So I finally got my journal going. What a feat... Anyhow, as you may have heard/found out, I'm working on a network management utility for managing moderately sized networks. My primary testing ground will be the Earlham College Computing Services. They manage the networks here at Earlham College. My goal is to create a utility which assists network and system managers to more easily manage the addition of new computers to a pre-established netowork. This of course was not my origonal choice for my project. My first choice was to combine this tool with a User Account management tool. But due to the potential complexitites of that project, I decided to narrow my focus down to just the Network Management tool. My problem space was thereby narrowed to a much easier project to acomplish in the time alotted. So-far, I've been able to establish a basic framework for the features and requirement which the users feel are needed. I have also been able to start my research into Software Enginering and Requirements Analysis which is going to be the focus of my paper. I will also be focusing on the Requirements Requirements Analysis stage of the development of my project. I'm really looking forward to doing more on this project and hope that I am able to finish it during the semester. Currently that is my largest concern. However, with good time management and sticking to established schedules, I should be good to go.
Now for the details (as detailed as they are). Earlham Computing Services (a.k.a. ECS) manages the network infrastructure on the Earlham College campus. They also provide support for the public labs and faculty machines. Throughout the year there is a continuous cycle of computers moving around and throughout the campus whether or not they be student owned or college/faculty owned. As they move around campus, their network requirements may change resulting in an almost constant reconfiguration of the college's network infrastructure. Currently, a majority of these changes are done by one or two people who manage the network and the underlying systems. My goal is to shift some of this "responsibility" onto others, thereby potentially making the job of the Network/System manager easier. Initial poling indicates that this would be a much welcomed step. This would require the manipulation of DHCP and DNS configuratoin files. There are many potential ways to acomplish this. Part of my idea is to create a tool which operates on both the command line and potentially a web browser which allows an authorized user to "register" a new computer on the network. This would be most useful for registering small numbers of machines individually, but not large fleets of machines since it still requires human interaction per machine. My solution for the larger problem of registering large numbers of computers at one time almost autonomously, would be to use some sort of automatic web based registration tool. A pre-existing solution to this may be the use of the Net-Reg program/interface. This allows an un-registerd computer to access a private network in which all addresses resolve to one address which in turn takes them to a registration page. Through that registration page, they can fill out a questionaire establishing their identity and registering their computer on the network. Combined these two tools allow for the two scenarios in which individual computers are rmoved/installed (faculty upgrades) and where large number of computers arrive on campus (students arriving on campus for the year/semester).
All of this is relatively simple(or seems that way) as long as you have a simple network infrastructure. Unfortunately, as networks grow in node size, it is imposible to manage within a simple infrastructure. Earlham's network has become such a network. Currently there are only a handfull of subnets on campus which are used to provide the network services. This is soon to change. With luck my tool will help manage the new infrastructure just as well. The new infrastructure will incorporate the heavy use of VLANs throughout campus. Machine names will be based upon the wiring closet to which they are connected (or that is one possibility).
So now I have had some sleep and can think clearly...
Now... Where was I? Oh yeah, I was explaining the way that the network here at earlham was changing. As I said, in some cases a computer's name will be determined by where it was/will be located. My tool has to support this and allow for the setup of a new computer in a different location than its final destination. This is to allow for ECS to set up the computer in Lilly before delivering it to the faculty's desk. To my knowledge, Student machines will continue to be set up in the same fashion as they are now where the owner's username is pre-pended onto student.earlham.edu. The later of these two scenarios is easier than the former in that there is no variability in the name. I still have lots to discuss with ECS to figure out what is possible and what they want to do. Hopefully I will get that out of the way soon and get further into the project development stage.