 |
 |
Archives of the TeradataForum
Message Posted: Thu, 13 Mar 2003 @ 18:40:01 GMT
| Subj: | | Re: FastExport Benchmarking |
| |
| From: | | Michael McIntire |
| | Does anyone know how this can be calculated - do you need to take into account the row length, the number of nodes, the number of
sessions logged on, number of rows, perm space of table transferring, Also, can anyone share their own benchmarking/performance testing or
real life figures. | |
A number of years ago, a client did some throughput testing with FEXP, bottom line was that in 5150 to Sun testing (100MBit FDX),
the max throughput was about 900K/card on the 5150 (20+ cards) side and about 7M/sec (millions)/card on the Sun side. Another card was added
to the Sun to improve performance...
Point is that FEXP throughput figures can easily be limited by hardware/networking configurations of both the host and the client. For
some simple testing, I think you can limit the potential number of network paths by limiting the TDPCOPXX entries in the client side HOSTS
file.
Talk to your NCR technical rep, who may be able to give you some rule of thumb performance numbers for your configuration.
| |
