Discovery and Assessment is the first stage of the cloud migration journey. In this stage, companies assess their cloud options to determine if it makes sense to move to the cloud from a cost-performance standpoint. Key objectives in Discovery and Assessment include:
- Calculate the TCO of moving to the cloud so you are confident that the cloud will be cheaper than managing your infrastructure on-premises.
- Know if your performance requirements for each workload will be met or exceeded at a lower cost in the cloud than on-premises.
Accurately addressing these objectives can quickly become overwhelming, and mistakes or oversights can lead to wrong decisions that have major consequences later in a company’s cloud journey. At Cloudamize, we recommend navigating the Discovery and Assessment stage by addressing the following six questions.
1. “What does our infrastructure look like?”
Complete and accurate discovery of every machine and application in your infrastructure is essential. Critical steps in migration planning – such as calculating TCO, predicting cloud performance, mapping dependencies, and designing the migration plan – rely on having a complete picture of current on-premises IT infrastructure. Unfortunately, many organizations begin with an incomplete picture because they rely on CMDBs. Since CMDBs are rarely up to date and many times do not account for shadow IT, relying solely on a CMDB will inevitably cause applications to be missed.
What’s the alternative to relying on an outdated CMDB or manually assessing on-premises infrastructure? At Cloudamize, we help companies obtain an accurate understanding of their infrastructure by tracking all commands and web server requests. This automated data collection allows our system to identify every application in an organization’s infrastructure. This includes potential Shadow IT, which we find by identifying dependencies going to IP addresses within environments that are out of project scope.
Additionally, we collect commands and web server requests at high frequencies. High frequency data collection is important because up to 80% of application and server connections are bursty or short-lived, so discovery solutions that do not collect data at high frequencies will miss these applications during discovery.
2.“What are our workload characteristics?”
Once you complete discovery, you’ll need a comprehensive performance analysis of each workload - this is absolutely essential to accurately calculating cloud TCO and predicting cloud costs. To understand workload characteristics, you’ll need to gather and analyze these six metrics:
- Peak CPU Utilization
- Allocated and Peak RAM usage
Observed Storage On-Premises (capacity and current occupancy)
- Disc IOPS and Bandwidth
- Usage patterns: Identifying how often compute and storage resources are on, idle, and unused
The key to accurately capturing workload characteristics is to identify short-term and long-term peaks rather than simply average them. Disregarding peaks and relying on averages while provisioning will inevitably cause performance degradation in the cloud. To incorporate performance peaks properly into your analysis, organizations need to consider detailed workload historical data over a period of at least two weeks.
3. “What cloud options are available for my workloads?”
After profiling the performance of each workload in your infrastructure, you’ll need to understand all of the compute and storage options available in the cloud. Cloud service providers provide multiple options to accommodate different use cases. These options include different combinations of CPU, memory, storage, and capacity to provide the optimal resources for all applications - and there are million of potential configurations. Understanding all of the options that are available is an essential step to being able to choose the right combination for each workload.
4. “What are the optimal settings for my performance targets?”
Now that your workload is characterized and you’re aware of the world of cloud options, the next step is to find matching infrastructure settings in the cloud (i.e. type of compute instances, storage volumes and throughput) and test them against performance goals. Essentially, project workload characteristics on every single available compute and storage option to see if performance targets are matched. The best practice is to run interactive, what-if scenarios against available cloud configuration options to identify the option that will deliver the performance and cost you’re seeking. This task involves data analysis, performance benchmarking, and predictive analytics. Without automation, it’s incredibly complex and time-consuming.
5. “Which pricing plan is right for me?”
After identifying the best compute and storage options in the cloud for each workload, the next step is to consider the different pricing plan options. For example, AWS offers an on-demand pricing plan and different types of reservation plans. Reserved Instance (RI) plans range from no up-front 1-year to 3-year all up-front RIs and can provide savings from 15% all the way up to 75% on top of on-demand pricing. Microsoft Azure offers pay-as-you-go subscriptions, and you can receive additional discounts based on your enterprise agreements. Meanwhile, Google offers a sustained usage model. If you know your performance profile, you will be able to identify the pricing plan that is extremely well-matched to your specific needs, which can help you cut costs significantly.
6. What’s the bottom line?
After answering questions one through five, organizations will know what they have in their infrastructure, their workload characteristics, their best options in the cloud, and their best pricing plan. Armed with this knowledge, it’s possible to calculate cloud cost, predict cloud performance, and confidently make the decision to move to the cloud and begin migration planning.
Finding answers to these initial questions during is a great starting point for beginning the cloud journey. The answers are important reference points for building a business case to move to the cloud, assessing cloud readiness, and planning migration. It can be a tough exercise, requiring a thorough understanding of your own environment and workloads, along with the options that are available through cloud service providers. The good news is that automated analytics can significantly improve the ease and accuracy of this process.