Re: Repository services and budget

[davidk@ISI.EDU]

Dave,

> Dave Crocker writes :
> 
> At 10:41 PM -0800 1/7/97, Hank Nussbacher wrote:
> >On Tue, 7 Jan 1997, Dave Crocker wrote:
> >But even with 20-30 registrars, you don't want the expensive technical
> >people doing the administrative tasks.  I keep thinking of Dilbert by the
> >xerox machine.  A $25K/yr (even part time) sec'ty is necessary.
> 
> 	There is a very serious downsizing trend for administrative staff
> in organizations.  It's been underway for, perhaps, 10 years.  Given
> voicemail, email, and other computer-based applications, the specific
> functions for an administrative assistant are considerably reduced.  They
> become cost-effective when a group is up toward 10 people, I believe,
> helping with general office and coordination functions, travel
> arrangements, and other logistics issues.  These are not serious
> requirements for a one- or two-person operation.
> 
> 	It can be VERY easy to inflate the requirements for the repository.
> I'm suggesting that it be kept as lean as is feasible.

I support the idea to keep everything as lean as feasible. But I don't
think that it is a good idea to save the money on personnel. I found from
my experience with running the RIPE database that it was of tremendous
value to have a 'secretarial-type/catch all' person in the office
(parttime is fine). Good personnel can save you more money by choosing a
cheap location (if that is needed at all ;-)) and reliable but cheap
computer equipment.

A 'catch all' type of person is very important during the start-up phase
when it is best to work with a small team of engineers to set the thing
up and you don't want to be bothered with all kind of questions that
delay the deployment or distract from the technical work. Also, such a
person is very valuable for informing the user community when there is a
problem as well as help with other unexpected things like getting a spare
part that is needed quickly or the like. We want to start up the shared
registry system as fast as possible and solve operational problems when
the repository is in full operation as quickly as possible. It's probably
worth it to spend some money to assist in that.

And then for the hardware: hardware costs should be kept as low as
possible. Qualified personnel will be able to find cheap and reliable
equipment (that can certainly mean PCs). The equipment needed will depend
on how one chooses to design the software and which data will be
represented. Qualified personnel will be able to make a good design that
will need a minimal of system resources. In any case, the hardware will
be (and should be) much cheaper then the personnel costs. Having a RAID
array or not is not such an interesting discussion. You will need at most
two of them and qualified personnel will be able to find an array that is
much cheaper then their own yearly cost and that is not that much more
expensive then a couple of good quality harddisks.

I feel that it is probably best to do an educated guess for the hardware
budget ($120,000.-) and to spend some more time on the most costly part
of the budget:

personnel costs

In addition to this, I feel that it was a very good suggestion to split
up the budget in setup costs and operational costs after setup. The IAHC
might even want to consider if they are able to get some grants/sponsors
for the setup costs.

David K.
---