Migration (information technology)

from Wikipedia, the free encyclopedia

In information technology, the term migration (from the Latin migratio , relocation) stands for the process of a fundamental change in an existing system landscape or an essential part of this landscape. The migration can relate to both hardware and software.

Definition of terms

The concept of migration is complex. On the one hand, it can stand for the conversion of a single system. On the other hand, it can also mean large migration projects in which many systems or several new applications are implemented. The migration from one operating system to another usually includes the migration of application software and data at the same time .

The term porting is used to distinguish it from migration for the transfer of a program from one hardware and / or software platform to another. To do this, the programs - which were written for a specific operating system or platform - are changed so that they run on a different operating system. The terms are sometimes used synonymously.

Media migration

Media migration describes a process in which the physical data carrier medium of a data object is changed within an archive. It is thus a type of procedure for maintaining a bitstream.

There are four types of implementation:

  1. Refreshment: here data is only copied to a data carrier of the same type. There are no changes to data or the storage infrastructure.
  2. Replication: here, as with refreshment, data is copied from one carrier to a new one. However, this can also be a different, newer data carrier. The difference to the refreshment is the change in the storage infrastructure. Example: data from a floppy disk to a USB stick.
  3. Repackaging: an archive package is changed here, i. H. data objects are rewritten themselves. Example: A compressed file in .zip format is compressed into a file in .rar format. The change in the example is therefore in the pack format.
  4. Transformation: here, similar to repackaging, data objects are also rewritten themselves. However, the contents of the archive package are changed here. Example: a text file in .docx format is changed to a text file in .odt format.

Refreshment and replication are all about maintaining existing data by changing the storage media. They thus represent media migration in the narrower sense.

In contrast, repackaging and transformation also change the content of the data. So there is a double function. On the one hand, when the data format is changed, a new data object is created that is usually saved on a new data carrier. However, the migration measure is also carried out with a view to future interpretability, i.e. H. Readability of the data. That is why one speaks here of media migration in the broader sense or format migration.

Software migration

Software migration can be defined as the process of changing from a previous to a new technological environment.

Migration goes beyond a simple update or upgrade and rather describes a fundamental change in the software infrastructure. Migration strategies form the basis of a migration . Ideally, utility programs are available for largely automated conversion.

The most frequent reasons for performing a software migration are the obsolescence of the software or the existing old system (" legacy system ") is no longer able to meet new hardware and software requirements. The latter can especially be the case in organizations if there is a lack of suitable IT specialists for older software.

Examples of software migration:

  • the transition from the operating system Microsoft Windows on Linux or Unix to Windows. But switching from an old AS / 400 to Linux is also a migration. Individual computer workstations or the software required for individual work steps are often migrated step by step .
  • A partial migration, on the other hand, would be to partition a new AS / 400 ( System i ) in such a way that OS / 400 and Linux run on it at the same time and software from both worlds can be used on just one server .
  • The adaptation of platform-bound software to another (hardware) system, for which there are tools such as the User State Migration Tool .
  • It is also a migration if you switch from a major release to the next higher release from the same software provider. Industrial customers who still have an old SAP R / 2 information system in operation and want to switch to SAP R / 3 or mySAP are faced with a challenging task. Both SAP versions are fundamentally different. Such migrations are therefore sometimes extremely difficult and can fail; it would be better to speak of a port here.
  • Legacy migration: An old application is converted to new application software (for example with a more modern basic technology or to standard software ) in order to ensure long-term further development. While such a porting project used to require reprogramming of the application code, automated tools are now available for certain migration paths. One example of this is the replacement of the outdated 4GL platform Gupta Team Developer by the .NET platform.

Data migration

The term data migration stands for the transfer of data from a source system to a target system. A data migration is carried out when a new system is to take over the tasks of one or more old systems. The replacement of an old system with an already productive system requires data migration.

The reasons for data migration are, for example, inadequate database structures, application systems that can hardly be maintained, the merging of several information systems into a uniform system and the desire to introduce new database technologies.

Examples:

  • A bank replaces a self-developed system with standard software . It is not enough to just install the standard software. Customer data, accounts and account balances must also be taken over.
  • When companies merge , the data from both companies must be merged.
  • The conversion to another character encoding
  • The transfer of databases
  • The transfer of text documents containing macros to another Office format
  • The transfer of spreadsheets that contain their own formulas

A data migration consists of several steps - starting with the analysis of the source systems to validating the results. The actual migration consists of three steps.

In the extraction step it is filtered which data should be transferred. To do this, the data is first analyzed, then extracted and compared. In addition, control data such as mapping tables are provided. Second, there is a transformation. The data is available in the data model of the legacy system. They have to be transformed so that they “fit” the data model of the target system. Relevant data is selected, incorrect data corrected and multiple existing data is merged. The mapping then follows. In the third and last step, the transformed data is loaded into the target system.

The three steps correspond to the ETL process of a data warehouse . But the goal is different. A data warehouse should provide new insights, e.g. B. to understand the development of sales figures. In the case of data migration, however, the data remains semantically unchanged. All (relevant) customers are still available. The account balances are also unchanged. Only the data model can change.

A data migration can be technically implemented, for example, using ETL tools, special migration tools with SQL scripts. Reliability plays an important role (no accounts should be “lost”). Likewise, a large number of object types often have to be migrated (customers, accounts, share custody accounts, stock exchanges, balance sheet data, etc.). A sequence control coordinates the ETL process for the various object types. A migration verification examines selected test cases manually (pars per toto), for example, and also uses statistics. Statistics make it possible to find a “needle in the haystack”, for example if a single account out of 10,000,000 accounts to be migrated is missing.

Application migration

As part of the application migration , an application is replaced with a new one. This process brings together elements of software migration as well as data migration ; often new hardware is also required. Careful planning (see also migration strategies ) and implementation is crucial for maintaining data consistency and for a smooth change of functionality from the old to the new application.

Hardware migration

The migration of existing systems to new hardware poses roughly the same problems as purely software-based migration and is usually inevitably associated with a certain software migration via interface drivers . Data migration is avoided as much as possible.

A practical example is the transition from a classic Ethernet network to ATM technology while retaining the structured cabling .

Hewlett-Packard carried out a hardware migration to a completely new microprocessor technology for existing customers of its server products from around the 2000s. The servers at customers' sites with the older Alpha processors and PA-RISC processors are gradually being converted to the Itanium processor technology developed together with Intel .

Live migration

As live migration, the move is a virtual machine (VM) refers, in which a VM on the fly from one physical host system ( host ) to another transfer or move is. Ideally, such a move takes place without impairing the VM, so that ongoing work in the VM can be continued without interruption. The goal of such migrations is to make hardware easier to maintain and to enable load balancing .

Changeover to newer interfaces and technologies

A function or a parameter of a program or for example SGML elements in markup languages which may no longer be available in subsequent versions, or even obsolete programming techniques are as deprecated / obsolete ( English deprecated ) classified.

The point of continuing this, however, lies in the upward compatibility . Because if an interface were simply abolished, exception errors could easily arise . The old processing of the input on such an interface is therefore replaced by a simple error handling routine, for example by giving a function a return value. The caller then receives z. B. not an error, but at least one - if perhaps useless - value of the expected old data format . This avoids problems that can follow if the caller had not implemented error evaluation on this interface. The choice of the new dummy value, however, requires careful selection (a parameter of the data type, text for example, would have to be "none"returned as) and knowledge of the original value range (0, for example, could result in division by zero ).

To support the changeover, some programming languages or development environments have the option of marking deprecated technologies with certain keywords.

The handling of complex interfaces can be quite time-consuming, otherwise the upward compatibility is simply lost. The "entrainment of contaminated sites" may develop into eminent problems in the course of evolution: A typical example is the 16-bit - compatibility of the operating system Microsoft Windows , which is still the OS / 2 - and DOS ensure compatibility needs. In modern Windows versions this means that a separate DOS operating system emulator has to be implemented.

Balancing between the two is one of the major problems in versioning modern software. Therefore, in new versions between small (minor) and a large update (major upgrade ) differed depending on the extent to which upward compatibility is guaranteed. A migration across several versions ( releases ) can cause problems much more easily or even require a new installation .

See also

literature

  • Knut Hildebrand: IT integration & migration . Dpunkt Verlag, Heidelberg 2007, ISBN 978-3-89864-455-6 .
  • Michael Willinger, Johann Gradl, Frank Densborn, Michael Roth: Data migration in SAP. 3rd, updated and expanded edition. Galileo Press, Bonn 2012, ISBN 978-3-8362-1808-5 .
  • John Morris: Practical Data Migration. British Computer Society, Swidon 2006, ISBN 1-902505-71-9 (English).
  • Jesús Bisbal et al .: A Survey of Research into Legacy System Migration. Technical Report. Trinity College, Dublin 1997, cs.cofc.edu (PDF; 200 kB), abstract .
  • Klaus Haller: Towards the Industrialization of Data Migration: Concepts and Patterns for Standard Software Implementation Projects. In: Pascal van Eck, Jaap Gordijn, Roel Wieringa (eds.): Advanced Information Systems Engineering, 21st International Conference, 2009, Amsterdam. Proceedings. Springer, Heidelberg 2009, ISBN 978-3-642-02143-5 ( PDF , English)
  • Carlo Breves, Eberhard von Radetzky: Application migration in the context of consulting projects . In: Zeitschrift für Unternehmensberatung , 8/2008, Erich Schmidt Verlag.

Web links

General information:

Data migration tools:

  • Scriptella - open source Extract-Transform-Load (ETL) and script execution tool.
  • ETL integrator Oracle Software Delivery Cloud.
  • Data Migration Toolkit (DMT) - GUI-based Java program for the migration of files and database data (free tool that demonstrates data migration in practice).

Individual evidence

  1. Federal Office for Information Technology: Migration Guide, Guide for the Migration of Software (Version 4.0). In: http://www.cio.bund.de . The Federal Government Commissioner for Information Technology, March 2012, accessed on August 15, 2020 .
  2. ^ Rouse, Margaret: Migration. In: SearchCIO. TechTarget, October 2016, accessed August 15, 2020 .
  3. o. V .: porting. In: http://www.it-administrator.de . Heinemann Verlag GmbH, accessed on August 15, 2020 .
  4. port. In: Computer Lexikon. Alexander Kirk, October 26, 1998, accessed August 15, 2020 .
  5. Dagmar Ullrich: Bitstream Preservation. (PDF) In: nestor manual. A small encyclopedia for digital long-term archiving. H. Neuroth, A. Oßwald, R. Scheffel, S. Strathmann, K. Huth, July 2010, accessed on February 1, 2018 .
  6. ^ Christian Luda: software migration . Concepts and practical implementation using the example of a music database, BA thesis . Offenburg 2011, p. 1 .
  7. ^ Christian Luda: software migration . Concepts and practical implementation using the example of a music database, BA thesis . Offenburg 2011, p. 5 .
  8. Porting Project migrates Gupta users to .NET .  ( Page no longer available , search in web archivesInfo: The link was automatically marked as defective. Please check the link according to the instructions and then remove this notice. In: Computerwoche , October 30, 2006@1@ 2Template: Toter Link / www.computerwoche.de  
  9. ^ Elvan Öztürk: data migration. In: https://wwwmatthes.in.tum.de/pages/t5ma0jrv6q7k/sebis-Public-Website-Home . Faculty of Computer Science, University of Munich, July 18, 2011, accessed on August 15, 2020 .
  10. K. Schwinn, R. Dippold, A. Ringgenberg, W. Schnider, and S. Fedtke: The successful data migration from old systems. Friedr. Vieweg & Sohn Verlagsgesellschaft mbH, Braunschweig / Wiesbaden, 1999, accessed on August 16, 2020 .
  11. a b c Jochen Knösels: Process of a SAP data migration. Gambit Consulting GmbH, accessed on August 15, 2020 .
  12. Bittner, Deborah: Migration in information technology. In: Netzorange. Retrieved August 16, 2020 .
  13. Alpha servers are history. ChannelPartner, August 11, 2010
  14. Server Upgrade and Evolution. HP.com, accessed March 6, 2015
  15. Live Migration . Glossary at DataCenter-Insider.de ; Status: July 21, 2010