Archives of the TeradataForum
Message Posted: Fri, 17 Jul 2004 @ 00:47:54 GMT
<-- Anonymously Posted: Friday, July 16, 2004 20:30 -->
I agree with Daves comments. The under lying protocol will keep changing as new versions come out. The encryption is one part that will have to change as computer power and algorithms improve. For one the encryption used in V2R5.1.0 was very poor and I found that without a lot of effort you could easily crack the passwords.
When I looked into writing a native interface into Teradata, I found that the protocol was pretty straight forward, even though it is not documented. I also found that you could easily get greater performance from writing your interface layer, than using the CLI provided. I would highly recommend that more effort is placed into making the CLI more performant, so that the likes of DBD::Teradata would choose to make use of this, rather than write their own. Another option could be t open up the CLI source code, say under the LGPL, to allow for people to improve the code on Teradata's behalf.
If I was Teradata, I would make changes to the underlying protocol on each version, to ensure only tested interfaces are used against Teradata, as non-tested interfaces could cause a Teradata system to react badly, which if I the Teradata DBA I would be pretty miffed at the prospect of being woken up in the middle of the night because a developer selected to used an unsupported product.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|