From pyeh@phys.sinica.edu.tw Mon Mar  8 18:45:28 1999
Date: Thu, 10 Dec 1998 13:23:33 +0800 (CST)
From: Ping Yeh / Academia Sinica <pyeh@phys.sinica.edu.tw>
To: Taiwan Group at CDF -- Yen-Chu Chen <chenyc@fnal.gov>,
     Gong-Ping Yeh <gpyeh@fnald>, Ming-Jer Wang <ming@fnald>,
     Paoti Chang <paoti@fnald>, Paul S. Chang <pschang@fnald>,
     Antonio Wang Chan <tony@fnal.gov>, Yi-Cheng Liu <liu@fnald>
Subject: Run II Tape Technology (fwd)


Hi friends,

    Related to the Run 2 tape technology evaluation distributed
earlier, Sliwa sent out this request to representatives.  Well,
I have no idea what HPSS is.  As many of you might already know,
currently it seems Enstore will be the choice of mass storage
system for production farm.  But for home institutions the
requirements are different.

    Since this will affect how we're going to perform data analysis
back at home.  Please take some time to think/explore about it
if you can.

							Thanks,
							Ping


---------- Forwarded message ----------
Date: Wed, 09 Dec 1998 09:07:16 -0500 (EST)
From: SLIWA@tuhep9.phy.tufts.edu
To: cdf_comp_reps@fnal.gov
Cc: SLIWA@tuhep9.phy.tufts.edu
Subject: Run II Tape Technology

I would like to request a following requirement to be taken into account when
the tape technology choice is made:

The chosen tape technology MUST be supported by HPSS. There are many good
drives and robots which are supported by HPSS at present, and which are used
at several HEP experiments in the world (CERN, STAR, Babar...). Addition of 
this requirement will provide CDF with a backup solution in case ENSTORE does
not satisfy the CDF requirements.

			Krzysztof

From GPYEH@fnald Mon Mar  8 18:45:38 1999
Date: Wed, 09 Dec 1998 23:42:54 -0600
From: GPYEH@fnald
To: chenyc, tony@fibi02
Subject: Re: Run II Tape Technology (fwd)


   HPSS    is the "High Performance Storage Systems" software
                  lead by  IBM  and  a dozen labs.
           Fermi Lab  was the leader with IBM.
           But,  it was performing very poorly for Fermi Lab.
           So, Fermi Lab decided to drop it.

           Meanwhile, a few other labs decided to use it
                 because they thought Fermi Lab was going to use it.

           An institution has to pay a lot of money to use  HPSS.
           So this applies to very few  CDF institutions.


   ENSTORE  is the new software being developed by Fermi Lab and DESY.
            DESY  has been using most of it, very successfully in the
            last couple of years.

            D0 has agreed to use this, 
               but CDF has not.

   I don't understand why either one of these have anything to do 
            with  remote institutions.

   I think the questions are:    high density, fast, affordable

   We have not decided what CDF will use in FCC,
               maybe combination of  Sony 
                             and     STK  Redwood

               STK  drives are very expensive

               One type of Sony drive is cheap

From pyeh@phys.sinica.edu.tw Mon Mar  8 18:45:54 1999
Date: Thu, 10 Dec 1998 14:39:47 +0800 (CST)
From: Ping Yeh / Academia Sinica <pyeh@phys.sinica.edu.tw>
To: GPYEH@FNALD
Cc: Yen-Chu Chen <chenyc@fnal.gov>, Ming-Jer Wang <ming@fnald>,
     Paoti Chang <paoti@fnald>, Paul S. Chang <pschang@fnald>,
     Antonio Wang Chan <tony@fnal.gov>, Yi-Cheng Liu <liu@fnald>
Subject: RE: Run II Tape Technology (fwd)


The question from the viewpoint of remote institutions is:

	What kind of tape drive should I buy?
	(issues on cost, speed, compatibility, ... etc)

That, however, may couple with the mass storage system in a
non-trivial way:  the production farm will put data in the
mass storage, and somebody has to copy them to tapes for remote
institutions.  The correspondence among files in mass storage,
files in tapes for remote institutions, and production database
doesn't seem to be simple.

Anyway, I guess the Data Handling group will give us a good
answer in the end.  But maybe we should really think more into it.

						cheers,
						Ping