Linux as
a network solution - During my Linux classes at CMCC, I'm always
amazed at the amount of effort that I have to make in order to pursuade
students, new to Linux, that it is another powerful operating system.
Most of the people new to Linux aren't new to computing - they usually
have a few years of Windows behind them. I've found out that's not
always a good thing. Now when I talk to new people, either students,
or people in general, I'm less likely to want to convince them or
pursuade them that they should make the leap over to Linux. That's
too much work, I'd rather wait until they have done some research
themselves and have decided that Linux may be the only way that
the job they need done can only be done with Linux. Minot Consolidate
School was such a job. All of the pieces that are necessary for
Linux to be the operating system of choice were at the Minot Consolidated
School.
Back to Top
Minot
Consolidated School - Minot, Maine
Project
-
The Minot Consolidated School has been surviving for the past few
years, like many other small school systems in Maine, with limited
or inadequate internet and email access, no internal school servers,
and plenty of extra unconnected Cat. 5 ethernet cable running "where-evaa"
above the ceiling tiles throughout the school.
Goals:
Improve existing network to include security, reliability, durability,
and consistent functionality, all with as little or no cost as possible
and with as little need for change for the existing user as possible
ie.. If a teacher feels comfortable with the Windows 98 operating
system on their desktop then give that person the option to keep
that and work with that.
Overall goal:
To create (as close as possible) a healthy, secure, minimal maintance,
easy to troubleshoot network with all
in-house existing hardware at little or no cost.
But first -
"Hunt down wires and make sure ends have labels."
Opportunity
#1
How do you take
an existing network (really a large bunch of single wires) that
should support about 50 desktop pc's and 80 laptops plus about 5
printers and make it operate effectively.
Baseline
- Initially only about 6 pc desktop users could connect to the internet
with reasonable consistency and the laptops - it depends.
Decision
- First, I felt we needed to find out what wires were being used
and where they were going. So, along with Mark (Director of Maintenance)
, Kevin his assistant, Rick Roy, and myself, we started the process
of finding end to end locations of wires. (This started over Christmas
break - Dave Bressler was busy working with pc desktop issues.)
When my new
semester (Jan. 14th) for Advanced Linux started I was able to recruit
more people to help. (Tom Pelletier, Luke Hague, and Bill Hickey
- Devin Giroux came later to help with the desktop computer issues.)
Conclusion
- Out of maybe 80 or more wires that had been run above the ceiling
tiles, we narrowed the need down to less than 20. That means that
most of the wires that had already been run were not needed. The
old headset was, if something was wrong or a new hookup needed to
take place then run a new wire not use an existing one.
Opportunity
#2
The computer teacher
had been going to auctions and buying extra equipment ie.. PBX system,
thin client systems etc.. which he intended to use for some improvements
to the school resources. Unfortunately, the PBX equipment, after
a bid from an outside contractor, would cost an additional $4,000
in order to use. The thin client systems were not set up to work
with Linux, which was the intended use, so until a solution could
be found, they were not useful.
Also, the school had
about 20 older Dell computers donated to them, but most were stacked
in the computer room not being used.
Baseline - Money
had already been spent and none of the added equipment was in use.
Decision - Initially,
I looked at the PBX equipment, I researched the cost and the implementation
strategy on-line. This equipment would be over-kill for this small
school. My recommendation was to find another less-costly solution
that would better suit their needs and see if some of the money
spent (about $1,500) could be recovered - maybe on ebay (Mark the
custodian's suggestion). Final outcome still pending.
Next, I started working
on the Terminal Server Project, which had been the intended use
for the Thin Clients (approx. $800). I immediately decided not to
waste a lot of time trying to make the thin client boxes work. I
went to the stack of Dell's and started selecting the best units
for the Terminal Server Project, along with a Packard Bell system
(temporary server) to use as the server and had the project completed
with (4) terminals accessible in the library. (Cost $0)
Finally, I talked to
Dave Bressler about having a goal of one computer per classroom
with at least 128 meg of RAM in each system and the ability to connect
to the internet. Anything more than that would have to come later,
if we had the time - and no promises beyond that.
Conclusion - PBX
system for school - still pending a decision. Linux
Terminal Server Project - completed, with the option
of expansion throughout the school if necessary. Currently, all
classrooms that want/need the desktop pc's are up and connected
to the internet - completed.
Opportunity #3
There are no servers
in the school.
Baseline - Nothing
needed to be changed or upgraded, nothing needed to be copied, nothing
existed. The school had been relying on MSLN (Maine School and Library
Network) to supply them with not only DHCP but everything else as
well. That meant that if any PC needed to connect to the internet
then it had to go to MSLN (Orono, Maine) to get all of the necessary
information, ie.. network addresses etc.. This not only risked the
possibility of not having a good connection but also of having too
much contention and delay on the network. Example: If I needed to
get a local address (in Lewiston) but had to go to Augusta every
time I needed it, then I risk the possiblity of my car breaking
down along the way or any number of things happening before/or if
I get back to Lewiston. What I really need is the ability to get
a local address, locally.
Decision - Build
the necessary servers (locally) and have them act as the decision
makers (locally).
Conclusion - I
selected one of the Dell's (600 Mhz) and loaded it with RedHat Linux
and made some changes to assist in network support. Loaded it to
act as the network Firewall/ local DHCP/proxy server (for the internal
NAT network) and have the option of making it work also as a local
mail server - if necessary. Because most of the systems at the user
level are loaded with some flavor of Windows, I have the ability
to use SAMBA for file sharing as well. Also, because of the Apple
laptops, I've included Appletalk - which may or may not be used.
Mission accomplished - all desktops and all laptops in the
school can access the internet and they all connect with greatly
improved speed.
Linux instructor - Don
Cunningham
Back to Top

This
is the current Linux Network. It is a basic network that fulfills
a really basic network function - it connects to the internet and
allows network browsing and email retrieval. It still needs a lot
of work to be considered a full functioning network such as, a backup
system, a print server, a logging server, network monitoring software,
etc..

Back to Top
Phase
1
1. Assess needs
and suggest necessary work to improve existing network in order
to consistently connect to the internet and read email on at least
one computer per classroom that wanted/needed that connection.
2. Schedule
dates, times, and manpower to facilitate work in phases of completion.
3. Divide responsibilities
up by need, desire, and ability.
4. Complete
necessary wiring/re-wiring of the network and re-work existing PC's
to insure at least one computer per classroom with at least 128
Meg. of ram and can successfully connect to the internet.
5. Install wiring
rack and punch down necessary wires and illiminate unnecessary or
extra cables.
6. Complete
network server(s) (DHCP,NAT,Firewall, etc...), print server, and
possible DNS, Web, and File server. Also install and test UPS's
etc.., write preliminary network policies ie.. security, backup,
maintainance etc...
7. Complete
work on the Library LTSP (Linux Terminal Server Project)
8. Test complete
network by the end of the first school break in Feb.
Back to Top
Phase
2
1. Install print
servers and determine adequate locations throughout the school.
2. Make sure
that the teachers in the classrooms are satisfied (as much as possible)
with their current computers, operating systems, and email solutions.
3. Start the
process of formally documenting all necessary paperwork ie.. Policies,
Procedures, Hardware Sheets, etc...
4. Load and
use available network monitoring software.
5. Test policies
and procedures ie.. backup policy, security policy, etc..
6. Address other
functional requirements within the school ie.. need for scanners,
library software, more LTSP labs, possible in-house webmail server
for local school mail, local content management site for information
sharing between teachers and/or administrators.
Back to Top
What's
next ?
I don't know! What should
be next is to finish the phase 2 process, but a number of things
are up in the air: support, our continued need, and just a general
understanding and appreciation of the work that has already been
done. I would like to think that what has already happened will
create an awareness that this approach to training is not only a
benefit to the students but it is a solution to local schools or
small businesses that really need this type of help - and I'm not
suggesting that this should be a free solution (meaning instructors
free time and work done totally for free by the students) but this
should be a solution that is worth investigation by the State or
whatever organization is tasked to help.
There are things that
I personally would like to have the opportunity to assist in, such
as a better and cheaper way of administering support for installed
networks, better use of available equipment, a cost effective strategy
to allow students to "earn while they learn", and some
master plan that uses the most cost effective tools in building
secure, stable, and reliable networks that need very little support.
I think I've been able to demonstrate that Linux should be the operating
system of choice for this type of need. By the way, I have no stock
in Linux - I just believe in using the right tool for the right
job. As for the future - We'll see.
Don Cunningham - Linux
instructor
Back to Top
|