Re: Exabyte Mammoth vs. 3570

1997-12-05 21:32:20
Subject: Re: Exabyte Mammoth vs. 3570
Date: Fri, 5 Dec 1997 19:32:20 -0700
be careful with the benchmarks.  it's important to understand real workload
vs. streaming rate performance.  if you can stream continuously, the mammoth
will likely win.  if there is a lot of start/stop, the mechanics and motors
of the mammoth are not robust enough allow for good performance.  this is
what the 3570 has been engineered for and it will likely outperform both
mammoth and DLT.

>To: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
>From: ian AT mainland.ab DOT ca (Ian Munro)
>Subject: Re: Exabyte Mammoth vs. 3570
>I know there is a benchmark which IBM has comparing Mammoth to DLT7000 to
3570 ... I'm currently trying to get my hands on it and will repost once I
find out it's location.  Based on the information I've heard, the 3570 was
the highest performing drive in this test, Mammoth was second ... likely
because of a high degree of start/stop activity ...
>>I'm in the process of evaluating an Exabyte 440 tape library
>>with Mammoth drives and a 3575 L06. 
>>My server platform will be WinNt and most 
>>of the clients will be connected via 16M/Bit Token Ring. (I know
>>we should be running ATM/FDDI, but over here it is VERY 
>>Has anyone used one of these libraries and can share 
>>some good/bad experiences with me?
>>I've heard some very good things about the mammoth 
>>in terms of future performance that will make the 3570
>>look like it's standing still.
>>Does not matter if your server is on another platform, the
>>main thing is the reliability of the drives/ATL and performance.
>>We will be rolling out ADSM big time and I don't want to 
>>choose the wrong library to do this with... 
>>Thanks in advance! 
>>Christo Heuër
>>ABSA Group 
>>South Africa
>>E-Mail: christoh AT DOT za
Ian Munro
Vice President
Mainland Information Systems
(403) 266-6611 (Office)
(403) 818-4141 (Cell)
<Prev in Thread] Current Thread [Next in Thread>
  • Re: Exabyte Mammoth vs. 3570, Ian Munro <=