The challenges of legacy migration projects are many and diverse. They depend on the scope of the migration process, for example, whether the process involves only the migration of data into a new system, or the migration of application functionality, or both. The challenges faced with a migration project also depend upon the state and nature of a legacy system that needs updating.
One of the major problems with data migration is handling the functional transfer of data that can be in various conditions. There is the challenge of evaluating the state and condition of the data in the legacy system, for example, looking at whether raw or incomplete data sets will have an adverse effect on migration or use. There is also the essential challenge of figuring out which data sets will have to be migrated — setting aside unnecessary data can decrease costs by quite a bit, but where to cut is the question. Data scientists or planners may need to spend substantial time and effort identifying key data sets to be migrated.
Another issue with data migration is that companies may face problems where it is not possible to automate the migration, and data needs to be entered by hand. The data entry element can present any number of problems, including the need to acquire the labor to do the data entry itself. In addition to hand-entering data for migration, project implementation staff may also need to hand-code parts of an architecture.
Other challenges involve dealing with the “containers” or systems that handle the data. Legacy software migration can involve many obstacles, some having to do with understanding the nature of the system as it was originally coded. Experts talk about problems with proprietary systems and nonstandard tooling or coding as being central to many legacy migration projects. Planners have to be able to understand how systems were set up (while some of them display the rudimentary design of earlier eras of IT) and how they will be translated through to a new environment or platform.
Other inherent challenges involve the details needed to plan for these projects. Another issue that experienced IT people bring up time and time again is scheduling and resource allocation. When not enough time is given for a key phase, the entire project suffers. For this and other reasons, experts suggest that offices need to have a detailed legacy migration plan on the table before attempting this type of project. Others talk about “data governance policies” or an “org structure” — where some resources involve delegating responsibility for key parts of the process. All of this planning helps to support a migration project that goes relatively smoothly, and one that is ultimately successful.
上一篇我們送上的文章是
What is ‘precision and recall’ in machine learning? ， _!在下一篇繼續(xù)做詳細(xì)介紹，如需了解更多，請持續(xù)關(guān)注。
本文由
日本NEC鋰電池中國營銷中心于2023-07-16 17:44:06 整理發(fā)布。
轉(zhuǎn)載請注明出處.