The course is part of these learning paths
Microsoft Azure provides a variety of cloud services in a variety of cloud service models: IaaS, PaaS, and SaaS. When businesses migrate to the cloud, they must choose which model is best suited to their needs and is the most cost-effective. This course is designed to assist Cloud Architects in identifying their current Azure expenditures and providing greater awareness of the costs associated with each deployment model as well as each aspect of an Azure deployment.
Optimizing Azure costs begins with knowing your current Azure expenditures. This course introduces you to the tools built into the Azure Portal that can help you understand the total overall expenditures in Azure as well as break down those costs by area: Compute, Network, Storage, Identity, and App/Cloud Services.
The remainder of the course drills down on specific costs associated with each area of Azure identifies the costs associated with each service and provides very clear and concise methods for reducing Azure expenditures. Many of the cost savings methods will require minimal changes to your Azure deployment and will take just minutes to implement while other cost savings methods may take a shift in your Azure strategy, such as moving from Iaas to PaaS. By the end of this course, you will have gained a thorough understanding of how charges are incurred, how to reduce or even avoid some of the charges, and you will have learned how to significantly reduce overall Azure expenditures and get the most out of what is spent in your Azure deployment.
Learning Objectives
- Identify current Azure expenditures
- Optimize compute costs
- Optimize network costs
- Optimize storage costs
- Optimize identity costs
- Optimize App Service and Cloud Service costs
Intended Audience
- Azure architects who are exploring options for reducing their Azure spending
Prerequisites
- Good understanding of Azure administration and management
To summarize, the strategy of Azure cost optimization begins with discovering where your Azure dollars are currently being spent. There are several tools to assist you with this process, the cost management and billing feature in Azure and the Azure Cost Management tool, also assessable from the cost management and billing feature. Once you've identified your current Azure costs, you can then prioritize and focus on reducing costs and maximizing your resource usage. If you have resources deployed on Azure IaaS virtual machines, you can reduce your compute costs by using virtual machine scale sets to add or remove server capacity on demand. It may also be possible to configure one or more of your scale sets as low priority and have instances deleted when demand is low. When it comes to disk storage for your virtual machines, it is more cost-effective to use managed disks. And of course you need to choose the appropriate disk type, premium, standard SSD, or standard HDD. Use the advisor feature in the Azure portal to help you to locate underutilized virtual machines and delete virtual machine disks that are no longer needed.
If you're currently using pay as you go pricing, seriously consider using Azure Reserved Instances. Whether a one or three-year commitment, you'll instantly realize significant cost savings. When analyzing your network costs, remember that in most cases, you're only paying for outbound data transfers. The exception to this is with VNet peering where you'll pay for both inbound and outbound traffic, so it's important to determine if VNet peering is absolutely necessary. The vest way to minimize network costs is to deploy all of your resources into a single region. Lastly, consider if ExpressRoute is a necessity in your deployment. ExpressRoute, while providing fast throughput, is expensive. Leverage your existing site-to-site capabilities whenever possible. Reducing storage costs begins with understanding that charges are based on per gigabyte per month. Carefully review what is currently residing in your storage accounts and delete unneeded files and virtual machines, as well as virtual machine snapshots. Instead of storing snapshots, use the Azure recovery vault in combination with a comprehensive retention policy to store your server backups.
In analyzing your identity costs, the key factor is in deciding what Azure AD edition is needed for your organization, Free, Basic, Premium 1, or Premium 2. Carefully determine which features are important in your organization and choose the lowest cost edition that supports those features. If your business qualifies, you can also save money by entering into an Enterprise Agreement with Microsoft. Reducing application deployment costs begins with determining whether your specific deployment requires that you have access to the operating system.
If so, Azure Cloud Services provides that level of access. Otherwise, the preferred method is by using Azure App Services. With either service offering, it's important to choose the right plan based on the demands of the workload deployed. Also, deploy all of your production web apps into a single plan, and deploy your development environment into another plan. And in many cases, the production and development deployments are best deployed in separate Azure accounts, and then you can delete the development resources when no longer needed. For Cloud Service deployments, remember that there are charges incurred for your resources deployed to any and all staging slots. Delete these slots, for example, after a successful and stable deployment of a development slot web app that was swapped to the production slot. By periodically auditing your Azure spending and implementing the best practices outlined in this course, you will be able to maximize your Azure expenditures and reduce your Azure spend.
Jeff is a technical trainer and developer residing in Arizona, USA. He has been a Microsoft Certified Trainer for the past 18 years, providing in-house development and training on Microsoft server operating systems, PowerShell, SQL Server and Azure. When he’s not developing and delivering courses on Azure, he’s photographing galaxies, nebulae and star formations from his computer-automated observatory in Chino Valley, Arizona using a 14” Schmidt Cassegrain telescope.