Cloud migration isn’t easy. In fact, just the opposite - it’s incredibly complex. There are a myriad of questions to address as companies plan a move to the public cloud, and answering each one accurately is critical for success. While not a comprehensive list of all of these questions, the following are 7 critical questions that you should ensure are on your cloud migration strategy checklist.
Here they are:
1. Discovery: What does my current infrastructure look like?
For a successful cloud migration, you need a thorough understanding of all applications and machines in your infrastructure. The Discovery phase is important to a successful cloud migration because knowing everything that exists in your infrastructure is critical to every other step in the cloud migration journey. Whether you’re calculating TCO or designing a migration plan, the only way to accomplish any of these tasks is by starting with accurate discovery.
Unfortunately, not all companies approach Discovery correctly. Many rely on a CMDB, which is a time-consuming and tedious process that is prone to errors because so few are kept up to date. Instead, leverage automation to track all commands and web server requests in order to identify every application and machine in the IT infrastructure, including Shadow IT.
2. Right-Sizing: Which compute, storage, and network configuration is optimal for each workload?
Many companies put off the process of properly right-sizing workloads until they’re already in the cloud. This leads to inaccurate decisions and poor performance in the cloud because right-sizing - identifying the compute and storage configuration for each workload that will provide the best cost-performance optimization - is critical throughout cloud migration planning. It informs calculating cloud costs, determining which workloads to migrate vs. keep on-premises, and planning migration phases (read Why Right-Sizing is Essential for Cloud Migration Planning).
Identifying the optimal compute and storage options for each workload involves a combination of in-depth performance analysis, usage analysis, cloud performance benchmarking, and predictive analytics. Without this exercise done precisely, initial workloads in the cloud will likely be over/under-provisioned, resulting in cost and performance issues that can halt your migration execution.
3. Dependency mapping: Which applications need to be migrated together?
Once you’ve identified all of the applications in your infrastructure, you need understand their dependencies and how they’re connected. For all applications, determine the other applications and servers they are communicating to, how often they communicate, whether the communication is bi-directional or unidirectional, and what the allowable latency is in communications between each.
Dependency mapping is where accurate Discovery becomes critical. Missing applications in Discovery, such as Shadow IT (which can account for as much as 15% of infrastructure), results in missing dependencies, which will then result in application connection breaks and security issues upon migration. Again, tracking all commands and web server requests at high frequencies ensures nothing is overlooked.
4. Application suitability: Are there suitable options for each one of my workloads in the cloud?
Some applications are better suited for the cloud than others. For example, certain legacy enterprise hardware may be more cost-effectively managed on-premises than in the cloud.
Also, applications that have bursty CPU and IOPS are much more cost-effective in the cloud because companies can take advantage of auto-scaling, which allows applications to be turned off when they aren’t being used.
Based on an application’s performance CPU, memory, and storage requirements, there may not be compute or storage options available in the cloud that match them well, and as a result may cost more or suffer performance issues in the cloud. Performance analysis, usage analysis, and right-sizing will help inform you on which applications you should migrate vs. keep on premises.
According to consultant Keith Townsend, three non-technical factors can also contribute to whether an application is suited for the cloud. These factors are:
- Licensing: If you would need an additional license to operate your workload from the cloud, it may be more cost-effective to keep it on-premises.
- Support: If your application provider is unwilling or unable to offer robust support in the cloud but your application is mission critical, keeping it on-premises may be the option that reduces the likelihood of encountering performance issues.
- Regulatory: From the perspective of most governments, not all data needs to be treated the same way. Applications that use or house certain kinds of sensitive data may not be able to store that data outside the borders of the country where the business operates.
5. Application Complexity: How do I identify applications that will be more or less complex to migrate to the cloud?
Organizations should prioritize less complex applications for phase 1 of their migration. These applications are less risky and easier to move. The best way to determine application complexity is to consider application classification, dependencies, and CPU usage. For example, certain applications that are not mission-critical such as disaster recovery, development, and staging tend to be migrated early on. Home-grown and mission-critical applications as well as those with more dependencies and greater CPU usage should be migrated later.
6. Workload TCO: What’s the estimated cost for each workload migrating to the cloud?
Nobody likes surprise bills. Knowing the cost of migrating each workload - before they actually migrate them - ensures companies stay within their budget for each phase of their migration. For instance, a Fortune 500 company leveraged Cloudamize to calculate what each workload’s annual costs in the cloud would be based on the right-sized configuration for each workload. This included compute, storage, and network costs for each workload as well as scheduled turnoff for those workloads taking advantage of auto-scaling. This analysis helped the customer eliminate workloads that were too expensive to migrate during phase 1.
7. Firewalls: Which ports need to be opened?
Understanding each application’s firewall rules allows you to configure your security groups. Not only will knowing your firewall rules help you understand which applications are talking on which ports, but they’ll also help you understand which ports you need to open to enable inbound or outbound application connectivity to work. Additionally, you’ll also want to understand your TCP and UDP ports so that you can build your firewall rules for your subnets.
Migrating to the Cloud is a Technical and Organizational Journey
The beginning of the cloud migration journey is overwhelming. Companies need to understand their on-premises infrastructure on a deep level while also understanding their best options in the cloud. Attempting this process by manual means is all but impossible and relying on outdated information can cause you to miss important information in the evaluation process. But by addressing the right questions and analyzing the right data, companies can migrate to the cloud faster and easier.