If you’re planning a data migration, then it should hopefully come as no surprise to learn that the road ahead could be a complex one. Whether you’re looking to replace or upgrade servers or storage equipment; move data to a third-party cloud solution; or consolidate your website, the process is likely to be far more complicated than simply moving data from one place to another. Whatever your specific aims, your ultimate goal will undoubtedly be to improve performance and deliver a competitive advantage.
Sounds great but, of course, data migration is a complex, detailed procedure that presents a unique set of risks. This article aims to outline the best practices for planning your data migration while avoiding the pitfalls.
Types of data migration strategy
For many companies, the process of data migration is part of a natural progression that allows for focussing on priorities, fuelling growth and ultimately modernising the IT infrastructure. But, before any type of migration is undertaken, it is vital to carefully analyse how the process could affect your daily operations and for how long. Of course, the type of strategy employed will come into play when attempting to answer such questions.
Several options are available depending on project requirements and timeframe, but most migrations fall into one of the two principal types of migration:
‘Big Bang’ Migration
In the ‘Big Bang’ system of migration, the full transfer is completed in a limited time frame. Live systems will experience downtime while the data is transferred to the new database. The process is a one-off event and, if planned for adequately, can be completed relatively quickly. This approach has obvious benefits as the migration is completed within the shortest-possible time, but it is not without risk.
While transferring from one system to another, there will inevitably be downtimes for critical business functions. Few businesses can last long without their key systems being operational and the process can be highly pressured, leaving little to no room for error. It is therefore, advisable that businesses plan at least one dry run of the migration process before the live event and have detailed contingency plans in place.
In contrast, and as the name suggests, the ‘Trickle’ system completes the migration process in smaller phases over a longer period of time. The old and new systems run in parallel while the data transfers, eliminating downtime and operational interruptions and relieving some of the pressure of the “Big Bang” approach. However, like the “Big Bang”, it is not without risk. It must be possible to track which data has been migrated making it a fairly complex procedure, it could also mean that old and new systems are operating in parallel, with users having to switch between the two.
How to carry out a data migration
Whichever strategy you opt for, it is vital to plan for every eventuality. The following list outlines the essential steps to undergo before carrying out a data migration.
The first step in any plan is to assess what is achievable. Once the desired outcomes have been identified, a timeline, resource plan and budget can be put into place. Data migrations are intricate undertakings and the aim is to achieve key goals with the smallest amount of disruption possible.
It is vital to know your data, so carry out a complete audit before any migration begins. This will give you the opportunity to address any concerns straight away. Bear in mind that data can undergo degradation over a period of time so ensure systems are in place to maintain data quality.
Once the planning and auditing stage is complete, it is vital to take time to study and understand your data. Without sufficient understanding of both source and target, transferring data into a more sophisticated application will amplify the negative impact of any incorrect or irrelevant data, perpetuate any hidden problems, and increase exposure to risk. Consider using additional software or third-party resources to help resolve issues if necessary.
3. Design and Build
When you have the information and metrics from the planning and auditing stage, it’s time to determine what data needs to move and the quality of that data. The results of the audit should form the basis for developing a series of rules for transferring all designated data.
In a staggered or ‘trickle migration’ process, the data can be mapped, tested and transferred in stages which could make it easier to stick to budgets and deadlines and deliver better results.
Before any actual code is written, the mapping specifications and supporting documentation must be clearly understood and signed off by the business.
4. Execute, Test & Validate
During the execution phase, your data is extracted from the source system, cleansed and loaded into the new system using the clearly identified migration rules. Once complete, testing of all key systems must occur before any migration can be signed off by the business. Run back through your original audit to assess any changes/issues and take appropriate action. Ensure that everything’s where it should be, clean up stale data, and double check permissions. Finally, demonstrate that the migration has complied with requirements and that the data moved is viable for business use.
5. Decommission and monitor
Run through the pre-conditions that need to be met before you can decommission your old system. Ensure that these are fully documented and agreed. Close down your project by passing over the process and technology adopted to measure data quality during the project.
If you are replacing hardware as part of the data migration, it is vital that you receive full and correct documentation from your IT Disposal Provider to confirm all the data was securely sanitised. Without this, old hardware could potentially pose a security risk for your business, with old, but still sensitive, data still searchable. Your IT Disposal Provider should also be able to indicate exactly how the data was sanitised, providing extra peace of mind.
The key to success
This may seem like a huge undertaking but as with any such process, it is vital to thoroughly assess what is necessary for your business. Each situation is unique, and you should approach the task in a way that fits your own set of needs.
One thing is for sure, whatever the reason for data migration, in order to succeed, the process must be given the thought and attention it deserves. It should be considered its own project rather than simply forming an element of a larger process. Without adequate planning, there’s every chance that the project will exceed budget, run over its allotted timeframe or even fail completely. If you’re still keen to go ahead then the following quick reference guide could come in handy.
Data Migration: A quick reference guide
- Define – clearly define the scope of the project and its desired outcomes
- Refine – through profiling and auditing, refine the scope of the project
- Minimise – minimise the amount of data to be migrated
- Profile – profile and audit source data before writing mapping specifications
- Set – set, and be realistic about, budget and timeline expectations
- Sign off – secure sign off from senior business stakeholders at each stage of the process
- Test – volume test all data as early as possible
- Resolve – factor in time for testing and issue resolving
- Manage – segment the project into manageable chunks
- Focus – remain focussed on the business objectives and benefits throughout
If your data migration process means you’ll have excess computing hardware your office no longer needs, we can help. Head to our homepage or call 0333 060 5203 to discover what CDL can do for you.