management >> Custom Solution - Database vs. NFS

by seijin » Tue, 05 Feb 2008 07:19:47 GMT

Howdy,

In building a document management solution, what do y'all think about
storing the scanned images in a database versus saving them in a
directory on the file system and storing a link to the file in the
database. I can certainly see where the database is needed to store
relevant textual information about the scanned image but I'm wondering
about the image itself.

I figured y'all would be one of the more knowledgeable places to ask.

management >> Custom Solution - Database vs. NFS

by billw@rjssoftware.com » Tue, 05 Feb 2008 21:23:38 GMT



I would highly recommend NOT storing the actual image file inside the
database with the search meta data. This is typically known as
"blobbing" the data and image together. The reason that I typically
recommend this is that if you are purchasing a solution for your
business and you go the blob database route, if you ever have to
migrate away from the imaging system becuase of fit, the software
provider going out of business or the product ending life and
support...then you would need to be able to migrate the images and
database information off to another system easily. Blobbing the
database with the images makes it very difficult to migrate away this
type of system and usually requires that you purchase a SDK or
migration toolkit from the software provider since they typically make
the data a little more proprietary or locked down. If you
could...purchase or build a solution that is database driven (SQL,
DB2, mySQL or otherwise) and then store the actual images on Optical
(for legal purposes), NAS, SAN, or on board DASD storage.

If you're looking for commercial options for document management
solutions...feel free and contact me. I work for a organization that
develops software for this market.

Bill Whalen
RJS Software Systems, CDIA+
952-898-3038
http://www.rjssoftware.com

Similar Threads

1. Custom SNMP monitoring solution versus commercial purchase

2. IPcop and dyndns.org custom dns a solution!

Maybe some of you tried to configure IPcop's Dynamic DNS updater to
update your dyndns.org custom DNS domain yourdomain.com and noticed
that it's just impossible to do since IPcop's Dynamic DNS web
interface just don't accept an empty HOST field.

After many experimentations I finaly found a solution. You will have
to edit your IPcop Dynamic DNS script file to accept a DOMAIN field
without a ".". Use WinSCP (or any clients) and find the file
/home/httpd/cgi-bin/ddns.cgi and comment the following line (put a #
sign in front of it) :

unless ($cgiparams{'DOMAIN'} =~ /[.]/) { $errormessage = $tr{'invalid
domain name'}; }

Now the IPcop web interface will accept a DOMAIN field without a "."

Exemple to configure yourdomain.com :
In the IPcop DDNS web interface enter:
SERVICE: dyndns-custom
HOST: yourdomain
DOMAIN: com
USER: your username
PASS: your password
ENABLE WILDCARDS: Y

Now IPcop will be able to update yourdomain.com

3. IPCOP and dyndns.org custom domains a solution

4. user space nfs vs. kernel nfs

Hi,

does the kernel nfs server support any method of user id mapping? I need to 
export a certain directory from a cluster node to a individual linux box ouside 
of the cluster with full rw access. Unfortunately the UIDs of the same user 
differ on both systems.

I know, that the user space nfs server has such a ability via NIS based user 
mapping like this:

/fs/user        host.foo.org(rw,map_nis=domainname)

in /etc/exports.But of course this works not with the kernel nfs server.

So what can I do to with the kernel nfs to map the UID?

Thanks for your comments

Regards,
Mischa

5. [NFS][2.4][ReiserFS] NFS and `nohide' vs. reiserfs - Linux

6. [PATCH 1/2] NFS: Manage NFS modularity vs FS-Cache modularity

7. + git-nfs-vs-nfs-convert-to-new-aops.patch added to -mm tree - Linux

8. Found Solution - Frontpage Master Site opening vs. Subsite

I discovered the issue.  If the option for "Open the last Web site 
automatically when FrontPage starts" is checked, and you happen to open 
the root web site and close FrontPage, there is no way to get back to 
the subsites without unchecking that option.  With that option 
Unchecked, it works like it is supposed to.  With the option checked, it 
keeps the root site stuck on open, even if you close the site and try to 
open another one.

I believe this is a bug because you should be able to open any site 
after closing the active one.

At least I figured it out.  Thanks for opportunity to post.  Is there a 
way to send this along to the MS development team for a hotfix?

Thanks,
Rich.

Rich wrote:
> Hi,
> 
> Have a strange situation.  We have 4 developers that are all working on 
> parts of a web site (say... www.mydomain.com).  If I open a site under 
> that site that is in another folder (say we have a department folder 
> under that site) that sites subfolder opens in mine ok.  So if I open 
> the site under the department folder called IT, then my path is 
> \\www.mydomain.com\department\IT.  Thats what show up in my Website 
> local web site path.
> 
> I get the IT site with all the navigation info fine & intact.
> 
> If another person opens the IT site after I'm out (not checking out any 
> of the pages), that person selects the IT site but opens up at the 
> \\www.mydomain.com main site instead of the IT site, even though they 
> opened it at the subsite level.  Their Website local web site path shows 
> the main site (not the department\IT part).  The folders on the left can 
> be navigated to the IT folder, but it doesn't maintain the other 
> information like it would if the site itself was opened.
> 
> I hope I'm making sense with this.  Because that person can't get the IT 
> site level, their is no navigation information showing.  The links for 
> the navigation don't show up, so when they publish the page for IT, it 
> loses that.
> 
> We have 2 people that it works correctly for and 2 that it doesn't. 
> We've upgrade all machines to the SP3 for Office 2003, so they match.
> 
> Any Ideas?
> 
> Thanks,
> Rich.