SAFRAN is a global provider of high-tech equipment mainly for customers in the aerospace and defence sector. For aerospace and defence activities, strong compliance requirements have to be fulfilled. One of these requirements is the long-term storage of all technical documentation for the different projects. Long-term in the SAFRAN business means quite often 30 or more years. This paper describes how these documents are managed inside Sagem, one branch of SAFRAN, why SAFRAN decided to replace PDF by PDF/A-1b as the long-term archiving format, how the migration from PDF to PDF/A-1b was managed and how SAFRAN plans to distribute this solution all over the SAFRAN group.
SAFRAN at a Glance
- SAFRAN is an international high technology group
- More than 10 billion euros sales in 2008 (800 million operating income, 250 million net income). 54,500 employees in over 30 countries (as of December 31, 2008)
- Three branches of activity: Aerospace propulsion, Aircraft equipment, Defence and Security
- Twelve major companies : Snecma, Snecma Propulsion Solide, Turbomeca, Techspace Aero, Aircelle, Hispano-Suiza, Labinal, Messier-Bugatti, Messier-Dowty, Teuchos, Sagem Défense Sécurité and Sagem Sécurité and a lot of daughter companies.
- World-wide no. 1 in: mainline jet engines with more than 100 seats (through equal partnership with General Electric), military and civil helicopter engines, helicopter flight control systems, landing gear, carbon brakes for mainline jets with more than 100 seats, digital fingerprint biometrics, space optics
- And a lot of other world-wide or European leadership positions
SAFRAN Informatique is the IT company for the Group. Its main mission is to deliver IT solutions and services that empower SAFRAN Group people and make a difference. SAFRAN Informatique manages more than 1,100 servers and 1,400 applications on more than 150 sites in the world and for more than 20,000 users. It manages more than 140 SAP environments.
SAGEM PLM and Document Management History
At the end of 1998, the top management of Sagem SA – a branch of SAFRAN – decided to deploy SAP PLM + ERP all over the Sagem Group (telecommunication, automobile and defence/aeronautics activities at that time) as a complete and integrated information system for more than 6,000 users. It was the first important SAP PLM deployment in France! (See picture 01).
Picture 01: Architecture of Sagem’s PLM and ERP solution
This new and single SAP PLM system replaced more than 40 TDM main applications and moreover the interfaces between them. The solution allowed access the to same, current and correct data in real-time, and is shared by all users of the company around the world (with China, Austria, Czech Republic, Tunisia, Brazil, and many more sites).
The Sagem SAP-PLM System provides a lot of important functions e.g.:
- components management
- all build of material (BOM) management (functional, definition, manufacturing, sailing, maintenance)
- document referral management (project, specifications, mechanical and electrical CAD files, manufacturing, maintenance …)
- workflow and process management
- change and configuration management
A very important component of this PLM solution is the integrated document management system which stores and controls all PLM related documents (see picture 02).
Picture 02: SAP DMS – ERP and PLM integrated document management system
All documents are managed in SAP DMS within the document information record (DIR) which contains all metadata for the document identification, status and classification attributes. All files (originals and PDF’s) are linked to such a DIR. All documents are stored in SAP vaults based and centralised on SAFRAN Informatique servers.
In Sagem’s SAP PLM solution, the link between documents and components/products is ensured with a BOM (bill of material) link to specific objects named “dossiers”, and each dossier contains only documents of one stage of the product lifecycle, e.g.: project dossier, justification dossier, definition dossier, manufacturing dossier, maintenance dossier…
The validation of the documents, including the validation of the files, is done through an SAP workflow and status cycle. This document status cycle includes an automatic PDF conversion of original files, just before the final validation and vault storage. This allows the project managers to do the validation without a CAD workstation but directly in SAP from his PC. The PDF generation should also guarantee long-term access to these documents, at any time, for any purpose.
Since the beginning, the SAP platform has used the SEAL Systems solution for SAP integrated printing and PDF conversion of all documents stored in SAP vaults.
Migration to PDF/A
In 2007, Sagem Defense and Security decided to replace the standard PDF conversion with a conversion process to PDF/A-1b. This decision was based on problems caused by the impossibility to visualise again old PDF documents from the SAP vault. These defective PDF files were not coming from the direct Microsoft Office conversion processes, but from paper to PDF scanned documents, from documents provided by external parties or from the direct PDF exports of the CAD applications. It was decided to reconvert all already stored and all in the future produced PDF files into PDF/A-1b, because PDF/A-1b had become the international ISO standard for the archiving of electronic data files. This guarantees for our documents:
- the long term storage and accessibility in the SAP vault, essential for aeronautic and defence activities
- the hardware and software independence
- the future capability to visualise these PDF/A-1b files free of charge with Adobe Reader, and without risk of lack of data because these files:
- include all character policies
- contain independent colour definitions
- support standard metadata for comments
- support full text retrieval and document structures
For the migration from PDF to PDF/A-1b we needed to replace the PDF generation processes by PDF/A-1b generation processes. But this ensured only the PDF/A-1b availability for processes where the PDF creation was under our own control. It did not include a solution for the PDF files already stored and for PDF files coming from external parties or PDF generators which were not under our control. We therefore decided to implement the PDF Longlife Suite from SEAL Systems.
With the PDF-Checker and PDF-Adjust modules of this suite, every PDF file can be checked against its PDF/A-1b compliance and corrected in case compliance is not fulfilled. With the direct integration of the Longlife Suite into SAP DMS it is guaranteed that every PDF document is checked and corrected if required, just at the moment when the document is checked-in into SAP DMS. We also built a process for all documents already stored in the DMS and checked and adjusted them to PDF/A-1b.
Currently 4,000 PLM users have access and create DMS controlled documents.
Up to now we have:
- More than 2 millions document information records (DIRs)
- More than 1 million PDF/A-1b documents stored in our DMS
- A global document storage volume of more than 2.3 TB
- More than 600 GB yearly growth of the storage volume
- New SAFRAN Project: PDF/A Implementation in the SAFRAN group
As described above, the PDF/A-1b conversion today is only implemented in Sagem companies and only for SAP vaults. Based on the positive results with PDF/A at Sagem, SAFRAN Informatique has now started a project to deploy the PDF/A-1b conversion processes to other DMS systems all over the SAFRAN group. This project is currently in the design phase. The following issues are already discussed:
- All Microsoft Office application formats and all PDF documents from any source application or DMS will be taken into account for the future PDF/A-1b conversion.
- The conversion process will be triggered by the user himself through an easy to use GUI or directly from other applications. In both cases the input files for the conversion process will be transferred to a dedicated directory of an FTP server. There the files will be picked up by the conversion processes. After the conversion the output documents will be stored in another dedicated directory of the FTP server. All Microsoft Office files can be sent directly to the conversion server, other applications need to provide PDF files (e.g. CAD by direct export, paper documents through scan processes, …) which will then be checked and adjusted to PDF/A-1b
- The conversion can be launched in real-time during the normal working day and from batch processes during night for mass conversion purposes.
- The converted files shall have the extension “_adj.pdf” to distinguish them clearly from the original standard PDF files.
- Not all PDF files can be successfully converted into PDF/A-1b. E.g. some errors in the metadata section cannot be adjusted, and non-embedded fonts which are not available on the conversion server will create an error. It has to be logged if a PDF/A-1b adjusting process runs into such an error or into a warning, but the general conversion process will not be stopped
- The solution should be based on a mixed infrastructure with some PDF/A-1b conversion servers centralised in SAFRAN Informatique facilities, shared between Paris area companies and small countryside entities, and some servers located in local server rooms for the larger remote companies but administrated by SAFRAN Informatique teams.
- The full PDF/A-1b conversion application will be managed by a SAFRAN Informatique administrator. The administrator will be responsible for the configuration and the distribution of the conversion server and the management of the running environment.
Picture 03: Planned architecture of the future PDF/A solution
As a provider of high-tech equipment mainly for the aerospace and defence industry, SAFRAN needs to guarantee long-term accessibility to all documents which are created and required in the product lifecycle. This long-term accessibility is not guaranteed by the native file formats of the original applications; it requires a neutral data format. Because of comfort features like easy multipage handling, colour support and full text retrieval and because of the availability of cost-free viewers on all major operating systems, SAFRAN decided originally for PDF as the format for digital preservation.
We discovered then that the PDF quality is strongly dependent on the source application from which it is created. Some of our original PDF’s were not reproducible or not reproducible without error. Therefore we decided in 2007 to migrate our existing PDF archive to PDF/A in one of our branches, and to use PDF/A as the new neutral data format for digital preservation. Our experiences with PDF/A in the last 2 years are justifying this decision. With PDF/A we have all of the advantages of PDF format combined with the guarantee of long-term accessibility.
A migration from PDF to PDF/A is not simple. All sources have to be identified and a strategy needs to be implemented on how the different PDF sources can be migrated to PDF/A. Typically there are 2 possibilities: direct creation to PDF/A instead of PDF or creation of PDF with an additional check and conversion process to PDF/A afterwards. Not every PDF file is convertible to PDF/A. However, experience has shown that even a partly adjusted PDF file is better for long-term archiving than the non-adjusted source document. Last but not least, the PDF documents already stored in the archives need to be converted to PDF/A as well. Even when the migration from PDF to PDF/A is a complex project, it is worth doing if a company needs to guarantee long-term digital preservation of documents. With the experience we made with PDF/A at Sagem, we have decided now to implement PDF/A all over the SAFRAN group.