IT Focus Area: infrastructure operations
April 27, 2016
10 Data Center Migration Red Flags and How to Avoid Them
When was your most recent data center migration? For many, the answer may be never.
There are some common, but not so obvious, items that can sidetrack or even derail some migration projects. You should take these into account as early as possible in your migration planning process to ensure a smooth and successful migration.
Find out how to take the chaos out of complex cloud deployments. Download your guide to creating and executing a successful cloud strategy.
Here are 10 data center migration red flags and how you can avoid them:
Red Flag #1: Viewing migrations as only an IT infrastructure project vs. a business initiative
When planning and executing a data center migration, it is easy to think of it as an infrastructure project for which the end goal is to move the physical and logical servers, storage, etc. While this is true, the purpose of those infrastructure components is to meet a business function or process through running applications. This means one of the keys for successful migration is the consideration of application and business capabilities in addition to infrastructure.
Red Flag #2: Inaccurate understanding of project scope
A complete and accurate understanding of both the assets and applications that are impacted by the migration is critical. A holistic approach needs to be taken that considers applications, their dependencies to multiple servers and the dependencies between applications. It is important to understand the business context of the applications, including business impact and allowed outage window for migration, along with the infrastructure and desired end-state.
Red Flag #3: Underestimating storage migration time and complexity
On the surface, storage migration appears to be fairly straightforward. You have a certain amount of data stored, and you want to move it to a new location. Yet it is not a simple one-for-one swap. The volume of data alone can make it time consuming and disruptive, especially if you need to shut down a production system in order to make the migration. Differences in architectures across vendors or product releases from the same vendor can get in the way.
Red Flag #4: The right tools are essential but they are not "auto magic"
Migration tools can be helpful. However, each tool is designed with certain assumptions in mind—assumptions that may or may not apply to your specific situation. For example, your storage migration requirements will take shape as you plan the migration. Factors such as the specific storage platforms you are migrating between, connectivity and available time will impact the fit of a tool for your migration.
With the appropriate tool(s) identified, there will need to be investment of both financial and personnel resources. Individuals will need to install and configure the tools. Establish a pilot platform, where possible, to confirm that the tools function as anticipated within your environment and for your intended purposes.
Red Flag #5: Not connecting the current data center with the new environments
Connectivity is a critical factor in data center migrations, but one that is easy to overlook as you focus on other issues. You may need a short-term increase in network bandwidth, which comes at a premium price and typically has a long lead time for engineering and installing. If you are running the old and new data centers in parallel as you migrate, you need to plan for sufficient bandwidth, acceptable latency, and maintaining quality of service for production traffic. This configuration needs to simultaneously support production requirements along with ongoing data migration and potential backups across the wire.
Red Flag #6: Failing to address all of the logistical details
This is another area that is easy to overlook until you are in the middle of a migration. You should ensure you have coordinated all the details. It is important to schedule all the pieces so they happen at the right time and in the right order. It is valuable to ensure that you understand the implications the schedule will have on the facilities. Often businesses and their data centers are in shared facilities, and plans made by other tenants or the property management company can significantly impact the logistics required for your migration. There are typically multiple vendors involved, each with its own requirements, so coordinating across all of them is critical. The better you understand and schedule the logistics, the smoother the data center migration will go.
Red Flag #7: Over-valuing uptime
In the UNIX world in particular, the longer a system has been 'up' the more favorable the up-time metric is considered. When a data center migration is looming, it is usually recommended to reboot the system to ensure all updates have been successfully applied and a smooth reboot occurs. This simple operation helps you avoid thinking there is a problem with the data center migration, when, in fact, it is a result of a system coming up in a different state than prior to the move. Performing the reboot prior to the migration is often easier said than done since these are typically critical systems and scheduling an outage window for the reboot may require approval from multiple areas.
A similar risk, but almost in reverse, can occur with network devices. Updates can be applied and active in RAM but they must, also, be saved to flash on the device. Otherwise, when the devices come up after the migration they may effectively be downgraded to a prior configuration. Additionally, saving the configurations to a separate location from where it can be restored is recommended. These are common practices among network administrators, but we are human and mistakes happen. Taking the time to confirm the completion of these steps prior to a migration will help avoid issues.
Red Flag #8: Not decommissioning old applications
Many companies do not take full advantage of the opportunity to optimize and/or rationalize their applications or operations when migrating. This can be a prime opportunity to have a discussion with the business to take the final (or initial) step in decommissioning an application that no one uses. The decommissioning should occur prior to the actual migration.
An honest and frank discussion about the value of an application or component vs. the cost to migrate should occur. For apps that are either redundant or at the end of their life, this discussion should be fairly straightforward. The harder discussion is for apps that are not quite at their end of life but close enough for consideration. This discussion is particularly valuable for organizations that have experienced mergers and/or acquisitions and are migrating to consolidate operations.
Red Flag #9: Missing opportunities to transform your IT envrionment
Beyond the hardware and software within the data center, a migration is both impacted by and can influence other processes within the IT organization. Tracking and closely monitoring changes which are occurring within the landscape of the servers and applications throughout the migration process leads to a successful migration. Put another way, take this as an opportunity to tighten your change control processes and have them become part of the of the entire IT organization vs. an afterthought. Significant data will be gathered during the data center migration project that can be incorporated into a configuration management system. It is essential to have the configuration management processes defined and consistently adhered to throughout the organization. A configuration system without effective processes is little better than proceeding without the system.
Similarly, standard operating procedures for data center operations should be confirmed and followed. There is no magic bullet. It requires intentional effort and support throughout the IT organizations.
Red Flag #10: Not confirming all assumptions
Throughout the planning effort, assumptions must be verified. Proceeding down a path too far without verification and buy-in from all parties can lead to a waste of time and money. An example is that all applications on a non-production server are non-critical. We have probably all seen situations of a pilot system slipping into a pseudo production usage by a key sponsor; often an executive of the project. Finding this out late in the game would likely alter the move plan causing additional planning effort and potential rescheduling.
Also, we must be cognizant of external business events impacting the project and the personnel available for working on a project. Some data center migrations occur as a result of a merger or acquisition. In these situations, there is a risk of increased attrition of personnel with a resulting loss of the institutional knowledge about applications and systems.
Transform your business with cloud. Download your guide to developing a successful cloud strategy.
Data Center Migrations Affect Every Part of Your Organization
Data center migrations don’t occur every year for many companies so it is important to work with a partner who can help you look at a migration holistically and how it will affect every part of your organization. By understanding the potential common pitfalls of data center migration projects, you can better ensure a smooth and successful migration that won’t jeopardize your reputation and internal brand to your company.