10 Considerations for Data Migrations and Translations
Are you faced with a complex data migration or translation? Do you have years of legacy data that needs to be migrated to a new system? Have you got old CAD data from a outdated system that is still being used?
If you have answered yes to any of these questions, you are facing the prospect of performing a migration or translation project. Here are 10 potential problems that you must look out for before starting:
- Underestimation of effort – too many projects are underestimated, primarily because the use cases for the translation are thought to be simpler then they actually are. For example, assemblies only need translation until someone remembers that drawings need to be included.
- “Everything” syndrome – Looking at a project, most organizations default to attempting to translate or migrate everything. In all cases, this is not necessary, as only a subset of the data is really relevant. Making this mistake can drive up both cost and complexity dramatically.
- Duplicate data – of everything that needs to be moved, how much of it is duplicate data (or same data in slightly different forms)? Experience shows that duplicate data percentages can be as high as 20 to 30 %. Unfortunately, identifing these duplicates can be difficult, but there are techniques to overcome this problem
- Accuracy of CAD translation – When looking at 3D CAD translations, how accurate a copy do the translated models need to be relative to the originals? Again, a blanket requirement of “identical” can drive up cost and complexity hugely. Some lesser target (say +- 2 mm) can improve success.
- Data already exists in Target – Some level of informal manual migration may have already occurred. So, when a formal migration is performed, data “clashes” can occur and result in failures or troublesome duplicates.
- Automatic is not always best – Developing an automated migration or translation tool can be costly, if the requirements are multiple. Sometimes, a manual approach is more cost-effective for smaller and simpler cases.
- Data Enrichment – Because the source data was created in an older system, it may not have all the properties and data that the target system requires. In this case, these have to be added during the migration or translation process. Forgetting about this step will prevent users from accurately finding data later.
- Loss of Data – For large data volumes, is it possible that some of the data is missed and deleted during the project? Very possible – to prevent this requires exhaustive testing and planning.
- Archive Solution – Once the translation or migration is complete, what happens to the original data? In some cases it is possible to delete it. However, in some environments (e.g. regulatory situations) this may not be allowed. In such a case, has an archive solution been put in place?
- Security – Legacy data may be subject to security (ITAR, competitive data, etc.). Does the migration or translation process expose sensitive information to unauthorized users? Often a process will take the data out of its protected environment. This problem has to be considered and managed.
Ask these questions before translations and migrations begin!