: 4582 | 108646 | 12353

Let's consider what DB2NSF does for us. 
Use this IdeaSpace to post ideas about Domino Server.

: -8
: 2
: 10
: Domino Server / Other
: db2, nsf, nsfdb2, dbsnsf
: Craig Wiseman21821 03 Dec 2007
: / Email

I like to stir the pot,and since one of the points of IdeaJam is to "gauge the marketability, popularity and viability" of N/D things let me propose this.

In a lot of ways DB2NSF seems like a holdover from the Dark Years when the goal was (seemingly) to upgrade the full functionality of a Domino server into the DB2/WebSphere stack.

DB2NSF in it's current incarnation has...issues. It's very picky about what DB2 you run and on what platform you run it on. It takes a whole lot of infrastructure and significant DB2 expertise to set up & run, it's not any faster than NSF for NSF-y things, and it seems to me that the SQL things it brings could be added in a much more efficient fashion to the NSF database directly. For example, NotesSQL does many of the verbotten things (joins, etc.) somehow and works quite well for reasonable datasets.
It seems to me that the development effort being spent on recreating NSF functionality could be/should be better allocated elsewhere.

So the idea is to add the basic SQL functions that DB2NSF brings directly to the NSF.

1) Corey Davis1744 (03 Dec 2007)
Is this different from Sean's idea? { Link }
2) Craig Wiseman21821 (04 Dec 2007)
@1 - Somewhat, but much broader in impact. It seems to me that the choice is either to spend the effort & dev time to re-develop and support NSF functionality in DB2 or add and support basic SQL functions to NSF.
Seems clear which path would take less dev & support effort for IBM and less effort and support for customers.
3) Slawek Rogulski8874 (04 Dec 2007)
@2 - This somewhat reminds me of Garnett and how it was going to be integrated into Domino.
4) Axel Janssen5023 (04 Dec 2007)
I've allways wondered which business value does dbsnsf aims at.
1. better performance for large data sets? (I am not sure if this is feasible)
2. support for rdbms transactions if the domino data is accessed transactionally from a platform which supports transactions? (might be great for certain use cases) .
3. easier integration with reporting tools (yup)
I think for point 3 it should be easier (at least I would guess) to have the data in an rdbms as the tools are geared to it, at least traditionally.
On the other hand, thinking Eclipse based, it might be possible to let BIRT directly operate on the nsf data. BIRT is a eclipse subproject for reporting. I played with it using relational database as data source and I kind of like it (am not the only one). There is a composite app sample in notes.net sandbox using BIRT, though I don't know how it accesses the data.
5) Craig Wiseman21821 (04 Dec 2007)
@3 - Are you sure Garent's history? { Link }
6) Slawek Rogulski8874 (05 Dec 2007)
@5 - Sorry, I I didn't get it entirely correct. My bad.
I would be happy if we could have computed views like CouchDb. As for relational storage I would love that but isn't it like trying to get cats and dogs to play nice in one pen? While additional infrastructure of the current solution is cumbersome I think it is a reasonable, not ideal, compromise.
7) Craig Wiseman21821 (05 Dec 2007)
@6 - A interesting point, but I'm gonna have to disagree. You CAN make NSFs do relational-like things. The bear is getting DB2 to do the NON-relational things that NSFs do easily.
Seems simpler to get the more flexible store to to be more rigorous than vice-versa. Plus, it's got to be way easier to port across platforms. DB2NSF at least an order of magnitude more complexity to an already complex product to gain benefits that are more easily gained through other, more direct, means.
My opinion, YMMV.
8) Slawek Rogulski8874 (05 Dec 2007)
@7 - If I got to replace the Domino store then I would put in a relational store and provide OR layer. As much as possible automate the generation of the OR layer. But allow it to be fine tuned. So your Domino app still sees documents not tables.

One storage engine means a compromise. Two separate specialized storage engines means additional complexity in setup and maintenance. Mind you making one type of storage appear as two is not trivial either although the complexity is hidden under the covers.

If I had to choose one it would be the relational storage. It does relational, obviously and can be made to look like an object store via an OR layer. This intermediate layer which domino would interface to for current style apps would need to be produced automatically as much as possible based on defined forms and perhaps some wizard type interface to collect the specification. Then based on the OR layer tables can be created automatically.


Welcome to IdeaJam™

You can run IdeaJam™ in your company. It's easy to install, setup and customize. Your employees, partners and customers will immediately see results.

Use IdeaJam to:

  • Collect ideas from employees
  • Solicit feedback and suggestions from employees and customers
  • Run innovation contests and competitions
  • Validate concepts
  • Use the power of "crowd-sourcing" to rank ideas and allow the best ideas to rise to the top

IdeaJam™ works with:

  • IBM Connections
  • IBM Lotus Quickr
  • Blogs and Wikis
  • Websphere Portal
  • Microsoft Sharepoint
  • and other applications.

IdeaJam has an extensive set of widgets and API's that allow you to extend and integrate IdeaJam™ with other applications.

Learn more about IdeaJam >>

IdeaJam developed by

Elguji Software Logo