Thank you for all adivce.
I know the 3490 drive is the compression used by IDRC.
However, the 3592 appeared to have a different compressed algorithm.
So I asked if any 3592 drives is compatible.
See the bellow:
https://www.ibm.com/docs/en/ts3500-tape-library?topic=drives-3592-tape
Data compression
The 3592 tape drives use the data-compression method known as streaming
lossless data compression algorithm. The compression logic for TS1120 and
later tape drives operates at more than twice the overall transfer rates of
the 3592 J1A tape drive.
Best regards,
Toyokazu Kobayashi
----- Original Message -----
From: "Radoslaw Skorupka" <r.skoru...@hotmail.com>
Newsgroups: bit.listserv.ibm-main
To: <IBM-MAIN@LISTSERV.UA.EDU>
Sent: Friday, April 23, 2021 4:15 AM
Subject: Re: 3592 hardware compression
W dniu 22.04.2021 o 16:38, Ed Jaffe pisze:
On 4/22/2021 1:31 AM, (K.K.Paradox)T.Kobayashi wrote:
There is a tape where a DFDSS dump was created on a 3592 drive (TS1130)
with the COMPACT feature enabled.
Q1. Can I read this tape on another 3592 drive (such as TS1130 or
TS1140)?
Are the 3592 hardware compressions compatible?
IDRC is IDRC. At only about 3:1, It ain't great. But, it's compatible...
Gentlemen,
Drive compression is *unpredictable* unless you know data to be written.
The true is same data on same drive will be compressed with same ratio.
BTDT. I spend a lot of time playing with compression ratios and preparing
data compressible 3:1, 6:1, 1:1, etc.
Long time ago I've got F37 abend, because my tape dataset exceeded 6TB -
but it was on JA tape in J1A drive. Native capacity is 300GB, so I got
over 20:1. I don't know exact value, because RMM was unable to record such
big numbers - that was a reason of F37.
3:1 is marketing value, which is more or less based on typical data.
However "typical" is not well defined, and you may not know how much
typical is your dataset.
Remark:
We know marketing 3:1 value. There is another one: speed. Is your drive
capable to write 360MB/s? Great, but you actual throughput will be lower.
Nominal value is true for big blocksize *and* proper compression ratio,
usually nominal one.
Worse compression (i.e. 1:1) mean worse throughput, usually you will get
same compressed stream, so in case of 1:1 it would be 3 times worse.
Better compression ...can be even worse, much worse! Very good compression
means very little compressed stream. Tape drive can slow down, there are
several gears (speeds), but when the stream is not enough to satisfy the
lowest gear the data buffer is exhausted, drive stops, rewinds a little
and starts again. Veeeery slow.
Of course the above covers real tape drives only, so it is rather
historical.
--
Radoslaw Skorupka
(looking for new job)
Lodz, Poland
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
