The world of IT infrastructure is being turned on its head with the rise of converged and hyper-converged infrastructure. Whilst the industry buzz is inescapable, it’s difficult to really know what’s right for you and which direction, if any, you should head in. We hope this article provides useful insight into the attributes of both strategies, enabling you to be better informed as you weigh up where to take your infrastructure next.
To provide some context here, data centres are composed of a vast mix of hardware, management platforms and vendors. Each is really a legacy stack of technology that is typically over-provisioned in terms of CPU, memory and storage to cater for the resident applications’ intermittent peak workloads. Whilst virtualisation has offered some redemption, the efficiency gains are isolated to the server infrastructure only. The result is silos of overprovisioned, under-utilised resources which are expensive to maintain, complex to manage, eat up real estate and consume disproportionate amounts of energy.
Converged infrastructure was the first response by the IT industry to solving this dilemma. These solutions are essentially building blocks of discrete components including compute, storage, virtualisation and networking. Each is closely integrated to work more harmoniously as one. As building blocks, each physical component still functions as intended for its primary purpose and can be broken out as such. Typically, converged infrastructures live in the same chassis and as demands on the infrastructure increase, components can be scaled accordingly i.e. more storage to meet growing retention demands. Generally, parts of the solution are pre-configured with resource ratios that optimise the infrastructure and create greater resource utilisation than legacy estates. Depending on the age of your existing infrastructure, there is sometimes scope to re-purpose hardware for deployment in a new converged design. If not, the danger is that choosing to converge creates another island of resource to be managed even if only in the short-term. Inevitably, part of the attraction of converging is cost – not least because buying the individual parts of most converged solutions will be significantly higher than buying as an integrated solution.
A step further
By comparison, hyper-converged infrastructure takes convergence a step further. In this infrastructure, everything is software-defined. There is no scope to break out server, storage or network – they co-exist together as a seamless pool of resource. Hyper-convergence delivers simplification and substantial cost savings by consolidating all resources into a single, dynamic and highly resilient infrastructure stack that expands and contracts as needed. Importantly, the architecture underpinning all of this is radically different to legacy and even converged architectures, with the sole intention of making infrastructure management simple and decluttering the datacentre. It goes far beyond servers, storage and networking however, bringing together other services like deduplication, cloud gateways and solid-state arrays into one solution. There’s a host of competing solutions in this space, including VMware EVO:RAIL, Nutanix and Simplivity, to name but a few.
For some, the thought of undermining existing investments in hardware infrastructure may be too much to swallow. That’s why we’ve been impressed by Atlantis, who offer a software defined storage system that converts direct attached storage into a pooled array – essentially, a SAN without a SAN. By increasing the number of VM’s that can share the storage, it effectively creates a hyper-converged infrastructure. If in time this proves a successful approach, organisations can grow using the Atlantis hyper-scale appliance integrating local storage, servers and virtualisation into a single platform, built on your chosen hardware brand.
It can feel like a minefield figuring out how to navigate through the numerous options available to shape the future of your infrastructure. As a pioneer in the world of virtualisation, we’re accustomed to working with all of the technologies illustrated through this document and how they can best be deployed to help you meet the goals of your organisation.
Read the full article here: First-time converging
If you’d like to know more, are planning a project or are “in-flight” and need some additional expertise, then please get in touch.