[AG-TECH] Fading AG usage?

[Jennifer Teig von Hoffman]

Man, this is an interesting conversation, keep it going!

George's comments made me want to chime in here too . . .

George Estes wrote:

> Just my personal opinion, but I think
> many of the sites falling off the AG wagon are sites that can't, or
> won't for whatever reason, put forth the effort necessary to make the
> AG work well for them.  

The AG is perceived as relatively high-maintenance. Whether or not this 
perception is accurate (a whole OTHER conversation), the perception 
itself does influence people's decisions.

Anecdotal evidence about this . . . at BU we've always had a lot of 
requests for AG usage from outside our university. People at other local 
universities and companies come to us, rather than setting up their own 
nodes. Last year we started charging fees for usage, in part because the 
volume of requests was high enough to be significantly impacting our 
workload (and it's hard to justify spending that much time supporting 
other people's meetings, without any reimbursement to cover the cost of 
that time).

Sometimes it makes a lot of sense for people to pay for using our node, 
rather than building one themselves. If you need a node that seats 10, 
you need it next month, and you don't have appropriate networking or 
hardware, it may be a lot easier and smarter to send us a check and use 
our node, than it would be to try to get all that work done in time 
(just *think* of the administrative nightmares some of us go through to 
get appropriate networking!).

But sometimes the scenario doesn't look anything like that. Sometimes 
it's one person who needs to join a meeting, and they're from an 
institution with plenty of bandwidth. I often think they'd be better off 
spending an hour and $150 to set up a Personal Node for whoemever needs 
to connect to a meeting -- and my suggestion that they consider this 
option usually is not well-received.

> And then there are probably some sites that
> got involved during the initial excitement but now don't have enough
> remote collaborations to support the Access Grid.  

Initial excitement and, in many cases, initial funding. A small number 
of sites were funded as a testbed, to set up AG nodes in 1999 and 2000 
through the National Computational Science Alliance. Since the Alliance 
has been gone for a while now (meaning that both that money and that 
particular bully-pulpit are gone), it would be interesting to know how 
many of those nodes are still being used at least, say, five times a 
month. We were one of the 1999 sites, and I know our original AG node is 
still pretty darn busy!

Cheers,
Jennifer