J
James
Let's say I have a stand alone app that was conceived/created decades
ago and has been updated through the years. It's a basic CRUD
application with the data stored in a positional flat file. After
data is entered, calculations are performed and are displayed to the
user. If all is well, the user submits the data for batch processing
and awaits results. During this batch processing, the data takes many
forms before again being sent away for even more processing. I'll
call this the legacy app.
Over the last few years, this legacy app has been accompanied by a web
based product providing most of the same features. Naturally, the
developers fired up their SQL Server and built an ASP based web site
with about 60 data collection forms. After data entry, the data in
SQL is transformed into the same positional flat file as in the legacy
app. Then processing is the same. The site's performance has
deteriorated under heavy load in recent years.
It's been suggested, as a way to improve performance, that we
eliminate the SQL storage and perform the CRUD operations directly on
the positional flat file by way of CGI. I guess the idea is to cut
out the middle man. My suspicion is that this was suggested on a whim
and someone ran with it.
This sounds like a large amount of file I/O and I would prefer to use
ado.net and SQL server but am meeting resistance. I don't mean to
start a flat file –vs- db flame war and if I'm in the wrong group
please let me know.
My question is this…
With about 15k users/day and about 400 pieces of data each, would
anyone really use a flat file as their data storage option, why or why
not?
-Cliff
ago and has been updated through the years. It's a basic CRUD
application with the data stored in a positional flat file. After
data is entered, calculations are performed and are displayed to the
user. If all is well, the user submits the data for batch processing
and awaits results. During this batch processing, the data takes many
forms before again being sent away for even more processing. I'll
call this the legacy app.
Over the last few years, this legacy app has been accompanied by a web
based product providing most of the same features. Naturally, the
developers fired up their SQL Server and built an ASP based web site
with about 60 data collection forms. After data entry, the data in
SQL is transformed into the same positional flat file as in the legacy
app. Then processing is the same. The site's performance has
deteriorated under heavy load in recent years.
It's been suggested, as a way to improve performance, that we
eliminate the SQL storage and perform the CRUD operations directly on
the positional flat file by way of CGI. I guess the idea is to cut
out the middle man. My suspicion is that this was suggested on a whim
and someone ran with it.
This sounds like a large amount of file I/O and I would prefer to use
ado.net and SQL server but am meeting resistance. I don't mean to
start a flat file –vs- db flame war and if I'm in the wrong group
please let me know.
My question is this…
With about 15k users/day and about 400 pieces of data each, would
anyone really use a flat file as their data storage option, why or why
not?
-Cliff