
Building off previous technologies, software-defined storage offers capabilities that are shared with their ‘legacy’ counterparts, but also new features. Software-defined storage often contains native features like fast cloning, deduplication, compression, encryption, and ‘tiered’ or fast storage. Some hyperconverged systems also include native backups and replication. These features differ in availability, capability, and speed, depending on the vendor.
The software-defined storage of hyperconvergence combines commodity storage, usually either a combination of spindle and flash, or all flash, into a pool of storage usable by all hosts cooperating in the hyperconverged system. It still offers the benefits of a storage area network such as redundancy and accessibility, without the separate, sometimes expensive requirements of a separate storage network (dedicated iSCSI or fibrechannel switches), clustered controllers, and disk shelves.
Use Case: VDI
One use case seeing adoption is a virtual desktop infrastructure. Because these systems are hosting many copies of the same operating system, the often-native features of deduplication and compression can offer huge storage savings over a traditional storage solution. Hyperconverged systems offer a scale-out design, making for easy expansion when the time arises. When adding additional nodes, you’re adding compute, network, and storage resources, so you don’t run into complications of scaling traditional environments; adding only compute or storage may leave you with a bottleneck or constraint on the other.
Use Case: DevOps
In a development-centric environment, a new term has arisen. DevOps was born from the idea of “agile system administration” or “agile operations.”
In a development-central environment, there is often the need to quickly allocate or free resources, which can be a complex and time-consuming task. The software-defined datacenter (compute, network, and storage) can be manipulated through an Application Program Interface (API) to automatically and efficiently perform these tasks. This can save time, speed up development, and deliver real value to an organization.
The agility of a software-defined datacenter, with the benefits of virtualized compute, networking, and storage, can boast lower CAPEX, OPEX, and TCO making this a real competitor in the marketspace. Hyperconverged infrastructure has many use cases, such as replacing traditional datacenter infrastructure, remote office/branch office, and virtual desktop infrastructures, and will be here to stay for the foreseeable future.