Number/size Of Packets At A Network Level For A MSSQL Query Vs Oracle Vs DB2.
Jul 24, 2007
Hello all,
I will try to supply all of the information available to me to help generate some answers to the question I have. I am doing some benchmark testing of an application of how it deals with various database queries, for example oracle 10g, DB2 and MSSQL 2005.
I have a very simple database called "student" that has various tables, with various columns, all as mentioned very simplistic. Each table has a few hundred rows of data, the most being one table with 2000 rows. The same database (structure and dataset), is replicated across all 3 DB's.
In order to do the testing, I have been recording the same queries (select * from table_a, select * from table_b etc etc) for each database at an interface which monitors traffic between the client and db server, the problem starts here; The size of the tcpdumps are vastly greater for MSSQL vs oracle or db2. Some examples:
100 query or transaction dump:
mssql=3.7mb
oracl10g=133kb
50,000 query or transaction dump:
mssql=1.8gb
oracl10g=0.6gb
db2=0.6gb
as mentioned each database's tables and data are identical to my knowledge and all queries are the same. The bloated dumps are making mssql performance numbers look bad.
So my question is: What could be the reason for MSSQLs client/server queries to contain so much information, has anyone else come across this? Is there any setting or something I could try to minimize it?
Does anyone know where there is a good summary of what you can do to encrypt data going between SQL Server and it's clients, and for that matter any other security things you can do with SQL Server ? I've been asked to write a document on this for someone and quite frankly I've got too much proper work to do...
Hi,I am getting the following error (When i am trying to update the column thru windows service) Please let me know the solution for the following.The following error is comming only when i set CommandTimeout to infinity (Commandtimeout=0).General network error. Check your network documentation. Number 11 Procedure ConnectionRead (recv()). Class 20 State 0 Source .Net SqlClient Data Provider Server Line number 0 Thanks and Best RegardsNagaraju A
My group has recently begun ramping up #of users from 5-10 to 50 concurrent. I'm currently using a 10Mb/s Ethernet but am not sure if there is a fiber or multiplexing solution that I could use to increase output. Basically, we're looking at up to 50 heavy duty data pulls (50MB+) at any one time.
Hello,I was wondering is the following is correct with the network packet size.I've installed sql server 2000 and made some changes in the configuration.Now i've read that sql server's default network packet size is 4096 bytes,but when i look up the advanced tab, in the optionsmenu of the enterprisemanager, the networkpacketsize is 0!Now i see that some connections are memoryconsuming, could the above hassomething do with this and should i change the network packet size.I would like to have some help with this.Thnx
Hi, I am using an Enterprise Manager and i have created a Database Maintainance Plan, can anybody guide me as to how to create backup on a remote machine.
I also mapped that device(say M/c B) onto my server box(M/c A) but i am not able to see that particular location in my backup activity screen.
We have an application that runs across the WAN to multiple locations. Performance is poor and we are looking at ways to improve performance. One suggestion from our Sr. Network Administrator is to change our Network Packet Sizes across all points, SQL Server & NIC to match the outgoing Router. This would be a size of 1440.
Does anyone have any thoughts or recommendations on this?
I am having a problem working this DTS packet. I have a local database table that takes in data and at a certain time of the day I have a DTS packet to run and insert into a remote SQL server. I have no problem running the DTS packet, but it is inserting duplicate data into the remote SQL server. How would I go about checking to see if the data already exits on the remote SQL server? Any help would be greatly appreciated.
hello, I built my application using SQLClient classes and MSSQL database i have another version of the database with Oracle, is changing to OracleClient classes the only change i have to do in order to run the program and the connection string? Thanks
Hi, I have to migrate the oracle SQL statement to MSSQL. I want to know how to write a sql which can retrieve: 1. the user table name 2. the user table column name in the user table 3. the type of the user table column
I have already done a Oracle compliant sql : ------------------------------------------------------------ select table_name, column_name, decode(data_type, 'VARCHAR2', 'java.lang.String', 'VARCHAR', 'java.lang.String', 'CHAR', 'java.lang.String', 'DATE', 'java.sql.Timestamp', 'NUMBER', DECODE(nvl(data_scale,0), 0, 'java.lang.Long', 'java.lang.Double'), 'LONG RAW', 'java.lang.Byte', 'BLOB', 'java.lang.Byte','java.lang.Object') from user_tab_columns ------------------------------------------------------------
Can anyone tell me how to change it to MSSQL sql which retrieve the same value as the above sql. Thanks.
i have a doubt here that would like to have some advice from u all. Can i get data from Oracle 10g database if i have an application running on MS SQL ? If can, would it be cumbersome afterall? Please advice.
hi all ms sql specialists.....i am a newbie on ms sql buti am used to oracle and toad ...and now a question (i guess this was asked xxx times befor but icouldn't find the answer easilly) :is there something powerfull like toad on oracle for mssql server? (notenterprise manager)
When I'm getting data from sqlserver using ado.net and a sqldataadapter, are the resultant network traffic packets secure? If I wanted to deploy my objects at a remote site, would I still be safe going straight to my sql server from there or should I build a web service and then auto generate 'remote' versions of my objects that will then communicate to the web service on https?
Hi all, Well, I have created an applicatin in asp.net and it connect to MSSQL database, and it work fine.
There is another application which was developed in Oracle and the database is oracle 8i.
Well, wat i want is whenever a form is submitted from the asp.net application, the data is inserted in the MSSQL server database and the Oracle8i database also.
Im working on a project to convert a C program from OCI(oracle) to ODBC(MsSQL).
One of the select statements uses oracle julian time:
"Select TO_NUMBER(TO_CHAR(TO_DATE('0101'||:styear,'mmddyyyy'),'J')) STYEAR from dual"
Its not a problem converting the statement, except msSQL uses 10/01/1900 as its origin for the julain calendar, where Oracle uses somewhere around 5000BC. my questions is, because this program is made to use oracle julian and need to keep it that way, how i can make the SQL conversion using oracles origin date?
Hi,I am using MSSQL 2k, and I have a linked server set up to an Oracle RDBversion 7. It goes thru an OLE DB provider for ODBC drivers on a systemDSN, which is using an Oracle RDB ODBC driver version 3.0.2.The problem occurs when I send a query that returns zero rows - queryanalyzer just does not complete nor return. This problem is not seenwhen there are rows being returned.I ran a trace and this is the error message I get-:Non-interface error: OLE DB provider MSDASQL returned an incorrectvalue for DBPROP_CONCATNULLBEHAVIOR which should be eitherDBPROPVAL_CB_NULL or DBPROPVAL_CB_NON_NULLAnybody who experienced this before has a solution?Thanks,Lawrence
I am currently migrating my database from oracle to MSSQL. My problem is that i have encountered the UTL_FILE function in Oracle and i am not finding its equivalent for MSSQL.
Can anyone help me with this plz? Is there any other way of reading and writting OS files in MSSQL?
I created a link from Oracle to SQL server. There is a trigger with insert statment in Oracle side. I got error message as follow when trigger is invoked:
SQL> insert into test1 values ('zerbra','brazi'); insert into test1 values ('zerbra','brazi') * ERROR at line 1: ORA-02047: cannot join the distributed transaction in progress ORA-06512: at "GGWEB.TRIGGER2", line 4 ORA-04088: error during execution of trigger 'GGWEB.TRIGGER2'
my simple trigger is as follows: create or replace TRIGGER TESTRI AFTER INSERT ON TEST1 FOR EACH ROW BEGIN insert into test1@sqlserver (city,state) values ('what','nine');
All,International company in Australia needs an allround DBA(development&Admin) in MSSQL and Oracle.If you have 5+ experience in development and (not OR) Administration inboth MSSQL and Oracle and like challenges, please send CV toJoin Bytes!.I will send the position description to suitable candidates.Cheers,Yak
Hi,I'm doing an Orcle to SQL port.Simply all I want to do is run a script / DTS package that reads Oraclesequences value and writes the current value of it in to a MSSQL table - ofSequenceName, ValueCan this be done ? If so how ? (Please provide example)Steve
I have a database that is 1.7terabyte in size with 136gb free and throws a "transport level error" telling me to discard the results when I run dbccshrinkfile ('DBNAME', size). I have tried various increments of size, from truncateonly to 1MB below its current value, and nothing works. I have tried to detach and reattach the db, restart the service, restart the server, and none have provided a solution. Any ideas?
I lost few days to solve a problem regarding the connection to an Oracle database server (10) with goal to execute a distributed transaction.
The environment was:
Windows Server 2003 Standard MSSQL Server 2000 Standard Oracle 10g Production Edition
The connection was quite easy to make with the Microsoft driver, but the error was: Server: Msg 7391, Level 16, State 1, Line 3 The operation could not be performed because the OLE DB provider 'MSDAORA' was unable to begin a distributed transaction. OLE DB error trace [OLE/DB Provider 'MSDAORA' ITransactionJoin::JoinTransaction returned 0x8004d01b].
I tried to use the oracle odbc driver, but all kinds of errors were raised, regarding the TNS name and stuff like this. The real cause was that when you create the linked server you must to check the €œAllow inbound process€? in the provider options.
The Oracle ODBC driver allow distributed transactions.
I can't map other sql servers without creating alias with proper port number on client network utility. Other users using same version of the client tools, MMC, SQL DMF etc. I need to map 70 sql servers on using my client tools. Any help is appreciated.
Hello every one I am getting this in my event log from time to time . not sure what is that ? Is this a hacker trying to send rubbish data to SQL through the port ? Any help is appreciated .
Length specified in network packet payload did not match number of bytes read; the connection has been closed. Please contact the vendor of the client library. [CLIENT: someip]
Hi.. I tried to setup MSSQL2000 trans replicatiom using push subscription method. My Publisher is also my distributor. But whenever I started the dstribution agent, I got the following error. It has problem connecting to subscriber's DB with the following error. I had tried that the login is 100% fine. But why this error appear ? Any help?
The name ' ' is not permitted in this context. Only constants, expressions, or variables allowed here. Column names are not permitted. (Source: Subsriber (Data source); Error number: 128)
Length specified in network packet payload did not match number of bytes read; the connection has been closed. Please contact the vendor of the client library. [CLIENT: xxx.xx.xxx.xx]
Client IP address is same as the server its producing the error on. I get these messages around 12pm everyday.
I have VC++ express and MSSQL SMS express and have an application working nicely locally. The Data explorer and data connections part works really easily.
Now, I want to make the application available to my home network.
I mapped the drive where the database is and called it Z: so I could put my "release" on my other network PC and assumed it would find Z: if I mapped the shared network drive on that machine and called it Z:
But: I can't even add the mapped connection on the local machine, I get:
The file "Z:databasescalorie.mdf" is on a network path not supported for database files. An attempt to attach.....etc"
It works fine on the original F drive.......
Am I approaching this the wrong way. How should I distribute to network PCs?
