Most businesses choose to run their applications either in the cloud or in a data centre. Sometimes, however, neither of these options are a good fit for the business.
Companies are now recognising the growing cost and restrictions of cloud deployments, while running a data centre in-house brings its own problems, such as the need for reliable connectivity and in-house skills. If you find yourself in this position, there’s a chance that converged infrastructure – a single, condensed and pre-built cloud appliance – could help.
If the term converged infrastructure is new to you, then this guide will give you a practical overview, including its history and the reasons why businesses adopt this approach as part of their wider cloud strategy.
The idea behind Converged Infrastructure is simple. The goal is to package the four core aspects of a data centre: compute, server virtualisation, storage and networking into a single, optimised computing appliance. It’s a flexible IT resource that delivers against defined service levels, with easy management and without breaking your budget.
You can think of converged appliances as a cloud in a box, bringing cloud-like benefits to on-premise deployments.
Instead of having many IT assets sitting in independent silos, converged infrastructure creates pools of servers, storage and networking resources so multiple applications can share them and be managed collectively, creating a private cloud environment that sits within the four walls of your office.
However, even the simplest of ideas can take a good deal of complexity to bring them to fruition. To understand why companies are adopting converged solutions, it’s important to look at what companies were doing before. It’s time for a (short) history lesson.
Back in the early 2000s, as the demand for new applications increased, so did the need for greater compute resource and data storage. Companies deployed servers and storage every time they required a new application. This resulted in silos forming that either supported one type of computing technology or a particular line of business application. Each IT domain – be it application, network or infrastructure – needed to be monitored, analysed and supported in its own silo.
Over time, this led to many different hardware and software configurations being supported throughout a single company. This rapidly became expensive and impractical: as much as two thirds of an IT budget could be spent on maintenance and operations, leaving very little for new projects and innovation.
Fast forward a few years and the introduction of cloud computing did not resolve the problem. The cloud often added to these silos as it became very easy for new instances to be spun up, with little consideration of their impact on the IT department.
At the end of the decade, many companies felt that their infrastructure was spiralling out of control. There needed to be a solution, one that could create a pool of virtualised servers, storage and networking capacity to be shared across workloads and applications.
In 2008, Oracle announced the HP Oracle Database Machine, an all-in-one device that would address this issue. It was soon followed by Cisco’s UCS; HP’s announcement of its Converged Infrastructure Architecture; and the arrival of VBlock – a joint venture between VMware, Cisco and EMC. The first wave of converged infrastructure had arrived.
Over the past decade, the converged approach has been validated with $3 billion spent worldwide on integrated systems, as estimated by IDC. Converged Infrastructure's growth runs in parallel to that of Platform as a Service; both having a place in a company’s IT strategy.
Apart from the fact it’s a single appliance capable that simplifies hardware procurement and can replace the in-house data centre, converged infrastructure brings many benefits to its adopters including:
If there are silos within your data centre, you know they all need separate monitoring and management. By consolidating these silos, you can centralise and simplify the monitoring and management while reducing direct costs and support overheads. You can even have it managed remotely by your cloud provider.
Beyond simply delivering all of the hardware components required to operate a cloud platform, converged solutions can come pre-configured, simplifying deployment for your in-house teams. Not the case with every converged solution, but the level of pre-configuration may mean that an appliance is simply delivered to an office location and goes live once it’s plugged in.
That’s the idea behind our solution, the Converged Cloud Stack: a fresh deployment can go from order to live platform in under a month. It takes care of the build of the hardware, environment and installation of applications, before it’s shipped out and switched on.
Compare this with architecting, ordering, racking, configuring and commissioning a traditional data centre solution, and one month in, you could still be waiting for supplier quotes – let alone deploying a live solution.
When a dedicated resource is deployed for a project, it is rarely fully utilised. If this underutilisation is multiplied across your organisation, then your company will be paying considerable sums for the unused resources. If the resource can be shared, as it is with converged infrastructure, then you can reduce costs.
Standardising what and how you deploy allows you to automate processes, improving efficiency and reducing time to value.
Despite sitting in an office location, if required, converged infrastructure can be managed remotely to the same level as a data centre deployment through a cloud provider.
Adding a machine element is as easy as placing a new order, with no major planning or configurations required. Without a complex setup, you could install a new system in a matter of minutes, ensuring you aren’t tied to just the resource within the appliance.
You can deploy converged infrastructure into locations where building out a data centre is not practical or viable, which is especially useful for short-term requirements. It’s also useful for sites that have poor connectivity, an issue which can rule out the cloud.
The cloud has a place within every business. Some organisations find many of their workloads are ideal to run on platforms such as AWS and Azure. However, there comes a point where the cloud (OPEX) may start to become more expensive than a CAPEX model. If your performance and storage requirements grow, your costs will too, which can cause challenges.
However, we often find that the decision to utilise converged infrastructure boils down to circumstance, for a company in a transitional period or for a specific project. Think about regulatory requirements that insist on the data being held in-country, for example, and converged becomes a compelling solution. Similarly, if a company is moving premises or consolidating their in-house infrastructure, an all-in-one appliance offers the power and agility to assist.
With converged infrastructure connected to your local network, security and IT managers can have confidence that data and applications won’t create a silo of its own while remaining accessible only to those with the appropriate credentials, privileges, and authorisation.
Moving on from our history lesson, and the idea of converged infrastructure has moved forward today, with the advent of hyper-converged infrastructure (HCI). Before we dive into what it is, HCI does not replace converged infrastructure as both contain unique benefits.
Hyper-convergence is the latest step in the pursuit of infrastructure that is flexible and simpler to manage. Converged infrastructure has similar aims but, in most cases, seeks to combine compute, storage, and networking into a single unit and provide a unified management layer. In a hyper-converged infrastructure, the technology is software-defined, so the technology is all integrated into a single device and cannot be broken out into separate building blocks.
Because hyper-converged infrastructure is software-defined, the infrastructure operations are logically separated from the physical hardware. This makes HCI suitable for even more workloads than converged infrastructure. This is because you can change how the infrastructure is configured at the software level and manipulate it to work for workloads that some pre-configured converged bundles can’t support.
As a rule of thumb, hyper-converged infrastructure is normally built with software-defined compute, server virtualisation, storage and networking, and provides a simplified scale-out architecture with commodity servers. When the capacity of one appliance is filled, another is simply added on, doubling the resource available.
Your company’s IT strategy is likely to consider in-house deployments, PaaS and cloud, and each has a role to play. Converged Infrastructure adds an additional dimension by bringing the convenience of the cloud to a private, in-house setting. It can be used where there is a tight deadline, lack of cloud or skills shortage on the ground in a particular location.
The Cloudhelix Converged Cloud Stack provides a private cloud in a box, offering agile delivery of powerful technology in just 30 days. Not just a hardware solution, it’s an end-to-end service including a project wrap through to ongoing managed services.
Delivered in a flight case, it is ready to ship anywhere in the world and we provide a project wrap to get your infrastructure up and running in under 30 days from project kick-off.