Home of the ColdFusion Community
|
Mailing Lists
|
Home /
Groups /
ColdFusion Talk (CF-Talk)
webservice limits
Author: Douglas.Knudsen
Short Link: http://www.houseoffusion.com/groups/cf-talk/thread.cfm/threadid:32240#161791
good info....danke
Doug
On Apr 30, 2004, at 7:33 AM, Douglas.Knudsen@alltel.com wrote:
----- Excess quoted text cut - see Original Post for more -----
There have been several recent threads about thinArrays, Javascript
Web Services, Neuromancer, etc. that indirectly address this.
The basic problem with a web service returning a large record set is
that the amount of WDDX or XML overhead often runs 250% of the amount
of actual data. This is a steep price to pay in bandwidth and
serialization/deserialization processor cycles.
In the application that you describe, the structure of the data is
known in advance by the consumer, and there is little need for the data
to be read by humans. Therefore, WDDX or XML is overkill.
What thinArrays attempt to accomplish is to exchange this type of data
with as little overhead as possible.
The record set is converted into a simple string with field separators
adding about 13% data overhead (as opposed to 250%). So. the actual
WDDX or XML packet contains only one field and associated tags.
Serialization and deserialization at the server are as faster or faster
than the equivalent XML or WDDX process.
Serialization/deserialization at the client are much faster than the
equivalent WDDX or XML process -- a single Javascript split command is
used to deserialize the thinArray into a Javascript array representing
the original record set.
That said, in the past I have often used WDDX to manipulate databases
on remote hosts, including copying entire (rather large) databases from
one remote host to another remote host (no admin access to the db at
either host). This was before the advent of web services but the
packet structure is similar.
The above performed surprisingly well.
But, If I were to do it today, I would use web services and thinArrays
to minimize the performance hit.
HTH
Dick
_____
Author: Dick Applebaum
Short Link: http://www.houseoffusion.com/groups/cf-talk/thread.cfm/threadid:32240#161788
On Apr 30, 2004, at 7:33 AM, Douglas.Knudsen@alltel.com wrote:
----- Excess quoted text cut - see Original Post for more -----
There have been several recent threads about thinArrays, Javascript
Web Services, Neuromancer, etc. that indirectly address this.
The basic problem with a web service returning a large record set is
that the amount of WDDX or XML overhead often runs 250% of the amount
of actual data. This is a steep price to pay in bandwidth and
serialization/deserialization processor cycles.
In the application that you describe, the structure of the data is
known in advance by the consumer, and there is little need for the data
to be read by humans. Therefore, WDDX or XML is overkill.
What thinArrays attempt to accomplish is to exchange this type of data
with as little overhead as possible.
The record set is converted into a simple string with field separators
adding about 13% data overhead (as opposed to 250%). So. the actual
WDDX or XML packet contains only one field and associated tags.
Serialization and deserialization at the server are as faster or faster
than the equivalent XML or WDDX process.
Serialization/deserialization at the client are much faster than the
equivalent WDDX or XML process -- a single Javascript split command is
used to deserialize the thinArray into a Javascript array representing
the original record set.
That said, in the past I have often used WDDX to manipulate databases
on remote hosts, including copying entire (rather large) databases from
one remote host to another remote host (no admin access to the db at
either host). This was before the advent of web services but the
packet structure is similar.
The above performed surprisingly well.
But, If I were to do it today, I would use web services and thinArrays
to minimize the performance hit.
HTH
Dick
Author: Douglas.Knudsen
Short Link: http://www.houseoffusion.com/groups/cf-talk/thread.cfm/threadid:32240#161770
Ok, so I'm trying to get my org here to use webservices and all. I keep getting
asked about setting up webservices like a DB feed. My usual response is that
webservices are for lightweight data, but I'm wondering. What are the limits of
webservices? Should they be used for large data sets?
I played around and created one that returns a 15,000 by 25 column data set. It
took about 2 mins.
Cheers!
--
Douglas Knudsen
Alltel ACI Rapid Response Team
"To fear love is to fear life, and those who fear life are already three parts
dead." Bertrand Russell
******************************************************************************************
The information contained in this message, including attachments, may contain
privileged or confidential information that is intended to be delivered only to
the
person identified above. If you are not the intended recipient, or the person
responsible for delivering this message to the intended recipient, ALLTEL
requests
that you immediately notify the sender and asks that you do not read the message
or its
attachments, and that you delete them without copying or sending them to anyone
else.
|
May 24, 2012
|
Latest Fusion Authority Articles
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
