Blog
No SQL Server. No Oracle. No DB2. No waiting. And no Cost.
Don’t get me wrong. I’m not bashing any of the above databases. They are all great database systems, and excel in their field. My only point is that for archiving documents neither of the above is the best option.
Serious license fees
One minor detail is the fact that when you outgrow the free entry versions you’ll be hit by some serious license fees. License fees that in some cases exceed the investment you have made in the complete document management solution.
Heavy hardware requirements
Running a document management system with millions of documents and hundreds of user based on traditional database technology requires several big servers with lots of disk arms and giant internal memory.
Sub-par performance
Even with some heavy iron in place you will experience challenging performance. Loading huge batches of document into a database simply takes time.
All of this is not caused by SQL Server, Oracle, DB2, or MySql being inferior products. By no means. They are simply designed for a completely different purpose.
Document management is by nature accumulative. Of course you change or delete documents, but the vast majority of documents you simply capture, store and later share. This is not the process databases are optimized for.
Haystack
That is the reason why we with Next went for an append-only storage engine based on Haystack.
Just as Facebook ™ did for their giant photo storage. And that is why Next outperforms any other document management solution on comparable hardware.
Read about our 4,600,000 documents/hour capture record on a single simple server running Windows.
Besides excellent performance our design in Next also gives us unmatched revision control, as all document versions are kept in the storage engine. No more Enron with documents gone missing.
Saving my clients tens or hundreds thousands euros on unnecessary database software is then only an added benefit.
Running with SQL Server, Oracle, DB2 and MySql
Tens or hundreds thousands euros on unnecessary database software
As I opened this blog entry by stating, I’m not bashing the databases. In fact we integrated tightly with data in SQL Server, Oracle, DB2 and MySql to validate and enrich your document metadata during capture. And in production setups of our BPMN workflow engine we are happy users of your preferred database.
My two cents. Use the right tool for the job.
