I'm in my final year of (3yr program) college and I need to create a large technical project to complete my program. I was going to do a web-based scheduling and grading system with PHP and mysql but I'm not really interested in that anymore.

I'm thinking of creating a syslog daemon which will recieve the facility data from devices on the network and then log appropriately and send alarms and email stats etc....

My question is: Would this be to complicated or to easy. I've been programming C++/C for about 3 years now. I have worked with MFC/Win32 API, Xlib, Athena and QT, and regular text apps on Linux and Windows. I plan to write the front end in Qt.

Also does anyone else have any ideas what else I could create? My friend is already making a sniffer and a front end for nmap port scanner so that out of the question. I have 4 months from September to complete the project, I will work on it while I'm in full time classes. The requirements are that we learn something thing and meet some of the goals we set out for ourselves, not that it works 100%.

any input appreciated!

Sounds like a memo from the marketing department to the coders at Microsoft.

why dont you try implementing a protocol or modifying an existing one. you will have yo get hold of some really good book for that. and the best one i can suggest is Douglas E Comer - TCP/IP illustrated. there are actually 2 parts, the first one has got all the concepts with detailed explanation and teh other one focuses on implementation details.

simulations are also a pretty good idea

Check previous projects and see what they didn't accomplish and capitialize on that.

For instance in my final year at my high school nobody knew anything about

java networking or Win32API in borland C++ and so I did those and they granted my great marks

as for your syslog daemon

I'm not sure what your saying, is it a database that recieves information from clients?

If so that doesn't seem too hard because because
its a small simple database and networking program. However, if you were to be able to create graphs from the data, have security features, limit bandwidth etc.. that would be a much better step. But again it really depends on your own programming abilities, your prof's expectations and your classes ability.

As for the webbased grading system it almost seems like an intro to arrays and networking, more like course work then creative programming (I'm sorry if I sound cocky, but in high school we did grading programs for 5 different languages, it was SICK!)

Personally I think it would be really cool if you could colaborate with your friend, him making "cyber criminal software" and you making "cyber patrol software"

From what I've learned over the years you basically only have 2 types of projects

BIG time consuming repeatitive (ie. data base)
or
Small creative original (something original to the class)

where both would garuntee a good mark

Another idea would be creating a 3D world with Win32API (C++, visual C++/MFC) on OpenGL or D3D ( I looked at OpenGL, it doesn't seem too hard if you really look at it but it does take some time)

anyway hope this helps

-John-e

What you want to do is not difficult. Of course you could also just use the inetd facilities in unix, and there wouldn't be any network programming involved. If you really wanted to institute a listener, however, you will need to become very familliar with the socket API.

Personally, i have always enjoyed client/server stuff. Almost anything that involves networking is inherently interesting. I'd find a syslogd clone to be a bit boring, but maybe it's the way you're describing it. It sounded like you were actually trying to do something similar to snmp.

Thanks for the input

Yea the grading thing was doomed from the start, it was a rush decsion so I could complete another course in January based on this project, I actually did a grade system in high school as well in C .

The bandwidth idea is great, however something very very simular is being done by another student, so I think I might avoid that area.

The syslog isn't terribly difficult or glamorous like a sniffer, or bandwidth control etc. I realize that UNIX uses a regular syslogd and their are others available.

I'm trying to examine interprocess communication techniques as well as network socket programming. I plan to have the daemon and then a completely separate GUI app done in Qt, and communicate between the two. The GUI will be used for monitoring, analysis and administration. I've even toyed with the idea of making it a sniffer type of daemon so that captures the messages in transit.

Basically one could point their router, VPN box or other network device to the machine running the daemon, and send their messages based on the priority and level set. Example someone logs into a VPN box and an email message could be instantly sent to admin or support staff. Messages would depend on the device sending the data. It would then monitor the messages sent and take appropriate action based on the rules defined. I became interested in this because for the summer I work for an organization with a very large WAN (1200+ routers), and I did a lot of work with setting up various logging tools for them.

I guess it's not overly exciting, but I think it covers the areas I'm interested in learning about, networking and processes running on UNIX.

Your syslog idea should be easier to implement than you might think. Most networked devices (computers, routers, printers, etc) implement a protocol called SNMP (simple network management protocol). Using this protocol you can ask a device for information or trap information coming from a device.
SNMP is well defined and somewhat easy to use.

I wouldn't go the sniffer (or cracking) route because it's already been done and there may be some liability involved. I've always suggested projects such as build a beuwolf cluster. Write a program that uses ICMP to discover a network layout. Take iptables, squid, and lvs (www.LinuxVirtualServer.org) and build a transparent caching proxying firewall with load balancing support. (You'll spend serious money buying M$ software or Cisco hardware to do that) It's very practical and a good learning experience. Build and demo a load balancing webserver. The possibilities are endless.