|
Problem |
If you are running JAVA applications in
Websphere ZOS that issue JDBC calls to DB2.
1.Where does the code translation occur to convert the DB2 EBCDIC data to
UNICODE for the JAVA program if DB2 is on the same LPAR with WebSphere?
2. Do we have to specify anything for the conversion, or is it part of
the JDBC driver to perform the conversion?
3. Where would the conversion take place if the DB2 table is on a remote
DB2?
4. What if the table data was stored in unicode-8 or Unicode -16?
5. Are there any performance metrics comparing EBCDIC storage vs UNICODE
storage? |
|
|
|
Solution |
When you bind the application, you can specify its
encoding scheme. If your application is in UNICODE and DB2 stores data in
UNICODE, then there is no need for conversion. Otherwise the conversion
will occur.
.
DB2, not the driver will do the conversion when you store or retrieve data
from DB2. DB2 will use OS conversion service. Please refer to info apar
II13277 for more information on how to enable the UNICODE CONVERSION
service.
.
When accessing remote database, according to DRDA architecture, the
receiving side will do the conversion. In this case, the coversion will
occur in DDF.
.
The conversion occurs regardless of UTF-8 or UTF-16, if you have UNICODE
application access EBCDIC data or EBCDIC application access UNICODE data.
.
The encoding conversion will have CPU overhead and it is CPU intensive
workload.
Note: Refer to the following RedBook: DB2 V7 Performance which discusses
the conversion will have large CPU cost, but no benchmark..
.
taken from PMR 40771,227,000 |
|
|
|
|
|
|
|