Businesses need to ensure that the money, time and talent they invest into cloud computing deployment isn’t wasted. Choosing the right costing model for the services running is a key component in the decision-making process, otherwise you’ll be paying for resources and/or services that aren’t being used.
For cloud workloads with constant resource requirements, reserved instances provide a discount of up to 72% compared to pay-as-you-go prices. Reservations are obtained by committing to a one-year or three-year plan for Virtual Machines, Azure Blob Storage, Database Commute capacity and many other resources and services.
A reservation can only be applied to Enterprise agreements, Microsoft Customer Agreements or by a Cloud Solution Provider (CSP). They can be bought with ease and flexibility, and can be paid for all upfront – or monthly at no extra cost. They also are best purchased by predicting a minimum amount of known capacity required and buying the reservation to suit.
Reservations can be easily scaled upwards as the demands of the business increase by signing a new one-year or three-year agreement.
Azure automation can be levered to power down virtual machines when they are not is use, which stops them billing on a pay-as-you-go model. This is especially useful if you know the virtual machines are not going to be used for large periods of time – for example this could apply to session hosts in Azure virtual desktop where you may want to ‘power down’ a subset of these out of hours.
Depending on the length of time these resources are powered down for can make them a more cost-effective option over reserved instances.
Azure Auto scaling can be used to scale virtual machines on demand, as resource requirements increase or decrease – meaning demand is constantly aligned to capacity. This is especially useful when applications suddenly spike to very high demand, despite having constant average low utilisation which can make them unsuitable for Reserved instances or Automation. This provides real time cost optimisation to your workloads.