[442] in Enterprise Print Delivery Team
Re: Pete's questions
daemon@ATHENA.MIT.EDU (Rocklyn E. Clarke)
Fri Aug 11 01:50:33 2000
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Content-Transfer-Encoding: 7BIT
Message-Id: <000811.013714.EDT.RCLARKE@MITVMA.MIT.EDU>
Date: Fri, 11 Aug 2000 01:34:15 EDT
Reply-To: "Rocklyn E. Clarke" <RCLARKE@MITVMA.MIT.EDU>
From: "Rocklyn E. Clarke" <RCLARKE@MITVMA.MIT.EDU>
To: PRINTDEL@MITVMA.MIT.EDU
In-Reply-To: <4.3.2.7.2.20000810150524.00dca9e0@hesiod>
Hi Everyone,
I would frame the issue in this way:
1. We obviously need to purchase whatever is required to handle the general
ledger statements, PLUS whatever is printed on the 8100s. I the scaleable
server is required to do this then the discussion is over. If the 170 is
adequate for this task then we need to consider the following:
2. The terms of our IPM software purchase specify that we have 6 months to
determine if we want to upgrade our configuration to handle the additional
printers. Our current plans call for us to "spec our" a project for
making this decision before we shut down our current project. I therefore
think that if we add the non-central printers at all, we are likely to
add them well before the end of FY 2001 (June 30, 2001). We must therefore
prepare for this likely development.
3. Nevertheless, it is perfectly reasonable to make the costs of any necessary
hardware upgrades a component of the "additional printers project". I
think that it would therefore be reasonable to pose the following question
to Theresa:
If you were satisfied that IPM was the right solution for handling the
distributed administrative printers, would you be willing to pay $X to
move them from their current home with Athena Print Services? Here the
$X figure represents either:
a) the cost of upgrading the RS/6000 server.
or
b) the cost of the RS/6000 upgrade PLUS the cost of the IPM license
upgrade.
Personally, in light of our new IS funding policy, I lean towards "b)".
I would therefore recommend that we determine the cost of upgrading from the
model 170 configuration to a configuration that can handle the non-central
printers, and share this information with Theresa. We should also verify that
the model 170 can handle our Phase 3 needs. If it can, we should go with the
model 170.
Rocklyn
-------------------------------------------------------------------------------
On Thu, 10 Aug 2000 15:46:50 -0400 Lynne E. Durland said:
>Greetings,
>
>I caught up with Peter this afternoon.
>
>He has one of the three quotes back. Part of what was behind his questions
>was the total price of the quote he has received. Basically it boils down
>to do we REALLY need the scaleable server, or would the model 170, which is
>a slower, single processor box, really do what we require, for say the next
>twelve months. Will we realistically be adding, or at least start adding
>the non central printers before the end of the year or will that actually
>not start until 12 months out?
>
>The 170 would still have the RAID. The rfq he sent out has six spindles
>but only 18 gig of usable for "data" or non operating system stuff.
>
>Part of the concern Peter was expressing is that the response he got was "a
>lot" of money, and if he does the expensive stuff for the print server,
>what does he have to say no to on other servers to make his budget work for
>the year.
>
>I expressed that I would send this email to the group and let the team hash
>out the question rather than just me trying to make the decision whether to
>stick with the scaleable box or go with the 170, single processor box.
>
>I did run some numbers around the general ledger print file sizes. What I
>came up with by, adding up the number of records in the 21 print files,
>then multiplying by 132, (assumption of 132 byte records), came out to
>about 834 meg of data. When I cross check this number with the size of the
>spool space, 2000 cylinders times JAD's magic number the spool space is
>about 1.32 gig, which allows the general ledger to hit the spool space and
>still have some room for "normal" print files. So if we round up to a gig
>for the general ledger, and for a rough guesstimate quadruple that for
>postscript, we only need 4 gig for the general ledger to hit the queue and
>then be retained. So the 18 gig of disk is sounding reasonable at this point.
>
>Lynne
>
>Lynne E. Durland
>Information Systems
>Database Services
>W91-109
>O:258-5857
>C: 617-293-8091
>B: 617-430-8762
>H: KB1FEM
>
> "Happiness often sneaks through a door you didn't know you left open."
>
> --John Barrymore
