Cloud computing has the potential to deliver dramatic cost savings to organisations, but there is a caveat: these savings are only truly possible through efficient use of those services, by taking the specific nature of your workloads into account and complementing them with the right platform and the most suitable cost model for your business.
Simply put: the key to driving down costs in cloud computing is to optimise the use of your platforms. It’s important to understand the nature of the application or workload that you’re intending to run.
In this blog, we’ll look at some key questions to ask when considering your cloud platform costs.
How many users are you expecting to connect at any one point in time?
The best place to start is by assessing your business requirements. This will inform the technical requirements of your application and, in turn, your cloud platform. If you expect only a few users, or customers, to connect to an application at any given time, then you’ll only need relatively few computing resources. But if you’re expecting hundreds, thousands, or even tens of thousands of connections – an e-commerce site, perhaps – substantial computing power will be required. It’s also important to keep the bandwidth required in mind when factoring in the number of users.
The more performance or bandwidth you demand from a platform, the more it will cost. That’s why it’s important to have a good understanding of your application. Without doing so, you run the risk of purchasing too much compute and bandwidth, and face overspending on idle resources. Worse still, you might underestimate your requirements and cut back costs, only to encounter a short-term increase in demand, one your shaved down platform simply cannot accommodate for. That means degraded performance, operational setbacks and grumpy end-users.
What are your organisation’s requirements in terms of location?
Where are my users located? Are they internal to my organisation? Are they external, public users? Do they need secure access to the environment, or is it a public facing website?
These are some of the questions that all enterprise CIOs should be asking. Are your users primarily employees accessing your workload from behind a firewall, over LAN or WAN, or are they accessing workloads via a VPN from a remote location? All these factors should be considered when choosing a platform, and each platform will have its own cost implications.
The geographic location of users is also important. Location affects the distance between your end-users and the hosted application, increasing or decreasing the latency. A poor latency will negatively affect your user experience, and crucially, latency sensitive applications may not function at all under a high latency environment.
Having the flexibility to choose your geographic location allows you to benchmark application performance against the cost of various cloud computing platforms in different parts of the world, which ultimately optimises costs.
Security: How much of a target are you?
Security can be expensive – depending on the structure of your organisation and your application or workload. But, given the South African threat landscape, you cannot go without it. A practical approach to security is to think about how likely your organisation will be targeted. Your organisation must consider the types of security threats, and the solutions you need to wrap around your application. This includes the type of data transmitted across your network, and the value it has for cybercriminals – paying extra attention to data like personal information, credit card details and login credentials.
It’s also important to factor in the attention your business is attracting. As the Official Technology Partner of Amaury Sport Organisation (ASO) which owns the Tour de France, we saw an astounding number – millions, in fact – of breach attempts made against Dimension Data’s website during the Tour de France, and not one was successful. That’s thanks to the correct preparation. Remember, if people are aware of who you are, or who’s involved, then it’s very likely that your risk profile will change, and your business will become a target.
While there are preventive and remedial measures that must be put in place, incidents can happen despite comprehensive security. Remedial support does cost more but it’s important not to overshoot on threat estimations.
How predictable are your workloads?
What is your application’s usage pattern? What’s the type of construct you need – from more stable Standard Virtual Machines to the more flexible public cloud. Each type has its own billing method and price points, suited to the particular types of workloads they account for.
Static, regular workloads are better suited to Standard Virtual Machines, while regular workloads that have predictable peaks and troughs in performance need a Virtual Data Centre. Both constructs use month-to-month fixed price contracts, but slightly different pricing points.
On the other hand, variable workloads are better suited for on-demand public cloud and is billed according to what you consume. However, it may not suit regular workloads, and could potentially cost a lot more if used around the clock.
It can be difficult to account for every individual business need, so prioritise what matters most to your business and then determine how this will affect your platform choice. This includes an IT budget, especially if reducing your IT expenditure is your top priority.
Skylight’s integrated cloud platform makes managing disparate cloud platforms easy and cost effective. If you’d like to know more about reducing the combined costs of your cloud solutions, download our free guide to IT department budget development today.
Read more blogs below