By John Bidgood, CTO Systal.
Hybrid cloud. It’s a headline-hitting concept, taking over technology and business publications alike as organisations in multiple sectors are advised to try and achieve the ‘best of both worlds’. But what does it actually mean? How can you best harness it for your organisation?
The concise answer is that hybrid cloud infrastructures combine elements of both public and private cloud environments, in a model that is utterly bespoke to each organisation deploying one.
Public cloud solutions such as Amazon Web Services (AWS) and Microsoft Azure are combined with private cloud assets placed in datacentres or co-location spaces, enabling greater choice and control over exactly where different assets and datasets are located. Typically, the most sensitive and business-critical applications will remain hosted in private cloud environments, where the organisation in question can retain the greatest levels of security and visibility, whilst less critical but perhaps wide-reaching resources are hosted in the public cloud, where costs are potentially lower and scalability is greater.
However, the possibilities for how this blend is actually achieved are nearly endless. Practically any combination of assets and applications can be hosted both publicly and privately, and with a wide range of providers to choose from and combine together, no two hybrid clouds are remotely alike.
If this sounds complicated – well, it is. And to make matters still more complex, hybrid cloud infrastructures also require orchestration technologies to enable workloads to move seamlessly between those disparate public and private platforms. Hybrid cloud infrastructures are, as such, extremely dynamic.
Plan for success
This dynamism means that to effectively reap the potential benefits of hybrid cloud (which we will be covering in subsequent blogs), it is vital for organisations to invest in their network architecture. And this begins before you actually start building your hybrid cloud. You need to assess and plan the network to deliver the necessary capacity, performance, resilience and security for both user and server interconnectivity.
First, you need to plan that public/private balance, considering which assets are best kept on-premise and which can most efficiently be farmed out to a public cloud service. As mentioned above, the security/service balancing act is key here – but so too are the levels of in-house cloud expertise you can support. Remember that managing an on-premise datacentre requires far more effort than managing a public cloud contract, so you need to ensure that you can manage the proposed architecture internally.
Next, you need to consider the peaks and troughs in demand that your network infrastructure needs to support. This is particularly important in certain businesses, for example online retailers who experience massive uplifts in demand around seasonal events such as Christmas and Black Friday. Physical datacentre space that is only used at select times of year is a waste of resource, so the stretchy, scalable public cloud can be really helpful here.
Next you need to consolidate your existing environment as much as possible whilst ensuring the underlying physical infrastructure (the plumbing!) can scale and deliver the service before you begin. There are several different aspects to this. Understanding and mapping all of your existing assets is the first step, and this should be followed by processes of deduplication, rationalisation and virtualisation.
The second step is to introduce Network Function Virtualisation (NFV) as an overlay to the physical network. As NFV is typically configured in software, your assets will maximise their agility and flexibility to scale your network in real-time.
Often this functionality will be deployed on the same physical compute assets as virtual servers and managed via the same Hypervisor software as your virtualised servers. Inversely some virtualised servers might actually be deployed on WAN/LAN network hardware platforms in remote sites.
The aim is to arrive at a point where every single network asset is genuinely required for business operations, is not repeated elsewhere or split between multiple locations, and takes up as little space as possible whilst maximising its re-use across the traditional IT support towers of compute, storage and the network. This will ensure that you have the simplest, most flexible and cost-effective architecture in place.
This is where managed services providers like Systal can be invaluable, helping you to seamlessly rationalise your existing architecture and provide the best possible foundation for a hybrid cloud deployment. We can advise, too, on brokerage, orchestration, service/security implications and portal software to automate precisely the right combination of your cloud processes.
So, if you’re ready to experiment with the best of both worlds, why not contact us today?