By Aaron Cox
When it comes to cloud migration, the odds are not in your favor....
Like most IT organizations you are probably either planning or considering a migration of applications and services to the public cloud. I hate to break it to you, but failure is more likely than success for many, the statistics are not in your favor. Whether it be an extended timeline, exasperated budget, or a complete failure in the act of migration. The pitfalls are many. Tech Republic and IDC recently shared some surprising insights related to cloud migration, in multiple surveys that interviewed 400 or more IT executive respondents. A few highlights:
- Cloud migrations often take longer than you expect - 73% of tech leaders expect their cloud migrations to take more than one year to complete, with 43% expecting the process to last more than two years. - Velostrata/Dimensional Research, November 2018
- Cloud Migrations are complex - 62% of companies said their cloud migration project was harder than expected, and 55% went over budget. - Velostrata/Dimensional Research, November 2018
- Enterprises underestimate the cost and effort of cloud migrations - 58% of workloads migrated to the cloud have returned to an on-premises hosting solution due to cost overrun or poorly performing architecture - Velostrata/Dimensional Research, November 2018
Based on the above statistics it is nearly a guarantee you will either experience missing your schedule, going over budget, or bring workloads back, and in some cases you may experience two or three of those scenarios.
With odds like this is your cloud migration doomed?
It doesn't have to be If you avoid the following myths, you can dramatically improve the success of your cloud migration.
Myth 1. Migrating to the cloud is quick and easy, just buy a tool.
We have explored the role of host migration tools already (two chucks and a truck) and how they fit into an overall migration program. They have a place, but cannot exist alone. Cloud Migration takes strategic vision and well planned architectural decisions in order to succeed. This will take time to accomplish, allow appropriate time to do the work right the first time, cutting corners or over reliance on tools will often negatively impact production and business operations. Start by building a schedule that allows you to collect the right level of information about your environment and to build a path to migrate using a data driven approach.
Myth 2. We have done a data center migration before, its all the same.
Today's Cloud Migration is not your father's Data Center Migration. We have migrated hundreds of data centers, and have learned that a successful cloud migration requires a different approach. The good news is that many basic principles still apply in order to set yourself up for success in the cloud, how you will operate when you get there is as important as the act of getting there, and your operational thinking will need to change. During the recent AWS Summit in Chicago, Wesley Story, Vice President of Infrastructure at Sysco foods stated, "“When it came to our migration we struggled when we treated it like a data center migration, we had to think differently and embrace the improved capabilities cloud gave us, when we used the easily accessible value of inexpensive experimentation and abandoned the set it and forget it sizing mentality of operating legacy infrastructure our success was immensely improved”.
In a traditional Data Center Migration the mentality is often to build large structured migration events, utilize tabletop testing methods, and migrate to infrastructure that will last the next 3 - 5 years.
In contrast Cloud migrations create the flexibility of separating applications and services into smaller digestible migration events that allow you to experiment with solutions, architecture, and methods before migrating production workloads and without interrupting your business. In addition, successful programs design public cloud infrastructure that is sized only for what is needed and allows you to grow as you go.
Expert Tip: Consider approaching transformational requirements in the same way. Addressing transformation in an iterative approach will help you accelerate your schedule, control your budget and transform as you mature to new capability in the cloud.
Myth 3. The cloud is cheap, and my budget constrained, migrating to the cloud will be the answer to all of my problems.
Can cost savings be found in the cloud? Certainly, but it is not inherently cheaper by simply picking up and moving your current environment as is (lift and shift) to the cloud (See: Why speed and agility are more important than unit cost in cloud). In fact, it will most likely be more expensive than it is today. To achieve meaningful savings careful analysis and design are required to utilize the services that would be cost prohibitive to build on-premise in a cost efficient manner rather than in a public cloud. Design for your objectives (cost, speed, agility etc.) and recognize opportunities to take advantage of the way cloud native tools change the methods of delivering redundancy, scale, and speed to market to free yourself from outdated architecture patterns.
Want the migration odds to be in your favor?
Consider replacing the previous myths with the following best practices -
Tip 1 - Do not become over reliant on tool(s). Migration tools are an important element of a cloud migration, however enterprises that succeed understand that it is a part of a bigger program. Build an effective Cloud Center of Excellence, establish objectives you can commit to, and build a schedule that supports gathering the data necessary to make data driven decisions and plans. Tools cannot give you the business context necessary to reduce impact to your organization during your migration. Rolling up your sleeves and getting involved with business leaders within your user community will yield surprisingly positive benefits and results.
Tip 2 - This is not a data center migration, build a migration approach that adopts low risk experimentation and functional testing in public cloud (before you migrate) and has a defined path into a cloud centered operational model. Once you have migrated you are not finished, the journey has only started, consistent fine tuning is a continuous initiative and will ultimately define the success of your project.
Tip 3 - Define an architectural roadmap that will allow you to transform in an iterative approach over time, after you migrate. Traditional architecture standards employ idle, redundant, always on infrastructure. Pivoting to Cloud Native continuous improvement cycles will promote driving toward the use of event driven services and architecture that reduce consumption while providing instant scale.
Tip 4 - Cloud cost awareness from day one requires well defined solution patterns, and then consist fine tuning to lesser expensive services and architectures.
Tip 5 - Everything as code, and I mean everything. utilizing infrastructure as code to develop templates, automate deployment and build efforts through native tools like Azure Resource Manager, AWS Cloud Formation or third party multi-cloud tools like Terraform, Salt Stack, etc. to reduce effort of deployment, reduce the risk of mistakes, while increasing the use of automation. New to writing code in JSON or YAML (typical infrastructure as code languages)? Consider tools like Stackery that make the coding process visual to assist you in the meantime.
Can you be successful in your cloud migration?
Absolutely, with careful planning, and change of perspective. Traditional data center migrations focus on migration from point a to point b. Once your journey to a target destination is complete and you have completed a handoff to operations, your objective is complete, and you move on to the next initiative. However that line of thinking will cost you when it comes to cloud migration. Embracing the arrival of your workload to the cloud is only the first step in making your cloud migration successful and winning the long game.
