Hadoop Document Transformations Using Adlib

In our series exploring the use of Hadoop for ECM, the best practice from our years of ECM experience tells us is that documents should be stored in both their native content as well as a PDF rendition of the content. Storing a PDF rendition allows consumers quick access to view the content, as well as being able to watermark and control the content to prevent consumers from altering the documents. This post will explore TSG’s partnership with Adlib and how we are using Adlib’s PDF conversion suite to transform documents being stored in Hadoop.

We have had a few conversations with clients about using Hadoop as an ECM platform, and there have been a few unique business scenarios that we could leverage the Adlib PDF Conversion suite. Adlib PDF delivers an enterprise-class, document-to-PDF conversion framework that offers the highest fidelity PDF rendering engine on the market, with accurate OCR capabilities.

The first scenario involves a large legacy system of scanned TIF and PDF files that are currently stored on a legacy system that makes it very difficult to find anything. There are over a million documents spread across many folders that are not full-text searchable, and are sitting around without adding any value to the business. By leveraging Adlib’s OCR capabilities we can:

  1. Read in the scanned TIF/PDF files from the legacy system
  2. Run the scanned files through Adlib’s OCR engine to extract the text from the document
  3. Migrate the documents into Hadoop using OpenMigrate
  4. Index the documents using Solr to allow for fulltext searching on the contents of the document

With the documents having been transformed by Adlib, stored in Hadoop, and indexed by Solr, the business can now leverage these documents to perform meaningful analytics on the documents.

The second business scenario involves a topic we have previously blogged on, which is running your ECM against Hadoop instead of a traditional RDBMS. In the case of storing documents in the ECM repository, best practice is to store the native content for authors, and PDF renditions for consumers. Since TSG leverages our OpenContent web services layer for all interaction with Hadoop, we have implemented the Adlib transformation engine to request PDF renditions of all documents when they are added to Hadoop. By leveraging the robust Adlib transformation suite that offers best in class transformations of over 400 file types, putting these documents in Hadoop with a PDF renditions ensures that all users will be able to view the documents being stored in the repository, regardless of the native application that was used to author the document. This is a very important consideration when leveraging Hadoop as your document repository that can be easily overlooked since ECM vendors have been transforming to PDFs for years:

  • Users expect mobile access to documents, and PDF is the only format every device manufacturer can reliably display
  • On PCs, there is no guarantee that everybody is going to have the correct version of every piece of software on their machine just to be able to view the document
  • PDFs are the industry standard for long-term storage of documents

The overall architecture for transforming everything to PDFs allows us to leverage the Transformation abilities of Adlib to properly store any type of content in Hadoop.

Hadoop with Adlib

We see Adlib as the perfect marriage with Hadoop to ensure proper storage of documents in Hadoop. Let us know how you are leveraging Hadoop in the comments below.