our customer finally set up a radius server for us to play with, but something must still be wrong with their setup. i can send them the right radius packets, but none of their responses get back to me at all. wonder what the heck is wrong with them? i mean, when i test locally on my PC, everything's ok, and when i test against kabo's radius server on his PC, everything's ok too. its just when my program 'talks' to the customer's radius server, that's when i dont get any response back. now their logs show positively that they're getting my packets and that my packets are all valid, and that they're trying to respond to me properly, but somehow i never get their response.... hmmmmmmmmmmmm.........

on the other hand, this idle time i'm enjoying has allowed me to finally catch up on the rest of the issues of MegaTokyo up to the present! W00T! T|-|3 3V||_ IS 3V3RYWH3R3!

alrighty then! my IVR module is DONE! authentication with RADIUS is DONE! previously, the IVR module isn't fully integrated with my radius module, but now they are! calls can already get authenticated and get through! i just earned 1 million points with my boss, hedwig! yeeehhaaaa!

all that's left now is for our customer to give us the details of their RADIUS system and to test against their own radius box. following this up with our customer isn't my job anymore, though, so.... until the customer sets up a test box for me to play with, i can sit back and relax....

thesis update
alright, so i'll be doing a packet sniffing agent. let me now summarize what i'm supposed to do with my thesis...

my thesis will cover three approaches to web services-based network management:

1) Service Oriented Architecture approach
2) Peer-to-Peer Approach
3) SNMP-SOAP Hybrid Approach

in all cases my agent should be able to provide info on a node, all its interfaces, and at the very least, bandwidth utilization on each interface. again, for bandwidth utilization monitoring, i'll need to capture packets on the interface.

i've already sent my new proposal to ma'm mida and sir andrew, but i'm not really sure if its still in time for propsal defense on Jan 27. i had forgotten that i should submit a proposal well before jan 27 to give enough time for my panelists to review my proposal. *cross fingers*!

packet sniffing agent
for on and off the past few months, i've been thinking about how to write an agent that measures bandwidth utilization for a network interface. i'm thinking it should make a good feature to put into my thesis, into the agent i'll be doing. i asked PLUG (Philippine Linux Users Group) about it, and someone pointed me to netdevice as a means of programatically talking with the network interface device and getting/setting its info (IP Address, subnet mask, etc). while this was really interesting, i had guessed back then that in order to measure bandwidth utilization, i'd have to capture packets on the interface. this morning i remembered to look up the topic once more, until i finally thought of looking at freshmeat for bandwidth monitoring software. i downloaded two such software and looked at their sources. in both instances, they used libpcap and actually captured packets in order to measure bandwidth utilization. so, that settles it then -- i am going to have to capture packets to measure bandwidth afterall. this should be really interesting!

just think about it -- i could create a SOAP agent that sniffs packets on a network! actually, for me to measure bandwidth, i WOULD have to sniff packets; its just that my agent should filter the packets so that in computing for bandwidth utilization, only packets destined to and coming from the network interface should be taken into account. now, take away that filter, and i'll just have a plain old packet sniffing agent! coolness!

PH33R MY L33T H4C|<|NG S|<|LLZ!

hehehehe.... got that one from reading Megatokyo. that online comic is darned funny!

anyway, i finally finished my IVR Module... sweet! the softswitch now answers something like "Welcome! Enter your PIN followed by the # sign...." and "Enter the number you wish to call, followed by a # sign". man, it works! of course, it also can actually take your pin and the number you're calling. it also validates the your call against a radius server. in other words -- IT WORKS!

once more, with more "largo" feeling:

PH33R MY L33T H4C|<|NG S|<|LLZ!

W00T!

ok so i sound like a pathetic script kiddie hacker wannabe. its just that, in the context of megatokyo, that is so funny! its not meant to be intelligent -- its SUPPOSED to be stupid, that's what's so funny about it. heck, just go browse megatokyo, willya?

alright, next on my "to do" list, is my thesis. i'll finish my document tomorrow. i hope i can still make it, but i'm not sure i can make the deadline for the defense on the 27th.

whoops! hedwig's here! time to boot to windows... *sigh*

alrighty!

i'm almost done with the IVR module for the softswitch. i think i just need to do some more stuff in the H323ConferenceSession class to accommodate the IVR module. i just haven't figured out where i'm getting this weird error... after it plays the opening greeting over the line, its supposed to collect the PIN, but it stops at sending the sound prompt for entering the pin ("enter your pin followed by the # sign" or something like that). i get this weird error that looks like a buffer overflow. hmmmmmmmmmmmmmmmmmmm.

in other news, me and my fellow "floaties" are scheduled to have to defend our thesis proposals this January 27. i should be able to finish my proposal by this weekend, with more than enough time to spare for polishing up before the 27th.

ooooh, and i got this new book, "Unix Systems Programming" by Robbins and Robbins. it rocks! it covers every POSIX system call in existence! i know most of the calls are in C, but hey, what do you expect? the book IS all about doing Unix system calls , afterall, which are nothing more than a collection of function calls and the associated structs that go with them. maybe i can do a complete object-oriented library around POSIX someday... cool!

*sigh* as for richard, he's got his head all filled up with that "Holy Blood, Holy Grail" garbage. so much for a potential "seeker of the Light"...

i'm beginning to think that revealing my occult background to him was a mistake. ah well, if it was a mistake, its just a small one (so help me God); i didn't reveal EVERYTHING to him. all he knows is i'm some sort of a long-time occultist, thats all. he doesn't know enough to do any real damage to me, in case he turns against me.

... and i'm off to work! well, shower first... then work!

thesis update
oh yeah, i forgot to mention... yesterday i managed to fix my "hello world" gsoap web service version. turns out, i had been doing the output parameter all wrong -- it should be:

[*] & result

i was doing it as

[*] * result

it seems i got mislead by one of the compiler error messages i once got, stating that the output parameter should be * * not &....

also, using apache axis' tcpmon tool to "sniff" the xml messages between web service and client really helped with my debugging!

my next step now is to delve deeper into linux systems programming. if i could figure out how to programatically configure the routing table, for example, i could make a web service with enough functionality to replace a typical snmp agent.

good luck to me!

Posted by lone_geek at 12:54 PM WST

Updated: Tuesday, 6 January 2004 12:56 PM WST

Share This Post

Post Comment | Permalink

i managed to convince kabo to change his snmp management console so that it just uses "create and go" to tell my agent to create a new row in a table. shortly afterwards, i figured out what was actually going on in my code, why it always messes up the config file when rows are added via "create and wait". fixed it all up this morning and now my agent can work with either "create and wait" or "create and go" requests.

after this i'll be doing that IVR module for the softswitch. and after that, i'll be doing the fax support module too.

this all reminds me.... i wonder when that new C++ guy will report to work? i would have thought that he should be here by now.... ???

HAPPY NEW YEAR!
and what a way to start the New Year by building my very first Zoid! i bought a Liger Zero Phoenix yesterday and i finished the Liger just in time for 12 midnight! YAY! i'll try to post some pictures soon!

on another note... i now consider my brother a lost cause. he's gone and become a born-again zealot. he's now one of those "chosen by God above all others" who think that everybody else is in a poor, pitifully bad way. ah well... i guess i'll have to warn our mother about this sooner or later...

speaking of mother, i lent my mother a copy of "Blessed Among Women". ok so all this isn't so geeky, but its the new year, there's no work today, and therefore theres nothing geeky to write! anyway, my mother seems a bit incredulous... she's already started reading it but she asked if this book wasn't a work of fiction or something... naturally i said, no, and that it isn't published under the category of "fiction" . i hope she reads it through and through. and i hope she really likes the book in the end.

i also hope my finger heals okay. of all the gosh-darned things to happen on new year's eve -- i cut my left middle finger while working on my new Zoid! i wont describe it in detail here; its too gross even for me to just remember it! suffice to say, typing on the keyboard now feels kind of funny.... *sigh*

ok, things to do this year... (or, if you really must put it this way, "new year's resolutions")...

number 1: sleep early! i really have to will myself to get rid of this insomnia!

number 2: be good! ok maybe that sounds like a tall order, but so help me God, i sure as heck will try not to entertain thoughts and feelings of criticism, condemnation, judgement and blame. there's a perfectly good explanation for this, but i don't think this is the right blog to put it in. i'll probably have to put up a "spiritual" blog for this if i really want to...

number 3: actually number 2 already pretty much sums everything up, so there IS no number 3!

looking ahead, i hope to finish my Master's Degree by May this year. i also hope to get into my sister in-law's fashion jewelry business later this year, and try to help out any way i can.

now beyond all this, i can't really look that far ahead so i'll just end by saying all in all, i feel good about the past year and i feel good looking at this new year and the months to come.

HAPPY NEW YEAR! PEACE ON EARTH AND GOODWILL TO ALL MEN!

got my paycheck today. i got docked one day and a few hours pay -- i timed out before my 8 hrs were filled in to go to our own xmas party and then i was absent the next day due to horrendous traffic. how's that for getting a formal timecard-bundy clock system implemented? since we're now into the timecard business anyway, i think we might as well implement formal leave policies -- such as leave forms and sick leave, vacation leave, emergency leave, etc. i think i'll email jaime about it in a polite way tomorrow or later this evening.

ah well. happy new year to you too!

of course i could say that at my pay level i ought not to have any real attendance requirements anymore, like most managers. but then bringing that up would open up a whole other can of worms, which i'm not quite ready to tackle yet.

anyway, when i email good ol' umbridge; i'll try not to wear my rebel bandana and take out my trusty .45 and my good ol' M4A1 while i'm at it. i'll just be a civilized person talking to another civilized person. i hope!