+++ Andrew Michael Lynn [19/06/08 14:26 +0530]:
>
>Requirements: Collaboration, Content-management, an Open Lab note-book,
>project management and workflow 
>
>1. OS
>Solaris 
>Enterprise Linux


That should have been "Redhat Enterprise Linux" (Redhat is not the only
entity making "enterprise" grade linux :-P ).

Is commercial support from Redhat/Sun going to be bought? If not, then
Debian would be a good rock solid alternative too. Freebsd would be as
good too.

Business angle:
Solaris/Opensolaris vs. Linux: 
1. How expensive is paid tech support?
2. How much support is typically needed in any of these options.
3. How much difficult is it to get admins for these options? And retain
    them?
4. As a consequence, how easy is it to get good documentation? And good
    community support?

>
>2. Application server
>GlassFish
>JBOSS
>Apache Tomcat
>Apache

Is the Tomcat/Jboss java stack being looked at only because of the JBPM
workflow tool below? By going in for a Java based stack has multiple
effects, technical as well as business wise:
1. Your hardware requirements increase dramatically.
2. Your LAMP options reduce dramatically. There is no sense running a
    LAMP application running on a machine running a Tomcat instance which is
    heavily used.
3. The business aspect - how easy is it to get Java talent from the
    market place and retain them as compared to LAMP? Whatever application
    you spend your valuable money on, will lock you into the platform that
    you choose. You can port between Unix OS, but you can't between
    languages later without significant investments. Are you willing in
    going that path?


>
>3. Portal Engine
>Sun Open Portal
>LifeRay
>
>Note: Most users/developers of computational biology applications
>web-enable their applications using a LAMP stack. Q: Can similar portal
>functionalities be provided by Joomla/Drupal/etc  ?

Yes, You should also consider FOSS options like Plone, Typo3 which has
workflow login built in.



>
>8. Data Warehousing (Do we need this?) The only suggestions in the first
>cut were proprietary solutions. 
>Data Warehouse – GreenPlum
>Data Integration/Extraction Tool – Informatica
>Design/BI - Microstrategy

If you are stuck with these options, look at the options of interfacing
with them using LAMP too.

>
>10. Workflow
>jBPM is built into JBOSS. 
>The taverna project (workflow developed for computational biology)
>



IMHO, my perceived biggest problem in organizations like CSIR is 
vendor/platform lock-in which causes expensive sustainability issues 
later. Having options which are nimble are as important as options which
are so called "enterprise grade".

The other problems are talent and support. Recruiting and retaining good Java
programmers/admins are a nightmare for commercial organizations in the
first place. I cannot imagine what it will be for organizations like
CSIR.

But again, it has been a while since I worked with such organizations.
Things might have got better since. :)

- Sandip

-- 
Sandip Bhattacharya
http://blog.sandipb.net

_______________________________________________
ilugd mailinglist -- [email protected]
http://frodo.hserus.net/mailman/listinfo/ilugd
Archives at: http://news.gmane.org/gmane.user-groups.linux.delhi 
http://www.mail-archive.com/[email protected]/

Reply via email to