As summarized in our series of migration posts, many of our clients are moving to Alfresco from legacy ECM solutions. For one of our current clients, we are moving from an onsite Canon ImageWare solution to Alfresco hosted within the Amazon Cloud. This post will share our experience migrating from ImageWare to Alfresco and include some specifics in regards to technical migration nuances.
Canon ImageWare – Why Migrate Now?
Our client has been using ImageWare since 2006. Originally, the solution was implemented as a simple document management tool to capture MS Office, PDF, and image documents. Some nice features included the ability to “print” from any application directly to ImageWare. ImageWare has been very successful within the client and has grown as more and more business processes were added including contract management, accounts payable, and tax certificate processing.
ImageWare presents a very basic interface similar to Windows Explorer. It allows users to set up cabinets and folders for different processes and functions. Implemented as more of a generic tool, ImageWare usage evolved to include specific business processes. ImageWare allows for some minimal configuration but most of the business specific logic (how to name documents, where to store) was left up to the business users to enforce.
For our client, the driving factor to move from the legacy system was that the ImageWare client would only run on legacy software – particularly Windows XP and that the ImageWare server could not be run in a clustered high-availability environment. To access the ImageWare system, users had to remain on a Windows XP machine or be provided access to a separate virtual machine running XP. This created a security problem as well since XP is no longer supported by Microsoft.
As our team started explaining how other clients use Alfresco, the client found more user and business driven factors to move from the glorified file cabinet approach offered by ImageWare. The ability to use workflows, advanced searching, and relate document types together by customer will improve the efficiency of their contracts process.
Migration Strategy
Starting with the need to move off ImageWare, TSG assisted the client with developing a migration strategy for both the documents and indexes (meta-data). In our discovery process we found several document specific items that required attention:
- Annotations – were stored in a proprietary Canon format in TIFF (fairly common) and would need to be preserved in the new Alfresco system
- Large file sizes – users had been concerned for years about system performance, particularly in remote locations. During our review, we determined that, while the “print to TIFF” was an interesting technology, it was resulting in uncompressed files, often times 80 times larger than a similarly compressed PDF.
- Document indexes – were unique to each cabinet in the ImageWare system and not always populated. A set of rules was developed to parse meta-data from the folder where the document was filed.
- Interface – In our discussions with the client, we presented both Alfresco Share as well as our HPI interface as options. Given the business focus of the client on documents and case management, the client chose HPI for the ImageWare migration.
Once these technical details were known, we quickly filled in the rest of the picture with our migration best practices.
Converting ImageWare TIFFs
Many of the files stored in ImageWare over the years had been saved using the “print” functionality. This created large TIFF files for emails and other documents. As the number and file sizes increased, the user experience and system performance decreased. To improve performance for the users and save storage, the migration process included a step to convert large TIFF files into PDFs. In some cases the files are being reduced by 40x to 100x. One sample test reduced the aggregate size of a set of files by 220x. We are anticipating that the reduced file sizes (combined with a modern ECM infrastructure) will dramatically improve performance without any loss of resolution or capabilities.
Preserving Annotations
Another critical point of the migration effort is preserving the annotations users had made in ImageWare. After some R & D into Canon’s annotation format we selected a process that would preserve the annotations as well as stage the large files for further processing. Utilizing the export functionality of the ImageWare client, annotations were “burned” into the file and then carried forward into Alfresco in PDF Format.
Reading Document Meta-data from the Database
The final complexity to the migration effort was getting the document index or meta-data from the ImageWare system. Unlike many of the other ECM systems we work with, ImageWare has an unconventional database structure. Our approach implemented a complex query to work backwards from the document name and folder location to find the meta-data.
Interface –HPI
The main business focus on the ImageWare system was the processing of contracts from the proposal through the approval process. This process focuses on a variety of documents including proposals, purchase orders, insurance certificates, tax certificates and invoices that need to be added to a folder until the contract is approved/executed. We found this process similar to our insurance clients and very consistent with what HPI can do with minor configurations. The users like that the interface has moved to a “business focus” with rental contract folders with specific meta-data.
Using Amazon Web Services (AWS) for Cloud Deployment
The final solution for the client will be built in the Amazon cloud and utilize the redundant database (RDS) and S3 storage capabilities to enable a high-availability environment without the need for a large amount of infrastructure management.
The modern architecture and user interface of Alfresco will reduce the performance and document indexing (attributing) errors encountered in the current system. Adapting Alfresco into the everyday tasks of the users will require minimal training since it is more focused on how the users do their tasks instead of being a “do it all” type of interface. You can read more in our post on how Alfresco simplifies ECM for users.
Summary
Overall, we found the migration from ImageWare to Alfresco fairly straightforward but involved some complexity in converting TIFF files with annotations to PDF. Our solution focused on reduced file sizes combined a modern browser interface with HPI all hosted in a robust environment in the cloud. The project is on schedule to complete in 3rd quarter 2014.
We’d like to hear your thoughts on this solution in the comments.
[…] simply procuring the necessary temporary space for the dump itself. As highlighted in an earlier post about a Canon ImageWare migration, the process required a dump and load approach since Canon did not supply the system API. Due to […]