DisableBind and other ways to make database faster

[Former Member]

I have an application that connects to various databases: SQL Anywhere, Sybase ASE, SQL Server, SQL Azure, Oracle, MySQL.

It has both inline and DataWindow queries and updates. One of the tables has a blob column.

On some of the connections I set DisableBind=1 but I am not really sure if I should be doing it. Also, sometimes in a database trace it will show the following:

(105c4878): BIND SELECT OUTPUT BUFFER (DataWindow): (0.018 MilliSeconds)

(105c4878): ,len=6,type=CHAR,pbt=1,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=51,type=CHAR,pbt=1,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=44,type=LONG,pbt=22,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=256,type=CHAR,pbt=1,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=44,type=LONG,pbt=22,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=31,type=CHAR,pbt=1,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=30,type=DATETIME,pbt=9,dbt=0,ct=0,prec=0,scale=0

(105c4878): ,len=256,type=CHAR,pbt=1,dbt=0,ct=0,prec=0,scale=0

Is this caused by the Bind setting and is it something that should/could be eliminated? Does it only get executed for DataWindows or will a inline cursor query do it as well?

Basically I am looking for somewhat non DBMS specific ways I can speed up the app before I look into adding indexes.

It is already very fast for in-house servers but when run over the internet I need it to be as fast as possible so every millisecond counts.

[Former Member]

Hi Roland,

I don't have the answer.

Knowing that you have multi-threading down to a fine art and inspired by smartphone apps that work with asynchronous server calls informing the user as appropriate; I would create a client data handling tier.

Just a thought.

Lars

[arnd_schmidt]

When your client is already "very fast", it seems that this is a bandwidth / latency problem.