Cloud migration has continually featured among the key talking points in town hall meets of CIOs and IT Heads for the past few years. A Gartner survey shows that about 75% of organizations today have at least some Cloud infrastructure presence. This blog takes you through the top pitfalls one has to avoid while implementing a Cloud strategy. But before I dive into that, I would like to start with various types of Cloud migration strategies.
Cloud migration can be categorized into six types:
- Re-hosting aka Lift and Shift
- Re-platforming
- Re-architecture
- Re-purchase
- Retire
- Retain
The most popular of these options, as chosen by organizations, is the “Lift and Shift” approach. The process involves using Cloud-native migration tools like AWS Server Migration Service, AWS Cloud Endure, Azure Migrate, GCP Velostrate (Migrate for GCE). The approach consists of replicating the Virtual Machines (VMs) to the equivalent Cloud VM along with the required databases, and you will be on the Cloud. However, it is observed that most organizations who opt for “Lift and Shift” neither get full cost benefits nor do they utilize the actual benefits of Cloud-native services.
Now that we know the various migration strategies, we can proceed to the top pitfalls one should avoid during migration.
Team Skills
The foundation of every IT shop is the engineers working in it. Not having the right expertise in the team would lead to decisions that are not cost-efficient, security flaws, and low on performance. To overcome these challenges, it’s best to induct team members with the required skills and certifications to form a Cloud Strategy and Migration team.
Big Bang Approach
One of the biggest blockers for Cloud Migrations to not achieve its required goals is that organizations start migrating heavy workloads at once. IT leadership has quite recently understood that with Cloud, a straightforward option would be to execute Cloud migration projects in phases. The more straightforward applications will go first, and an architectural review on what option would be best suitable for new applications.
Not using Cloud Native features
VMs are often inefficient when you are running a simple application on a high-spec machine. Migrating such a VM with like-to-like specifications would lead to a massive jump in your bills. A Cloud strategy team’s approach could be to group the applications based on technology types, complexities, criticality, and the volume of transactions. Based on the above parameters, they can consider going on a server-less architecture, on containers, or VMs.
Software Licensing
Software licensing is another area where people often go more wrong than right. With procurement, Software Asset Management, and IT teams working in silos, the problem quickly turns into an Achilles’ heel. With proprietary software like MSSQL or Oracle, or any other ERPs, it’s best to review the software license agreements before proceeding further.
Process Management
More often than not, your processes are more focused on your on-premise infrastructure. The challenge in addressing Cloud-related issues lies in the need to update your process books. Process change also means one has to change the culture. The approach of allocating a considerable number of resources to VMs and then not using any of them for a long time often leads to cash burn. IT Support teams should be trained on the importance of cost management. By building standardized stacks and proper monitoring of utilization, one can run Cloud operations with the least possible cost.
Security and Compliance
Security is one of the more challenging areas which either prevents or slows down your Cloud adoption. The worst-case scenario being, migrating to Cloud and then getting caught in the compliance web. The Cloud Strategy team should have Cloud Security Architects on board and have a defined strategy before starting the migrations. This would include understanding the compliances like PCI/DSS or HIPAA. One should also have policies to provide “Just enough” and “Just-in-time” privileges.
Another area of focus should be your connectivity between on-premise infrastructure and private Cloud. The transit data has to be well secured. One should also consider implementing solutions like threat detection analytics and vulnerability assessment. Depending on your DR needs, a proper strategy would ensure that your business is up and running at all times, no matter what issue/incidents occur.
Conclusion
Cloud migrations appear easy and straightforward due to the abundance of tools in place. The actual risk often occurs when you migrate to the Cloud infrastructure.
Here are some of the key takeaways to make your digital transformation journey less painful:
- Build a “Cloud Strategy Team” consisting of infrastructure, application, and security architects. This team should define the standards, tools, disaster recovery approaches, etc.
- Consider the “Infrastructure as a Code” approach to providing the infrastructure. This controls your costs and maintains standards
- Create a template of the most commonly used infrastructure stacks
- Train your development and support teams
- Start small, learn and then expand your scope
- Maintain regular processes of Cloud monitoring and cost usage
- Build Cloud-specific support run books so that the IT Support team has all the details to support the infrastructure.
Remember, Cloud Transformation is not ‘The Destination’; it’s ‘A Journey’
Good Luck!