Why take an application-first approach to cloud?

Moving to public cloud is a double-edged sword

On one side you get all the benefits of a decentralised infrastructure, self-service and autonomy for your internal departments, and no more of the headaches that come with managing a technology platform. On the other (sharper) edge, you can easily lose visibility of who is provisioning what and, if there is no governance controlling how services are deployed and managed, costs can quickly spiral out of control.

In fact, many organisations would be surprised at the amount of wasted spend in their cloud environments. The fault primarily lies in taking a cloud-first approach but applying your current legacy infrastructure thinking.

If you are still building and managing workloads in the traditional way, it will end up costing you more to move to the cloud. You won’t be optimising your environment and taking advantage of all the features of an elastic, on-demand and dynamic cloud platform.

So, with that perhaps controversial opening gambit, here are the management and process considerations you should incorporate into your cloud strategy, and the ways in which you can generate cost savings from your transition to a multi-cloud environment.

Redevelop your Workloads

“Lifting and shifting” your workloads to the cloud is a low risk approach and there may also be some initial advantages. It’s a quick sugar hit in terms of savings, but you will quickly lose this advantage if you don’t also apply an application-centric approach and if you aren’t constantly iterating and looking to make improvements.

Taking a Microsoft SQL database as an example, it’s a lot better to re-platform these applications for platform as a service (PaaS) than just move your virtualised servers from on-premises to the cloud.

Alternatively, you might also consider completely rebuilding your applications for a cloud world, by using containerisation or microservices. That way, you can adopt an Agile model to speed up development cycles and time to market with new products and services.

Introduce Governance and Compliance

Organisations today are typically using on average five different cloud providers, which can be difficult to manage and control. However, rather than being your organisation’s gatekeeper and restricting the accessibility and flexibility of the public cloud services in use, make sure you have a governance platform that gives you full visibility into your overall cloud activity, and the compliance policies in place to manage it.

Then, enforce tagging on everything that is spun up. It’s one of the simplest ways to take control and enforce lifecycle management on any cloud application, instance and service. Who owns it, which cost centre, purpose, type and duration of use, etc.

That way it’s much easier to manage services from creation to deletion, and to identify orphaned resources – hard disks, snapshot backups that are created as part of an application, but then forgotten about when that service is shut down or redeployed. Identifying and switching these services off can generate immediate savings in a pay-as-you-go, on-demand public cloud environment.

Consistent tagging also makes it much easier for you to group applications with common attributes and compare costs between different cloud providers or with on-premises or hosted infrastructure running similar services.

Leverage Cost-Saving Approaches

Lifecycle control is a great way to introduce greater cost savings for public cloud. Mandating expiry dates and alerts on any applications or instances spun up in the cloud will ensure that you don’t end up paying for unnecessary services.

In the traditional infrastructure world, the opportunities for just in time provisioning and automation have been limited. Generally, you’d have to build in a lot of headroom into your virtual machine (VM) platform for applications and services that might have sudden or unexpected peaks in demand. This is a costly exercise, especially when this capacity is underutilised for the majority of the time. Autoscaling is available with most major public cloud providers and, if switched on, will automatically trigger extra capacity for any sudden surges in activity and scale the services back down once the surge is over.

You can automate a range of other activities, such as eliminating storage that has been inactive for a set period of time, or power scheduling, to shut down workloads at certain times such as on weekends or out of hours. Non-essential infrastructure, such as test and development environments could be configured so that they are only powered up when they are needed by the developers working on the project.

Each cloud provider has different discounting mechanisms, such as Microsoft Azure’s enterprise agreement (EA), AWS reserved instances and Google’s committed use discounts. For example, with AWS reserved instances, where you know that certain applications will be used a lot, you can generate hefty discounts by paying upfront and locking in for fixed single or multi-year terms.

How We Help

Working with a cloud partner like Interactive ensures that you are not just optimising on one public cloud platform, but that you have adopted the most cost-efficient and productive multi-cloud strategy.

Having managed Australia’s largest private cloud for over a decade – and with extensive experience of both Azure and AWS environments – we’re well placed to assist you. From business case development to ongoing cloud management services, our cloud experts have the deep experience necessary for you to experience the full benefits of cloud. On average, our team saves cloud customers 30% through optimisation and management of their cloud. We are here to help and can find what suits your business, and budget, best.