Posted on November 1, 2017 by Eric Gardiner
It is important to recognize the impact Azure cloud usage and cost management will have on your success in the cloud. A thorough gap analysis will help you make prudent decisions and avoid costly and unnecessary surprises.
As many as 57% of IT leaders report exceeding their cloud budgets. Like misplaced dominoes, early mistakes sizing cloud requirements compound during planning and deployment. The result is migration failure or at the very least, costly overruns. Poor planning around Azure cloud usage and cost management is a slippery slope. As deployments scale past a workload or two, the reality of the problem only becomes evident when it is too late. At this stage, the usual recourse is to pay up, and start over.
Part of the challenge lies in properly assessing your cloud needs. This is more complex and time-consuming than many think. The devil is in the details and there are plenty of ways to get this wrong. Here are some of the recommendations we make to implement effective Azure cloud usage and cost management strategies.
1. Understand ‘pay-for-performance’
Appreciating the difference a pay-for-performance model entails should be job one. The traditional IT mindset has long been one of bigger, faster and better. Designing on premise infrastructure for the peaks in your business is widely seen as a best practice. In the cloud, however, provisioning for peaks is a recipe for increased costs.
Instead, the focus should be on usage. Look closely at workload performance. In many cases, you should actually downsize cloud resources relative to your on premise infrastructure. Base provisioning on day-to-day usage, and use bursting to support instances of increased demand. That way you only pay for exactly what you need and when. Careful monitoring is essential since usage and peaks in demand may change over time and with them, your monthly cloud bill.
2. Know your software licensing costs
Even at this early stage, it is important to investigate the changes required to the software you are considering migrating to Azure or AWS. Many versions of software are not cloud-ready and will need upgrading. In some cases, this can be costly, making these applications poor choices for migration. This is often the challenge with versions of SQL Server, which is one of the first applications many consider moving to the cloud.
Just as important, do not forget about the inter-dependencies between applications. If two applications communicate with each other, migrating only one could result in performance delays and user dissatisfaction. In these instances, you have to consider the software licensing costs of moving both.
3. Plan for Azure cloud security expenses
A lot of confusion persists around cloud security. Many organizations incorrectly assume the provider owns the entire security stack. Not recognizing where your cloud provider’s responsibility ends and yours begins can result in unanticipated costs. Beyond paying to license security software, there is also the impact of fees and fines associated with non-compliance.
4. Account for Cloud Data Costs
If you build it, data will come. Using cloud as part of your storage and disaster recovery strategy is a natural starting place for many organizations. However, as you move more workloads to the cloud, data feeds from other tools move along with them (think CRM and analytics). Storage costs can quickly get out of hand.
Expect this phenomenon and factor it into your planning. Keep in mind that once data is in the cloud, it often remains there even after the removal of its associated workload. This ‘zombie data’ has an ongoing cost and may no longer offer value to the enterprise. This is where effective governance comes into play.
5. Enable Effective Azure Cloud Governance
Users have a tendency to put workloads in the cloud and then forget about them. It is common for IT to add cloud resources to support application development and then fail to decommission these resources once the project is complete. The result is “compute-sprawl” – one of the most common drivers of cloud cost overruns.
As early as possible, consider who will have privileges to deploy workloads and be held accountable. Create policies and procedures for requesting, provisioning and decommissioning workloads. This will help reduce unexpected costs and unnecessary finger pointing. Of course, good governance will only get you so far. To connect the dots, having the right tools to monitor cloud usage – whether by user, department, or by project – is essential.
A proper gap analysis entails a fair bit of time, effort and expertise. Beyond avoiding costly surprising, there is another hidden benefit. Planning a migration is often the best time to de-clutter applications, modernize and determine usage requirements. While this makes migrating easier, it also helps right size and improve the performance of on premise infrastructure.
Continuously monitoring cloud usage and costs is essential to effective cloud management. It should be an integral part of your cloud regimen. And it starts with effective planning.