When it comes to the cloud, there’s a million and one options, configurations and solutions out there. Once you’ve decided whether your current project or workload requires public, private or dedicated infrastructure (in this post, we’re going to focus on private), you then have to think about the implementation and approach.
For the sake of keeping this blog to a reasonable length, we’ll focus on the private cloud because, if you’re working within a corporate IT function considering a DIY cloud build, it’s primarily going to be utilised by your business, therefore it falls into the territory of the private or dedicated cloud. It’s not the public cloud because, quite simply, you’re not likely to be creating a pay as you go service to rival AWS or Azure. Unless you are, then we wish you the best of luck.
A private or dedicated cloud architecture allows more customisation and flexibility than the public cloud, allowing a good amount of control depending on which flavour...
private = shared with low contention
dedicated = not shared, with no contention
We won’t dwell on the reasoning behind choosing the private cloud in this post, you can read more about that on our private cloud product page or you can check out this blog post, which covers the difference between private and public cloud. However, to cover it off in a sentence, it’s favoured by businesses looking for a level of performance that would make the public cloud expensive to run or for situations where security requirements rule out using an open and shared platform.
It seems like a logical idea — if you want to utilise the cloud, why not build it yourself? 100% company built, run and owned. A very valid argument, yes, but the logistics will take corporate IT teams out of their usual remit to deliver it. If you are a company like Walmart, with huge digital infrastructure requirements and - let’s face it - a whole lot of budget, building your own private cloud might well be a good option. However, for the majority of tech teams working within average enterprises, an at-scale DIY approach is unlikely to be the right solution.
First off, securing the budgetary buy-in for all of the hardware to build a robust private cloud with solid availability might be difficult. CAPEX spending in the hundreds of thousands can be a hard sell to a financial director, especially if you’ve been utilising pay as you go services or monthly recurring costs for infrastructure up until now. The budgetary side of things is only one issue but is important because it highlights that, even though this is a technical project, there’s a whole lot of buy-in it required from the wider business.
Using Walmart as an example once again, their DIY private cloud endeavours were part of an initiative to compete with Amazon in the e-commerce space, and so the infrastructure project was part of a wider aim to join up their business with technology. It wasn’t just a cut and dry decision to favour private cloud, more something that worked for the particular use case. It highlights that even though the public, private or dedicated decision is a tech decision, there’s a risk in making that decision in isolation from the wider business.
The pace of change and a DIY cloud approach For many corporate IT teams, while infrastructure is key to their success, it’s not their sole focus, and so creating your own cloud platform would likely put a resource skew towards infrastructure that simply wouldn’t be necessary.
For example, the security landscape is always changing, with methods of attack becoming increasingly sophisticated while more and more endpoints are accessing your network. With your infrastructure built and managed internally, it can be a real challenge for the average sized corporate IT team to stay up to date and safe.
If there’s a way for your business to security concerns by utilising a trusted solution from a particular provider, it becomes the duty of your partner to ensure compliance and enforce security, with their teams carrying out patching, monitoring, management and more on your behalf.
Of course, simply utilising a provider or cloud solution doesn’t guarantee security, but through due diligence and utilising trusted, ISO accredited services, you can pull on the deep security expertise held within companies whose business it is to focus entirely on infrastructure.
Deciding what to do with legacy applications is never a quick and easy process. They’re usually pretty important, hence where built using or still utilising legacy technology.
The first port of call for many businesses will consider making the leap from on-prem to public cloud hosted, which can sound costly depending on how much resource you have available to refactor and decouple your application.
The process of modernising your application for the cloud can be considered high-risk or perhaps too large of a first step for businesses. Especially those who are used to having their applications running on servers within the same building.
In this situation, a managed hosting provider offers a lot of value, as they can take on the initial, temporary hosting of your application, taking it out of the care of your tech team and so minimising the risks associated with it being hosted internally. There will be a monthly (likely fixed) cost for this hosting, but with it in the care of your provider, you can begin to have them look at ways and means of modernising it without rebuilding it or taking it offline. With it in the care of a provider who can really get to know and understand the ins and outs of your application, you will have more perspectives and approaches on offer than when going it alone, and you can offhand any modernisation work to your provider.
We won’t labour this point, given that it’s very well-trodden ground, but if you’re considering going it alone or partnering with a provider, it’s worth mapping out the responsibilities that are key to your technology function and who would carry them out. Managed hosting includes all of the fundamentals to run workloads and applications, all of which would be internally handled should you be building your own cloud platform.
Key managed cloud provider responsibilities
Hardware setup and configuration Software setup and configuration Patch and update management Monitoring and management Technical support (24/7) Security (managed firewall, DDoS mitigation and more) High availability/ uptime guarantee Backup/ DR/ offsite replication.
This covers the fundamental tasks undertaken as part of a managed services agreement, however, within hardware setup, scalability needs to be considered. I.E, if you build your own platform, you need to have enough resource ready and to account for spikes in demand. Of course, cloud providers have a whole heap of resource on-hand due to the nature of their business, whereas for a business, it probably doesn’t make economic sense to have servers live, ready and waiting for spikes in activity.
Somewhere between the extremes of building an entire platform internally and handing over the control to a cloud provider is an unmanaged hosting solution. In essence, this is what the big public cloud providers offer; access to infrastructure that’s not necessarily owned by your business, but hosts your applications, managed by your internal team.
While the public cloud isn’t always suitable, the majority of managed hosting providers who operate a private cloud will offer a cheaper, unmanaged option. The benefit here is that your business doesn’t have to invest heavily upfront to build a platform, as this is supplied by the provider. Aside from CAPEX and cost, there are a few added benefits to unmanaged hosting, such as hardware refresh, access to support (which might not be free, but is still there when required) and the ability to utilise said provider’s resource for projects or activity as and when required. It’s much quicker and easier to kick off a project this was as your provider has knowledge of your infrastructure.
Even if you are leaning toward a DIY cloud or a managed solution, if you’re a company currently managing your own infrastructure, the unmanaged hosting option could provide a useful stepping stone while your business fully gets to grips with what it needs and how best to operate.
Of course, as a company specialising in expert managed services, it’s always going to seem like we favour the managed approach, but it’s not that straightforward. No decision that means a fundamental change to your infrastructure should be made lightly, and it really does come down to your situation, current provision, budget, staffing, departmental aims and business goals. A few consulting days from a trusted source will be able to provide a steer, a plan or some potential options.
Posted in Managed Cloud on Oct 17, 2018