This blog was co-authored by Dave Schroeder and Joshua Moye of Kalypso: A Rockwell Automation Business.
Think about a data migration expert as a professional data mover. Our job is to package, transport, cleanse, and adapt business information in a way that allows an organization to keep the data it needs when employing new solutions that enhance their digital thread capabilities. These solutions must be custom designed around current data and processes to be truly effective. The data migration process, described in the sections below, can be leveraged for new technology implementations, such as PLM or IoT, to help organizations move from where they are to where they want to be.
Data migration frequently presents a risk to implementation projects. To understand why, consider the challenges inherent to the migration workstream. At Kalypso: A Rockwell Automation Business, we’re often asked to move data from decades-old legacy systems that are poorly documented and only understood by a few key staff members. Additionally, these systems are more likely to contain inconsistent or inadequate attribution over time, further complicating the data model. Finally, even after an organization has begun a data migration, circumstances change as new data is created daily. If there are any corner cases that are missed, or if data is moved incorrectly, it could prevent someone from being able to complete their day-to-day work.
A successful data migration allows organizations to improve their data quality while moving away from previously siloed and segmented data. It also reduces swivel chair activities that are commonly associated with using multiple systems for a single task or process, thus increasing overall productivity. Most importantly, a successful data migration doesn’t cause significant interruptions to existing business processes.
There are many similarities between moving business data from one system to another and moving to a new home. In either case, the first thing you need to do is pack, or extract, your belongings. In the data extraction phase, your business can hire professional data movers to identify and package all the necessary and relevant data that will be moved later. As a rule, however, your business should only take the data it needs. So, when preparing for your move, you must determine what information is no longer needed and what data can be thrown away or cleansed. Doing this will be beneficial for your business and your end users as it rids your system of inaccurate or outdated information that may cause inefficiencies later on. When working with older systems or data, these can be difficult decisions to make. Some sets of information may have different or limited metadata, further complicating the migration.
As you extract that data, it can be useful to index or categorize it, making it easier to search for and find later on. Compare this to how you might label certain boxes once they’re packed to make them easier to find and unpack later. When captive data like old photos or documents on a network drive without any attributes or a system to manage them is freed, it is much more useable and valuable to a connected enterprise.
It's also important to know how data will be used once it’s migrated to the new system, not just where it will be stored. It's common for a business to streamline and standardize complex or redundant processes when implementing a new system. In this scenario, data may need to be transformed to fit the new system and ways of working. Every corner case matters and the new system will be designed accordingly. In some cases, data may not easily fit into the new system or process as it currently exists and will need to be transformed. Compare this to how you might expect to rethink your floor plan when moving into a new home. As you evaluate your new space, you may expect to shift some furniture or enhance your living space with some new accent pieces.
Often, we are asked about what factors make data migration more complicated, especially regarding volume of data. In a way, migrating one million drawings of the same type is fairly straightforward. In contrast, migrating 1,000 drawings of 1,000 different types is much more difficult. We will likely have to transform much of that data to fit the target system and the number of transformations needed is a key component in determining the effort. Additionally, some object types require more effort than others. For instance, migrating a set of documents is much simpler than moving a set of CAD data, or a set of parts and documents with linkages to other objects, or even external files. Lastly, we must consider how much of each data set's history we require in the new system. Due to the increased complexity, migrating any object's full history is not recommended, except in rare cases.
In larger systems, it is practical to split datasets and move them in pieces. This is also a more difficult process and means the data must live in at least two systems for some time. However, it allows large data migrations to happen faster, as trying to migrate all an organization's data at once could take years.
It's moving day! The final stage of the data migration process aims to load the newly transformed data into your target system. To ensure that there aren't any unknowns on moving day, Kalypso's data migration experts practice moving all the data 3-4 times during a rehearsal process. With each iteration of this process, Kalypso captures specific if-then-else rules that dictate where data is stored. Ideally, there is a place for everything, but this is often not the case. You may be migrating from a system that doesn't align with the target system’s data model. For example, in Windchill, multiple revisions of a part can have the same name and number. In other systems, each revision may have a different name and/or number.
When dealing with uncertainty, you can end up with messy or misplaced data – compare this to the boxes that may end up in the basement after moving into your new home. Consequently, Kalypso will work with someone on the client side who knows the data well enough to make a judgment call. This data expert will decide on an individual basis where each remaining piece of data should be stored. Finally, using the rules we discerned and the guidance from the client’s data expert, we execute the migration and prepare to go live on a new target system. When we’re done, organizations can enjoy the benefits of having more of their data under one roof – increased visibility, efficiency, and reliability, all without breaking any of their normal business processes.
For any organization to achieve a digital thread enabled enterprise, they will need to embrace emerging technologies and shutter legacy systems. This will require some level of data migration activity. Although it is a complex and inherently risky workstream, proper data migration results in more useable data, housed in a more powerful system, for a more connected enterprise.
To learn more about Kalypso’s data migration capabilities and joint initiatives with PTC, visit their PTC Partner page.
Stephen Jones is the Technical Senior Manager, Kalypso: A Rockwell Automation Business. Stephen has nearly 15 years of PLM experience specializing in Data Migration and RPA. He has been at Kalypso for 3.5 years and has taken over as the Data Migration Service Line Lead.