Legacy ECM and Enterprise Search
Federated Search Approach
Migration of legacy system is one of the need of our big data giants due to reason below;
- No Api
- Old database
- No support
- Performance
- Retention old policies
- Restructuring old structure
And due to digital transformation applications or legacy Enterprise Content Management Systems (ECM’s), Federated Search has been a very often used solution among content management giants. If we want to build a new system that accesses the contents of the existing ECM and we don’t want to use an extensive workforce to migrate those Legacy into sweet and easy to use ECM. And due to legacy reasons it was thought many times whether to use federated search. Those reasons can be;
- Legacy users
- Legacy contents
- Legacy integration and
- Legacy technology
Federated search has been used to solve all the problems above adding interoperability and transformation of legacy systems and building new systems. As example, in a bank their mortgage is on Documentum, contracts are on IManage and account is on Open Text. Federated Search can make it possible for loan teams to be able to show loans from Documentum and contracts within IManage by connecting the new system as done by many new ECM vendors.
Is that soooooooooo easy?
It is fairly easy to say by marketing wing of a ECM consultants that why should we move migrate the old system if we can search by federated query and it is very easy to query certain system, extracts the result and finally merged, why do we need unnecessary additional index, however there are cons.
Cons:
- Performance
- Merging
- Security
- System Fault Tolerance
- Logistics
Hence, depending on ECMs to ECMs and their data, their patience of compromising the performance and juggling capacity with security, some one can still praise the Federated search or they can research another approach.
Publishing Approach
In terms of querying the legacy ECM directly and merged back, there is another approach that I am going to recommend is, Publishing approach. Same theme where businesses want to access the data but hesitant to replace those legacy systems. Publishing approach builds and parallel systems where it pulls the data of multiple systems in extra storage index to search and any ways these days storage is progressively cheaper.
In this approach a messaging system can push the documents and metadata on topic for search whenever a new document creates into legacy and a parallel consumer can pull those data from the topic. With this pull and pull the new repository will have metadata and copy of the document as well.
By this way, businesses can still maintain their old system to store their new data as well as can provide the shared search repository to search data for other systems while integration.
So publishing approach gets edge over federated approach in following areas.
- Integration
- Performance
- Security
- Data Format Enhancement
- Can be stored into multiple repositories.
I will recommend using Solr/Lucene to use a publishing approach, however choose your own convenience.
Cheers! Happy Searching :)
