Place that Modem on the Deck at Your Feet and Step Back Away from It...
| The case for dismantling modem pools, andespecially not building
new ones, in the 21st century |
Access to the campus network and computing and information resources is
the name of the game, and colleges and universities need to support remote access
to their networks. But what d'es it cost to maintain outdated communications
technology? Here, Brian Voss details why institutions should consider abandoning
modem pools.
Recently, I’ve seen some inquiries from CIOs about the wisdom of installing
dial-up modem lines at their institutions to serve student, faculty, and staff
with off-campus access to their campus networks, and the Internet beyond. I
thought for a moment I had been whisked back in time to the 1980s or early 1990s,
but it appears that now in 2004, dinosaurs may still roam the earth.
Let me say up front for the record: I think establishing a modem pool in the
21st century is a bad, bad idea. I’m not sure it was ever a good idea.
But I can understand how it might once have appeared a reasonable thing to do.
Establishing a modem pool is sort of like buying a boat. It sounds like a good
thing, and you reap some instant joy. But in the end it becomes a hole in the
water you throw money into. Or in this case, a hole in your network you throw
money into. I'll use Indiana University as a case study to try to explain. A
short history lesson—a sort of a parable—follows.
Propagation of the Pool
In the 1980s, IU’s academic computing organization established a small
(<20 lines) modem pool to serve a small and select community of faculty,
researchers, and grad students who were big into computing and who wanted to
be able to access those time-share machines from home to run jobs, check outputs,
or monitor operations. It was a modest pool that cost very little to establish
and support, and it paid great dividends in terms of making a small group of
influential and powerful people happy. I was not involved in this decision,
but I would have done the same thing. At the time I was on the administrative
computing side, and we established just a few lines to allow our staff to get
into machines for emergency purposes from home. Time, and technology, marched
on. Computing centers and organizations merged. And the population of users
who desired to get into on-campus resources grew. And new applications started
to arise, such as electronic mail, bulletin boards, and other user-focused services
(as opposed to techie machine programming and management).
And so the modem pool grew, continuing to “feed the bears” As it
became larger, so did the funding required to support it. And funding, always
in short supply, meant that modem supply always was outpaced by dial-up demand,
and users got busy signals. So much for making a group of influential users
happy!
As the 1990s moved on, along came more and more applications that people could
access from off-campus. Then came the Web/Internet, and the need expanded beyond
campus resources to the world at large. By the mid-to-late 1990s, everyone needed
to be connected regardless of time and place. So the over-burdened, always-busy
modem pools at IU became a constant source of irritation and outrage to the
user community.
Every year, in our annual user satisfaction survey, modem availability was
the Number One complaint. After all, this remote access was in support of university
missions for teaching, learning, research, and service. It became more than
an entitlement, it became a “right” of faculty especially, but also
of students trying to get online to do assignments for that same faculty. And
those rights being denied (through constantly consumed modem pools) by the IT
organization led to even greater dissatisfaction. Our users were ‘madder
than hell and not going to take it anymore!’
In 1998, IU’s groundbreaking IT Strategic Plan (http://www.indiana. edu/~ovpit/strategic/),
written by that same user community as a roadmap for advancement in technology
infrastructure and service by the central IT organization, included a major
recommendation on access to IT resources and services. And a key direction,
action item #5, involved remote access (i.e., modems). It said, in very easy-to-understand
language in the interpretation comments: “No Busy Signals!”
With that mandate, strategic funding was directed into the modem pools at the
two core campuses (Bloomington and IUPUI in Indianapolis), and the pools grew
enormous: ~1,800 at IUB and ~1,100 at IUPUI. And so did the investment in this
service—more than $700K per year in PRI telephone line charges, and over
$500K in hardware.
With this investment, we discovered the legendary “peak in the demand
curve”—and we met that demand. No busy signals. Ever. Room to spare.
We were her'es! Other items rose to the top of the user complaints in the annual
satisfaction survey. But busy signals were banished. All was right with the
world. Right, except for that $700K annual cost.
One good thing about the IU IT Strategic Plan’s action item #5 was that
it didn’t just say, “Add modems.” In the actual action item
charge, it wisely talked about the more generic issue of reliable access. At
the time it was published, the very first inklings of the next stage of technology
were emerging. This, of course, meant that our charge was to develop new, high-speed
services. These high-speed services (Broadband: DSL, cable, etc.) would not
be something the university itself could get into the business of providing.
Rather, they would be vendor-provided services in our communities, coming into
users’ homes along with the telephone lines and TV programming. It became
apparent that the natural evolution of remote access, as dictated by the IT
Strategic Plan, would provide a revolution in the historical role of Institution
as ISP. Bravo!
After the Apex
At IU we have seen significant migration off our low-speed modem pools and onto
vendor-provided, user-contracted broadband services. Of course, no longer is
Internet access really considered by most faculty, staff, and students just
an item required by their association with the university. It is now being viewed
increasingly as a utility for one’s home—just like water, gas, sewer,
phone, or cable TV.
It is no longer used by the household university-affiliate; it is a commodity
used by everyone—spouse, children, visitors, and others. As of Spring
2004, we’ve seen about half of our users establish broadband connectivity
in their homes, and they no longer use our modem pools. Because really, surfing
today’s multimedia-rich Internet on a 56K dial-up is as productive and
exciting as watching grass grow. And through this natural migration, we've managed
to reduce the modem pools (and still retain the ‘no busy’ service
level)to their pre-IT Strategic Plan levels (about 1,000 lines at IUB and 700
at IUPUI).
So how about our short-term view for the next 12-18 months? Our plans are still
forming, but in reality, the key people who rightfully drove the process of
expanding the modem pools in past decades—the IT-savvy faculty, and our
student community—have moved on. If you ask our primary IT advisory faculty,
nearly all of them have broadband access at home, right along with their HBO.
In fact, many have been able to justify and fund the cost of the broadband service
by dropping the second home phone line they had (so that they could be online
on the IU modem pool and not, themselves, create busy signals for someone trying
to call them at home).
As for the students, every time I guest-lecture on campus, I always ask if
anyone is using modem pools. Rarely, maybe one or two hands will sheepishly
go up. Most times, the students look at me like I’m from Mars (“What’s
a modem pool, Mr. Voss?”). And those students who don’t live off-campus,
well, they’re in our wired-to-each-pillow residence halls, which also
get wireless access beamed from new-technology phase-array antennas through
their windows. In truth, we’re still getting demographics regarding just
who is using the modem pool today. But it appears from anecdotal evidence that
it’s not leading IT faculty, and it’s not the majority of the student
population.
So perhaps the time is about here when the rallying cry ‘No Busy Signals!’
is less a cry and more a whimper. I don’t see us ever fully eliminating
the free access modem pools. But I am confident we will scale them back to much
smaller levels, and deal with the fact that they may, once again, have busy
signals.
Morals for Modems
What’s the moral? Let me give you some itemized arguments to use when
someone suggests that you start a modem pool in the 21st century. Or you can
use them as ammunition if you’re looking to dry one up that was established
by you, or your long-gone predecessor, in the 1980s or 1990s.
Dead technology. Low-speed dial-up is becoming useless. You can’t effectively
surf the Web of today with it. You can’t use it to do one of the most
popular activities today—downloading. Lower speeds were fine when you
were in character mode; but what is left that’s like that today? And if
there still is any of that, it won’t be around long.
Oh, what about those five-times-faster claims of some ISPs for their dial-up
modems? Can’t universities install that? Smoke and mirrors, friends. They’re
simply using compression algorithms, and that d'es little for improving overall
performance. It’s slick marketing meant to squeeze every last drop out
of an already-made investment by ISPs who can’t get into the broadband
game.
Scarce funding resources. Higher education is, across the board, being
pressed financially. Modem pools are not cheap. Given the dead technology issue,
do you really want to invest your scarce resources in this area? Wouldn’t
the funding be better spent on more Internet capacity and network security?
New wireless access infrastructures? Building fiber on-ramps from the institution
to the Internet? And these are just in the telecom area of IT. Add in all the
demands for new funding for information systems, user support and services,
and research-focused computation and storage. Where d'es low-speed dial-up fall
in that list?
Institutional role. Given all the broad funding pressure, is it really
in the institution’s best interest to become an ISP for its faculty, staff,
and students? Outsourcing has become, and continues to be, a big topic in higher
education; since when is “being an outstanding ISP” part of any
higher education institution’s mission-centric strategic plan?
Equity. Is it really fair that the institution provides open, free
access for what has become a household utility, while most members in the university
community pay their own freight? Oddly—and this is just a suspicion—my
guess is that the biggest remaining user base of low-speed modems are faculty
and staff. So this is less about access for students to learning resources,
and more about basic Internet access for these employee households. I don’t
mean to widely cast aspersions; I’m sure there are some underpaid individuals
who use the resource to do their jobs and nothing else. But in general terms,
the day of this being a faculty-entitlement, or an employee benefit, seem to
have passed.
Funding reallocation. For those like IU that have had huge modem pools
(and the equally huge budgetary investments) I’m not advocating that the
funds saved from their devolution will go to party-hard. We have so many unfunded
technology mandates. Wireless networking is a major new one, because it d'es
not replace well-wired networks on our campuses; it expands and complements
them.
Network security—from firewalls to Virtual Private Network (VPN) servers
(that secure those wireless and broadband remote services)—is another
new area of investment. With Internet expansion, costs go down, but demand continues
unabated. More investment is needed, including the building of optical on-ramps
that will better connect the institution to its neighbors and collaborators,
and the developing national cyberinfrastructure. All these areas are essentially
new money initiatives. These services are, in the view of most IT people and
most users of IT services, critical to the continued ability of IT to be a strategic
enabler of the larger university mission in teaching, learning, research, and
service. Wouldn’t it be grand if we could look at our modem pools as a
funding source to provide these new services?
In the end, I do believe we must retain a focus on remote access. We should
enable it through working to leverage the buying power of our institution(s)
in trying to get better rates for broadband services from the competing providers,
into homes in our communities.
We should work with providers to establish peer
relationships, which will, in the end, help improve performance for both our
users on their networks, and their other users (as load is shunted off directly
to our campus networks). We need to continue to work on securing our networks
while making it possible for “anytime, anywhere” access to institutional
data and resources through technologies like VPN. Let’s have our IT organizations
do these things, instead of running modem pools.
In the end, let's consider modems another bit of historical technology, past
its useful era, that needs to go onto the technology discard heap. Along with
dumb terminals, floppy disks, mono-chrome monitors, non-GUI applications, SNA
networks, and a whole generation of x86 technology computers. You wouldn’t
give more than a nanosecond of thought to implement any of those technologies
for on your campus, right? Same for modems.