In our previous blog, you could discover the reasons why the future is on the Microsoft Azure platform. In this blog you get an overview of the step by step approach to ensure an optimal infrastructure for your SAP environment.
The right size for the perfect fit: Microsoft Azure Virtual Machine sizing
It’s important to understand the different available machines on the Microsoft Azure Platform and to decide which machines meet your SAP landscape requirements the best. The virtual machines are priced according to size, and also have dedicated levels of service with respect to how many data disks you can attach, the storage I/O quotas applied, network quotas, and as well, the size of the temporary virtual hard disk increases with the larger virtual machine sizes.
Determining the right size, step by step
The size of the virtual machines your SAP environment is running on and your SAP System Architecture have a significant impact on performance and management. That’s why we take it step by step, to ensure an optimal infrastructure for your specific requirements.
- Map your current landscape: Build an inventory of your existing infrastructure and SAP environment. Evaluate the CPU, Memory and storage (Size and IOPS) requirements for your current environment. A good practise is to map your current CPU to SAPS from your existing infrastructure (based on vendor specific conversion table) and map those against the new CPU architecture, taking parallelism of processes into account. At the same time you can compare your existing SAPS with these benchmarked for the Azure virtual machines.
- Database or Application Server SAPS: Make a precise distinction between database SAPS (for scale up scenarios only) and SAP application server SAPS (which can scale out). Dedicated database SAPS are typically less than 20% of the total number of SAPS.
- System availability: Draft your business requirements towards availability and disaster recovery and take this into account when designing your deployment model on how you best meet your organizations RPO/RTO requirements by distributing machines over different availability sets, availability zones or even different Azure regions.
- Performance: Evaluate if your current system performance meets up with the business expectations. Moving to the cloud can bring great improvements towards practically on-demand scaling capabilities and storage performance improving application response.
- Design Microsoft Azure Virtual Networks: Keep in mind that multiple network adapters inside a virtual machine do not increase the network throughput. Also think about security between different networks for production, quality and development by making use of subnets or different virtual networks.
- Test on-premise to Microsoft Azure latency: Microsoft Azure offers a very large set of datacenters around the globe. Connection to these datacenters can be established over internet or a private connection connection. Depending on the latency requirements you can consider to connect trough IPVPN or ExpressRoute.
- Right-size your environment: Microsoft Azure deployments do not require a large sizing buffer like on-premise architectures do, as resources can be added as the business volume increases, for example at year-end, or after acquisitions. For large instances you carefully need to plan ahead.
As infrastructure experts, the engineers at Flexso can help you determine the present and future requirements of your SAP environment, and guide you towards a future proof architecture in Microsoft Azure that is perfectly aligned with your business goals and transition roadmap.